v EPA
United States February 2006
Ag™ " EPA 600/R-05/004aF
Air Quality Criteria for
Ozone and Related
Photochemical Oxidants
Volume I of
-------
EPA 600/R-05/004aF
February 2006
Air Quality Criteria for Ozone and Related
Photochemical Oxidants
Volume I
National Center for Environmental Assessment-RTF Office
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
PREFACE
National Ambient Air Quality Standards (NAAQS) are promulgated by the United States
Environmental Protection Agency (EPA) to meet requirements set forth in Sections 108 and 109
of the U.S. Clean Air Act (CAA). Sections 108 and 109 require the EPA Administrator (1) to
list widespread air pollutants that reasonably may be expected to endanger public health or
welfare; (2) to issue air quality criteria for them that assess the latest available scientific
information on nature and effects of ambient exposure to them; (3) to set "primary" NAAQS to
protect human health with adequate margin of safety and to set "secondary" NAAQS to protect
against welfare effects (e.g., effects on vegetation, ecosystems, visibility, climate, manmade
materials, etc); and (5) to periodically review and revise, as appropriate, the criteria and NAAQS
for a given listed pollutant or class of pollutants.
In 1971, the U.S. Environmental Protection Agency (EPA) promulgated National Ambient
Air Quality Standards (NAAQS) to protect the public health and welfare from adverse effects of
photochemical oxidants. The EPA promulgates the NAAQS on the basis of scientific
information contained in air quality criteria issued under Section 108 of the Clean Air Act.
Following the review of criteria as contained in the EPA document, Air Quality Criteria for
Ozone and Other Photochemical Oxidants published in 1978, the chemical designation of the
standards was changed from photochemical oxidants to ozone (O3) in 1979 and a 1-hour O3
NAAQS was set. The 1978 document focused mainly on the air quality criteria for O3 and, to a
lesser extent, on those for other photochemical oxidants (e.g., hydrogen peroxide and the
peroxyacyl nitrates), as have subsequent revised versions of the document.
To meet Clean Air Act requirements noted above for periodic review of criteria and
NAAQS, the O3 criteria document, Air Quality Criteria for Ozone and Other Photochemical
Oxidants, was next revised and released in August 1986; and a supplement, Summary of Selected
I-ii
-------
New Information on Effects of Ozone on Health and Vegetation, was issued in January 1992.
These documents were the basis for a March 1993 decision by EPA that revision of the existing
1-h NAAQS for O3 was not appropriate at that time. That decision, however, did not take into
account newer scientific data that had become available after completion of the 1986 criteria
document. Such literature was assessed in the next periodic revision of the O3 air quality criteria
document (O3 AQCD) which has completed in 1996 and provided scientific bases supporting the
setting by EPA in 1997 of the current 8-h O3 NAAQS.
The purpose of this revised air quality criteria document for O3 and related photochemical
oxidants is to critically evaluate and assess the latest scientific information published since that
assessed in the above 1996 O3 AQCD, with the main focus being on pertinent new information
useful in evaluating health and environmental effects data associated with ambient air O3
exposures. However, other scientific data are also discussed in order to provide a better
understanding of the nature, sources, distribution, measurement, and concentrations of O3 and
related photochemical oxidants and their precursors in the environment. The document mainly
assesses pertinent literature published through 2004, but also includes assessment of a few
additional important studies published or accepted for publication in 2005.
A First External Review Draft of this O3 AQCD (dated January 2005) was released for
public comment and was reviewed by the Clean Air Scientific Advisory Committee (CASAC)
in May, 2005 to obtain. Public comments and CASAC recommendations were then taken into
account in making revisions to the document for incorporation into a Second External Review
Draft (dated August, 2005), which underwent further public comment and CASAC review at a
December, 2005 public meeting. Public comments and CASAC advice derived from review of
that Second External Review Draft were considered in making revisions incorporated into this
final version of the document (dated February, 2006). Evaluations contained in the present
document will be drawn on to provide inputs to associated O3 Staff Paper analyses prepared by
EPA's Office of Air Quality Planning and Standards (OAQPS) to pose options for consideration
by the EPA Administrator with regard to proposal and, ultimately, promulgation of decisions on
potential retention or revision, as appropriate, of the current O3 NAAQS.
Preparation of this document was coordinated by staff of EPA's National Center for
Environmental Assessment in Research Triangle Park (NCEA-RTP). NCEA-RTP scientific
staff, together with experts from other EPA/ORD laboratories and academia, contributed to
I-iii
-------
writing of document chapters. Earlier drafts of document materials were reviewed by non-EPA
experts in peer consultation workshops held by EPA. The document describes the nature,
sources, distribution, measurement, and concentrations of O3 in outdoor (ambient) and indoor
environments. It also evaluates the latest data on human exposures to ambient O3 and
consequent health effects in exposed human populations, to support decision making regarding
the primary, health-related O3 NAAQS. Lastly, the document also evaluates ambient O3
environmental effects on vegetation and ecosystems, surface level solar UV radiation flux and
global climate change, and man-made materials to support decision making on secondary
O3 NAAQS.
NCEA acknowledges the valuable contributions provided by authors, contributors, and
reviewers and the diligence of its staff and contractors in the preparation of this document.
I-iv
-------
Air Quality Criteria for Ozone and Related
Photochemical Oxidants
VOLUME I
Executive Summary E-l
1. INTRODUCTION 1-1
2. PHYSICS AND CHEMISTRY OF OZONE IN THE ATMOSPHERE 2-1
3. ENVIRONMENTAL CONCENTRATIONS, PATTERNS, AND
EXPOSURE ESTIMATES 3-1
4. DOSIMETRY, SPECIES HOMOLOGY, SENSITIVITY, AND
ANIMAL-TO-HUMAN EXTRAPOLATION 4-1
5. TOXICOLOGICAL EFFECTS OF OZONE AND RELATED
PHOTOCHEMICAL OXIDANTS IN LABORATORY ANIMALS
AND IN VITRO TEST SYSTEMS 5-1
6. CONTROLLED HUMAN EXPOSURE STUDIES OF OZONE AND
RELATED PHOTOCHEMICAL OXIDANTS 6-1
7. EPIDEMIOLOGICAL STUDIES OF HUMAN HEALTH EFFECTS
ASSOCIATED WITH AMBIENT OZONE EXPOSURE 7-1
8. INTEGRATIVE SYNTHESIS: EXPOSURE AND HEALTH EFFECTS 8-1
9. ENVIRONMENTAL EFFECTS: OZONE EFFECTS ON
VEGETATION AND ECOSYSTEMS 9-1
10. TROPOSPHERIC OZONE EFFECTS ON UV-B FLUX AND
CLIMATE CHANGE PROCESSES 10-1
11. OZONE EFFECTS ON MAN-MADE MATERIALS 11-1
I-v
-------
Air Quality Criteria for Ozone and Related
Photochemical Oxidants
(cont'd)
VOLUME II
CHAPTER 2 ANNEX (ATMOSPHERIC PHYSICS/CHEMISTRY) AX2-1
CHAPTER 3 ANNEX (AIR QUALITY AND EXPOSURE) AX3-1
CHAPTER 4 ANNEX (DOSIMETRY) AX4-1
CHAPTER 5 ANNEX (ANIMAL TOXICOLOGY) AX5-1
CHAPTER 6 ANNEX (CONTROLLED HUMAN EXPOSURE) AX6-1
CHAPTER 7 ANNEX (EPIDEMIOLOGY) AX7-1
VOLUME III
CHAPTER 9 ANNEX (ENVIRONMENTAL EFFECTS) AX9-1
I-vi
-------
age
Table of Contents
List of Tables I-xviii
List of Figures I-xxi
Authors, Contributors, and Reviewers I-xxviii
U.S. Environmental Protection Agency Project Team for Development of Air
Quality Criteria for Ozone and Related Photochemical Oxidants I-xxxix
U.S. Environmental Protection Agency Science Advisory Board (SAB) Staff Office
Clean Air Scientific Advisory Committee (CASAC) Ozone Review Panel I-xlii
ABBREVIATIONS AND ACRONYMS I-xlv
EXECUTIVE SUMMARY E-l
1. INTRODUCTION U.
1.1 LEGAL AND HISTORICAL BACKGROUND 1-1
1.1.1 Legislative Requirements 1-1
1.1.2 Criteria and NAAQS Review Process 1-3
1.1.3 Regulatory Chronology 1-4
1.2 CURRENT OZONE CRITERIA AND NAAQS REVIEW 1-8
1.2.1 Key Milestones and Procedures for Document Preparation 1-8
1.3 ORGANIZATIONAL STRUCTURE OF THE DOCUMENT 1-11
1.3.1 General Document Format 1-11
1.3.2 Organization and Content of the Document 1-12
REFERENCES 1-14
2. PHYSICS AND CHEMISTRY OF OZONE IN THE ATMOSPHERE 2J_
2.1 INTRODUCTION 2-1
2.2 CHEMICAL PROCESSES INVOLVED IN OZONE FORMATION
AND DESTRUCTION 2-2
2.3 METEOROLOGICAL PROCESSES AFFECTING OZONE 2-8
2.4 RELATIONS OF OZONE TO ITS PRECURSORS 2-15
2.5 THE ROLE OF CHEMISTRY-TRANSPORT MODELS IN
UNDERSTANDING ATMOSPHERIC OZONE 2-18
2.6 TECHNIQUES FOR MEASURING OZONE AND ITS PRECURSORS .... 2-22
2.7 SUMMARY 2-24
REFERENCES 2-27
3. ENVIRONMENTAL CONCENTRATIONS, PATTERNS, AND
EXPOSURE ESTIMATES 3-1
3.1 INTRODUCTION 3-1
3.2 AMBIENT AIR QUALITY DATA FOR OZONE 3-3
3.3 SPATIAL VARIABILITY OF OZONE IN URBAN AREAS 3-11
3.3.1 Small-Scale Horizontal and Spatial Variability in
Ozone Concentrations 3-14
I-vii
-------
Table of Contents
(cont'd)
Page
3.4 DIURNAL AND SEASONAL VARIABILITY OF OZONE 3-17
3.5 TRENDS IN OZONE CONCENTRATIONS 3-32
3.6 RELATIONSHIPS BETWEEN OZONE AND OTHER SPECIES 3-39
3.7 POLICY RELEVANT BACKGROUND OZONE CONCENTRATIONS . . . 3-44
3.8 OZONE EXPOSURE IN VARIOUS MICROENVIRONMENTS 3-55
3.9 SUMMARY OF KEY POINTS 3-76
REFERENCES 3-80
4. DOSIMETRY, SPECIES HOMOLOGY, SENSITIVITY, AND ANIMAL-TO-
HUMAN EXTRAPOLATION 4-1
4.1 INTRODUCTION 4-1
4.2 DOSIMETRY OF OZONE IN THE RESPIRATORY TRACT 4-2
4.2.1 Bolus-Response Studies in Humans 4-4
4.2.2 General Uptake Studies 4-7
4.2.3 Dosimetry Modeling 4-10
4.2.4 Summary and Conclusions - Dosimetry 4-13
4.3 SPECIES HOMOLOGY, SENSITIVITY, AND ANIMAL-TO-
HUMAN EXTRAPOLATION 4-15
4.3.1 Summary and Conclusions: Species Homology, Sensitivity,
and Animal-to-Human Extrapolation 4-22
REFERENCES 4-23
5. TOXICOLOGICAL EFFECTS OF OZONE AND RELATED PHOTOCHEMICAL
OXIDANTS IN LABORATORY ANIMALS AND IN VITRO TEST SYSTEMS 5-1
5.1 INTRODUCTION 5-1
5.2 RESPIRATORY TRACT EFFECTS OF OZONE 5-2
5.2.1 Biochemical Effects 5-2
5.2.1.1 Cellular Targets of Ozone Interaction 5-2
5.2.1.2 Monooxygenases 5-5
5.2.1.3 Antioxidants, Antioxidant Metabolism, and
Mitochondrial Oxygen Consumption 5-5
5.2.1.4 Lipid Metabolism and Content of the Lung 5-8
5.2.1.5 Ozone Interactions with Proteins and Effects
on Protein Synthesis 5-11
5.2.1.6 Differential Gene Expression 5-12
5.2.1.7 Summary and Conclusions—Biochemical Effects ... 5-13
5.2.2 Lung Host Defenses 5-14
5.2.2.1 Clearance 5-14
5.2.2.2 Alveolar Macrophages 5-15
5.2.2.3 Immune System 5-17
5.2.2.4 Interactions with Infectious Microorganisms 5-20
5.2.2.5 Summary and Conclusions—Lung Host Defenses ... 5-21
I-viii
-------
Table of Contents
(cont'd)
age
5.2.3 Inflammation and Lung Permeability Changes 5-22
5.2.3.1 Time Course of Inflammation and Lung
Permeability Changes 5-23
5.2.3.2 Concentration and Time of Exposure 5-24
5.2.3.3 Susceptibility Factors 5-25
5.2.3.4 Mediators of Inflammatory Response and Injury .... 5-28
5.2.3.5 The Role of Nitric Oxide Synthase and Reactive
Nitrogen in Inflammation 5-30
5.2.3.6 Summary and Conclusions—Inflammation
and Permeability Changes 5-32
5.2.4 Morphological Effects 5-35
5.2.4.1 Acute and Subchronic Exposure Effects 5-35
5.2.4.2 Summary of Acute and Subchronic
Morphological Effects 5-39
5.2.4.3 Subchronic and Chronic Exposure Effects 5-41
5.2.4.4 Summary and Conclusions—Subchronic and
Chronic Morphological Effects 5-44
5.2.5 Effects on Pulmonary Function 5-45
5.2.5.1 Acute and Subchronic Exposure Effects on
Pulmonary Function 5-45
5.2.5.2 Summary and Conclusions—Acute and
Subchronic Effects on Pulmonary Function 5-47
5.2.5.3 Ozone Effects on Airway Responsiveness 5-47
5.2.5.4 Summary and Conclusions—Effects on
Airway Responsiveness 5-55
5.2.6 Genotoxicity Potential of Ozone 5-56
5.2.6.1 Summary and Conclusions—Genotoxicity
Potential of Ozone 5-57
5.3 SYSTEMIC EFFECTS OF OZONE EXPOSURE 5-57
5.3.1 Neurobehavioral Effects 5-57
5.3.2 Neuroendocrine Effects 5-59
5.3.3 Cardiovascular Effects 5-59
5.3.4 Reproductive and Developmental Effects 5-61
5.3.5 Effects on the Liver, Spleen, and Thymus 5-63
5.3.6 Effects on Cutaneous and Ocular Tissues 5-63
5.3.7 Summary and Conclusions—Systemic Effects of Ozone 5-64
5.4 INTERACTIONS OF OZONE WITH OTHER CO-OCCURRING
POLLUTANTS 5-65
5.4.1 Ozone and Nitrogen Oxides 5-66
5.4.2 Ozone and Other Copollutants 5-67
5.4.3 Complex (Multicomponent) Mixtures Containing Ozone 5-69
I-ix
-------
Table of Contents
(cont'd)
age
5.4.4 Summary and Conclusions—Interactions of Ozone with
Other Co-Occurring Pollutants 5-76
5.5 EFFECTS OF OTHER PHOTOCHEMICAL OXIDANTS 5-77
5.5.1 Summary and Conclusions—Effects of Other
Photochemical Oxidants 5-79
REFERENCES 5-80
6. CONTROLLED HUMAN EXPOSURE STUDIES OF OZONE AND
RELATED PHOTOCHEMICAL OXIDANTS 6-1
6.1 INTRODUCTION 6-1
6.2 PULMONARY FUNCTION EFFECTS OF OZONE EXPOSURE IN
HEALTHY SUBJECTS 6-3
6.2.1 Introduction 6-3
6.2.2 Acute Exposure for Up to 2 h 6-4
6.2.3 Prolonged Ozone Exposures 6-6
6.2.3.1 Effect of Exercise Ventilation Rate on FEV]
Response to 6.6 h Ozone Exposure 6-6
6.2.3.2 Exercise Ventilation Rate as a Function of
Body/Lung Size on FEV] Response to 6.6 h
Ozone Exposure 6-7
6.2.3.3 Comparison of 2 h IE to 6.6 h O 3 Exposure
Effects on Pulmonary Function 6-7
6.2.4 Triangular Ozone Exposures 6-8
6.2.5 Mechanisms of Pulmonary Function Responses 6-10
6.2.5.1 Pathophysiologic Mechanisms 6-12
6.2.5.2 Mechanisms at a Cellular and Molecular Level 6-14
6.3 SUBJECTS WITH PREEXISTING DISEASE 6-16
6.3.1 Subjects with Chronic Obstructive Pulmonary Disease 6-16
6.3.2 Subjects with Asthma 6-16
6.3.3 Subjects with Allergic Rhinitis 6-18
6.3.4 Subjects with Cardiovascular Disease 6-20
6.4 INTERSUBJECT VARIABILITY AND REPRODUCIBILITY
OF RESPONSE 6-21
6.5 FACTORS MODIFYING RESPONSIVENESS TO OZONE 6-23
6.5.1 Influence of Age 6-23
6.5.2 Gender and Hormonal Influences 6-24
6.5.3 Racial, Ethnic, and Socioeconomic Status Factors 6-25
6.5.4 Influence of Physical Activity 6-25
6.5.5 Environmental Factors 6-26
6.5.6 Oxidant-Antioxidant Balance 6-27
6.5.7 Genetic Factors 6-28
6.6 REPEATED O 3 EXPOSURE EFFECTS 6-29
I-x
-------
Table of Contents
(cont'd)
Page
6.7 EFFECTS ON EXERCISE PERFORMANCE 6-30
6.8 EFFECTS ON AIRWAY RESPONSIVENESS 6-30
6.9 EFFECTS ON INFLAMMATION AND HOST DEFENSE 6-32
6.9.1 Introduction 6-32
6.9.2 Inflammatory Responses in the Upper Respiratory Tract 6-32
6.9.3 Inflammatory Response in the Lower Respiratory Tract 6-33
6.9.4 Adaptation of Inflammatory Responses 6-38
6.9.5 Effect of Anti-Inflammatory and Other Mitigating Agents 6-39
6.9.6 Changes in Host Defense Capability Following
Ozone Exposures 6-40
6.10 EXTRAPULMONARY EFFECTS OF OZONE 6-42
6.11 EFFECTS OF OZONE MIXED WITH OTHER POLLUTANTS 6-43
6.12 CONTROLLED STUDIES OF AMBIENT AIR EXPOSURES 6-44
6.12.1 Mobile Laboratory Studies 6-44
6.13 SUMMARY 6-44
REFERENCES 6-47
7. EPIDEMIOLOGIC STUDIES OF HUMAN HEALTH EFFECTS ASSOCIATED
WITH AMBIENT OZONE EXPOSURE 7-1
7.1 INTRODUCTION 7-1
7.1.1 Approach to Identifying Ozone Epidemiologic Studies 7-2
7.1.2 Approach to Assessing Epidemiologic Evidence 7-3
7.1.3 Considerations in the Interpretation of Epidemiologic Studies
of Ozone Health Effects 7-5
7.1.3.1 Exposure Assessment and Measurement Error
in Epidemiologic Studies 7-6
7.1.3.2 Ozone Exposure Indices Used 7-10
7.1.3.3 Lag Time: Period between Ozone Exposure
and Observed Health Effect 7-11
7.1.3.4 Model Specification to Adjust for Temporal
Trends and Meteorologic Effects 7-14
7.1.3.5 Confounding Effects of Copollutants 7-17
7.1.3.6 Hypothesis Testing and Model Selection in
Ozone Epidemiologic Studies 7-18
7.1.3.7 Impact of Generalized Additive Models
Convergence Issue on Ozone Risk Estimates 7-21
7.1.3.8 Summary of Considerations in the Interpretation
of Ozone Epidemiologic Studies 7-23
7.1.4 Approach to Presenting Ozone Epidemiologic Evidence 7-25
7.2 FIELD STUDIES ADDRESSING ACUTE EFFECTS OF OZONE 7-26
7.2.1 Summary of Key Findings on Field Studies of Acute
Ozone Effects from the 1996 Ozone AQCD 7-26
I-xi
-------
Table of Contents
(cont'd)
Page
7.2.2 Introduction to Recent Field Studies of Acute Ozone Effects 7-27
7.2.3 Effects of Acute Ozone Exposure on Lung Function 7-28
7.2.3.1 Spirometry (FEV^ Studies in Outdoor Worker,
Exercise, Children, Elderly, and Asthmatic Panels . . . 7-29
7.2.3.2 Peak Flow Meter (PEF) Studies in Asthmatics
and Healthy Individuals 7-40
7.2.4 Respiratory Symptoms 7-48
7.2.5 Acute Airway Inflammation 7-55
7.2.6 Acute Ozone Exposure and School Absences 7-57
7.2.7 Cardiovascular Endpoints 7-60
7.2.7.1 Cardiac Autonomic Control 7-60
7.2.7.2 Acute Myocardial Infarction 7-63
7.2.7.3 Cardiovascular Endpoints in Human
Clinical Studies 7-64
7.2.7 A Summary of Field Studies with Cardiovascular
Outcomes 7-64
7.2.8 Summary of Field Studies Assessing Acute Ozone Effects 7-65
7.3 ACUTE EFFECTS OF OZONE ON DAILY EMERGENCY
DEPARTMENT VISITS AND HOSPITAL ADMISSIONS 7-66
7.3.1 Summary of Key Findings on Studies of Emergency
Department Visits and Hospital Admissions from the
1996 Ozone AQCD 7-66
7.3.2 Review of Recent Studies of Emergency Department Visits
for Respiratory Diseases 7-66
7.3.3 Studies of Hospital Admissions for Respiratory Diseases 7-71
7.3.3.1 All-year and Seasonal Effects of Ozone on
Respiratory Hospitalizations 7-72
7.3.3.2 Potential Confounding of the Ozone Effect on
Respiratory Hospitalizations by Copollutants 7-79
7.3.4 Association of Ozone with Hospital Admissions for
Cardiovascular Disease 7-80
7.3.5 Summary of Acute Ozone Effects on Daily Emergency
Department Visits and Hospital Admissions 7-83
7.4 ACUTE EFFECTS OF OZONE ON MORTALITY 7-83
7.4.1 Summary of Key Findings on Acute Effects of Ozone on
Mortality from the 1996 Ozone AQCD 7-83
7.4.2 Introduction to Assessment of Current Ozone-
Mortality Studies 7-84
7.4.3 Single-Pollutant Model Ozone-Mortality Risk Estimates 7-85
7.4.4 Meta-Analyses of Ozone-Mortality Risk Estimates 7-94
7.4.5 Seasonal Variation in Ozone-Mortality Risk Estimates 7-97
7.4.6 Ozone-Mortality Risk Estimates Adjusting for PM Exposure . . . 7-100
I-xii
-------
Table of Contents
(cont'd)
age
7.4.7 Ozone Risk Estimates for Specific Causes of Mortality 7-103
7.4.8 Ozone-Mortality Risk Estimates for Specific Subpopulations . . . 7-108
7.4.9 Summary of Acute Ozone Effects on Mortality 7-110
7.5 EFFECTS OF CHRONIC OZONE EXPOSURE 7-111
7.5.1 Summary of Key Findings on Studies of Health Effects and
Chronic Ozone Exposure from the 1996 Ozone AQCD 7-111
7.5.2 Introduction to Morbidity Effects of Chronic Ozone Exposure . . 7-111
7.5.3 Seasonal Ozone Effects on Lung Function 7-113
7.5.4 Chronic Ozone Exposure Effects on Lung Function and
Respiratory Symptoms 7-115
7.5.5 Chronic Ozone Exposure and Respiratory Inflammation 7-122
7.5.6 Risk of Asthma Development 7-124
7.5.7 Respiratory Effects of Chronic Ozone Exposure on
Susceptible Populations 7-126
7.5.8 Effects of Chronic Ozone Exposure on Mortality and
Cancer Incidence 7-127
7.5.9 Effects of Ozone on Birth-Related Health Outcomes 7-131
7.5.10 Summary of Chronic Ozone Exposure Effects on Morbidity
and Mortality 7-134
7.6 INTERPRETIVE ASSESSMENT OF THE EVIDENCE IN
EPIDEMIOLOGIC STUDIES OF OZONE HEALTH EFFECTS 7-134
7.6.1 Introduction 7-134
7.6.2 Ozone Exposure Indices 7-135
7.6.3 Confounding by Temporal Trends and Meteorologic Effects
in Time-Series Studies 7-137
7.6.3.1 Assessment of Ozone Effects after Adjusting for
Temporal Trends and Meteorologic Effects 7-138
7.6.3.2 Importance of Season-Specific Estimates of
Ozone Health Effects 7-141
7.6.4 Assessment of Confounding by Copollutants 7-148
7.6.4.1 Relationship between Personal Exposure to
Ozone and Copollutants 7-149
7.6.4.2 Assessment of Confounding Using
Multipollutant Regression Models 7-150
7.6.5 Concentration-Response Function and Threshold 7-154
7.6.6 Heterogeneity of Ozone Health Effects 7-159
7.6.7 Health Effects of Ozone in Susceptible and Vulnerable
Populations 7-163
7.6.7.1 Health Effects Associated with Ambient
Ozone Exposure in Asthmatics 7-164
7.6.7.2 Age-Related Differences in Ozone Effects 7-168
I-xiii
-------
Table of Contents
(cont'd)
age
7.6.7.3 Vulnerability of Outdoor Workers and Others
Who Participate in Outdoor Activities to Ozone
Health Effects 7-171
7.6.8 Summary of Key Findings and Conclusions Derived from
Ozone Epidemiologic Studies 7-174
REFERENCES 7-178
8. INTEGRATIVE SYNTHESIS: OZONE EXPOSURE AND HEALTH EFFECTS 8J,
8.1 INTRODUCTION 8-1
8.1.1 Chapter Organization 8-2
8.2 AMBIENT OZONE AIR QUALITY IN THE UNITED STATES 8-3
8.2.1 Current Ozone Concentrations and Spatial Patterns 8-3
8.2.2 Diurnal and Seasonal Variations 8-4
8.2.3 Long-Term Trends 8-4
8.2.4 Interrelationships Between Ozone and Other Ambient Pollutants . . 8-5
8.2.5 Policy Relevant Background (PRB) Ozone Concentrations 8-6
8.3 FACTORS AFFECTING HUMAN EXPOSURE TO AMBIENT OZONE . . . . 8-8
8.3.1 Personal Exposure 8-8
8.3.2 Indoor Concentrations 8-9
8.4 SYNTHESIS OF AVAILABLE INFORMATION ON OZONE-
RELATED HEALTH EFFECTS 8-10
8.4.1 Integration of Experimental and Epidemiologic Evidence 8-12
8.4.1.1 Cross-Cutting Issues Relevant to Assessment/
Interpretation of Ozone Health Effects 8-12
8.4.1.2 Dosimetry 8-14
8.4.2 Experimental Evidence for Ozone-Related Health Effects 8-15
8.4.3 Biological Basis for O3 Health Effects Assessment 8-28
8.4.4 Epidemiologic Evidence 8-32
8.4.4.1 Acute Ozone Exposure Studies 8-33
8.4.4.2 Chronic Ozone Exposure Studies 8-39
8.4.4.3 Summary of the Epidemiologic Evidence 8-40
8.5 ASSESSMENT OF POTENTIAL THRESHOLDS 8-42
8.6 BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE
FOR OZONE-RELATED HEALTH EFFECTS 8-44
8.6.1 Acute Ozone Exposure-Induced Health Effects 8-45
8.6.2 Chronic O3 Exposure-Induced Health Effects 8-50
8.6.3 Mortality-Related Health Endpoints 8-51
8.6.4 Health Effects of Ozone-Containing Pollutant Mixtures 8-53
8.7 SUSCEPTIBLE AND VULNERABLE POPULATIONS, AND
POTENTIAL PUBLIC HEALTH IMPACTS 8-55
8.7.1 Preexisting Disease as a Potential Risk Factor 8-56
I-xiv
-------
Table of Contents
(cont'd)
age
8.7.2 Age-Related Variations in Susceptibility/Vulnerability 8-59
8.7.3 Vulnerability of Outdoor Workers and Others Who
Participate in Outdoor Activities 8-61
8.7.4 Genetic Factors Affecting O 3 Susceptibility 8-63
8.7.5 Potential Public Health Impacts 8-65
8.7.5.1 Concepts Related to Defining of Adverse
Health Effects 8-65
8.7.5.2 Estimation of Potential Numbers of Persons
in At-Risk Susceptible Population Groups in
the United States 8-70
8.8 SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS . . . 8-73
REFERENCES 8-83
APPENDIX 8 A: Summary of New Animal Toxicology, Human Clinical,
and U.S./Canadian Epidemiologic Studies of Health Effects
Associated with Ambient or Near-Ambient Ozone Exposures 8A-1
9. ENVIRONMENTAL EFFECTS: OZONE EFFECTS ON VEGETATION
AND ECOSYSTEMS 9-1
9.1 INTRODUCTION 9-1
9.2 METHODOLOGIES USED IN VEGETATION RESEARCH 9-3
9.3 SPECIES RESPONSE AND MODE-OF-ACTION 9-6
9.4 MODIFICATION OF FUNCTIONAL AND GROWTH RESPONSES 9-7
9.5 EFFECTS-BASED AIR QUALITY EXPOSURE INDICES 9-11
9.6 OZONE EXPOSURE-PLANT RESPONSE RELATIONSHIPS 9-15
9.7 EFFECTS OF OZONE EXPOSURE ON NATURAL ECOSYSTEMS 9-19
9.8 ECONOMICS 9-22
REFERENCES 9-25
10. THE ROLE OF TROPOSPHERIC OZONE IN UVB-RELATED HUMAN
HEALTH OUTCOMES AND IN CLIMATE CHANGE 10-1
10.1 INTRODUCTION 10-1
10.2 THE ROLE OF TROPOSPHERIC OZONE IN DETERMINING
GROUND-LEVEL UV-B FLUX 10-1
10.2.1 Factors Governing Ultraviolet Radiation Flux at the
Earth's Surface 10-1
10.2.1.1 UV Radiation: Wavelengths and Energies 10-2
10.2.1.2 Temporal Variations in Solar Flux at the
Earth's Surface 10-4
10.2.1.3 Atmospheric Radiative Interactions with
Solar Ultraviolet Radiation 10-5
I-xv
-------
Table of Contents
(cont'd)
age
10.2.1.4 Data Requirements for a Surface
UV-B Climatology 10-15
10.2.2 Factors Governing Human Exposure to Ultraviolet Radiation ... 10-16
10.2.2.1 Outdoor Activities 10-16
10.2.2.2 Occupation 10-18
10.2.2.3 Age 10-18
10.2.2.4 Gender 10-19
10.2.2.5 Geography 10-20
10.2.2.6 Protective Behavior 10-20
10.2.2.7 Summary of Factors that Affect Human
Exposures to Ultraviolet Radiation 10-21
10.2.3 Factors Governing Human Health Effects due to
Ultraviolet Radiation 10-21
10.2.3.1 Erythema 10-22
10.2.3.2 Skin Cancer 10-24
10.2.3.3 Ultraviolet Radiation Exposure and the
Incidence of Nonmelanoma Skin Cancers 10-25
10.2.3.4 Ocular Effects of Ultraviolet Radiation
Exposure 10-32
10.2.3.5 Ultraviolet Radiation and Immune System
Suppression 10-34
10.2.3.6 Protective Effects of Ultraviolet Radiation—
Production of Vitamin D 10-35
10.2.4 Summary and Conclusions for Ozone Effects on UV-B Flux .... 10-37
10.3 TROPOSPHERIC OZONE AND CLIMATE CHANGE 10-38
10.3.1 The Projected Impacts of Global Climate Change 10-39
10.3.2 Solar Energy Transformation and the Components of the
Earth's Climate System 10-43
10.3.3 The Composition of the Atmosphere and the Earth's
Radiative Equilibrium 10-44
10.3.3.1 Forcing of the Earth's Radiative Balance 10-45
10.3.4 Factors Affecting the Magnitude of Climate Forcing
by Ozone 10-47
10.3.4.1 The Global Burden of Tropospheric Ozone 10-48
10.3.4.2 Background Concentrations versus Regionally-
Oriented Ozone Enhancements 10-49
10.3.4.3 Ozone Trends: Globally and in North America .... 10-51
10.3.4.4 The Sensitivity of Ozone-Related Forcing
Surface to Albedo 10-54
10.3.4.5 The Altitude Dependence of Forcing by
Tropospheric Ozone 10-54
I-xvi
-------
Table of Contents
(cont'd)
Page
10.3.4.6 Co-occurrence of Ozone with Particulate Matter ... 10-54
10.3.5 Estimated Forcing by Tropospheric Ozone 10-55
10.3.5.1 Direct Climate Forcing Due to Ozone 10-55
10.3.5.2 Indirect Forcing Due to Ozone 10-57
10.3.5.3 Predictions for Future Climate Forcing by
Anthropogenic Ozone 10-58
10.3.6 The Impact of a Warming Climate on Atmospheric
Ozone Concentrations 10-59
10.3.7 Conclusion 10-59
REFERENCES 10-61
11. OZONE EFFECTS ON MAN-MADE MATERIALS 11-1
11.1 ELASTOMERS 11-1
11.2 TEXTILES AND FABRICS 11-3
11.3 DYES, PIGMENTS, AND INKS 11-4
11.4 ARTISTS' PIGMENTS 11-5
11.5 SURFACE COATINGS 11-12
11.6 CONCLUSIONS 11-13
REFERENCES 11-15
I-xvii
-------
List of Tables
Number Page
1-1 National Ambient Air Quality Standards (NAAQS) for Ozone 1-5
1-2 Key Milestones for Development of Revised Ozone Air Quality
Criteria Document 1-10
3-1 Summary Statistics for the Spatial Variability of O3 (in ppm) in Selected
Urban Areas in the United States 3-13
3-2 Previous Estimates of Background O3 in Surface Air Over the United States 3-48
3-3 Personal Exposure Concentrations 3-64
3-4 Indoor/Outdoor Ozone Concentrations in Various Microenvironments 3-65
5-1 Summary of Studies that Evaluated Morphological Effects of a Single
Acute O3 Exposure 5-40
7-la Field Studies that Investigated the Association Between Acute Ambient O3
Exposure and Changes in FEV] in Adults 7-30
7-lb Percent Changes in FEV] (95% CI) Associated with Acute Ambient O3
Exposures in Adults, Ordered by Size of the Estimate 7-31
7-1 c Cross-day Percent Changes in FEV] (95% CI) Associated with Acute
Ambient O3 Exposures in Adults, Ordered by Size of the Estimate 7-32
7-2a Field Studies that Investigated the Association Between Acute Ambient O3
Exposure and Changes in FEV] in Children 7-33
7-2b Percent Changes in FEV! (95% CI) Associated with Acute Ambient O3
Exposures in Children, Ordered by Size of the Estimate 7-34
7-2c Cross-day Percent Changes in FEV] (95% CI) Associated with Acute
Ambient O3 Exposures in Children, Ordered by Size of the Estimate 7-35
7-3 Difference in Annual Percent Increases in Lung Function from the Least
to the Most Polluted Community in the Children's Health Study by Time
Spent Outdoors 7-119
I-xviii
-------
List of Tables
(cont'd)
Number Page
8-1 Acute O3-Induced Physiological and Biochemical Changes in Human
and Animals 8-29
8-2 Gradation of Individual Responses to Short-Term Ozone Exposure in
Healthy Persons 8-67
8-3 Gradation of Individual Responses to Short-Term Ozone Exposure in Persons
with Impaired Respiratory Systems 8-68
8-4 Prevalence of Selected Cardiorespiratory Disorders by Age Group and
by Geographic Region in the United States (2002 [U.S. Adults] and 2003
[U.S. Children] National Health Interview Survey) 8-71
8-5 Acute Respiratory Conditions per 100 Persons/Year by Age Group in the
United States (1996 National Health Interview Survey) 8-72
8A-1 Short-Term Ozone-Induced Health Effects Observed in Controlled Human
Exposure Studies 8A-2
8A-2 Effects of Acute O3 Exposure on Lung Function in the U.S. and Canada 8A-6
8A-3 Effects of Acute O 3 Exposure on Asthma Emergency Department Visits
in the U.S. and Canada 8A-10
8A-4 Effects of Acute O3 Exposure on Total Respiratory and Asthma Hospital
Admissions in the U.S. and Canada 8A-13
8A-5 Effects of Acute O3 Exposure on All-Cause Mortality in the U.S. and Canada .... 8 A-18
8A-6 Toxicological Effects of Acute Ozone Exposure in Animals 8A-23
10-1 Examples of Impacts Resulting From Projected Changes in Extreme
Climate Events 10-41
10-2 CTM Studies Assessed by the IPCC for its Estimate of the Change in
Global and Total Column O3 Since the Preindustrial Era 10-52
10-3 Tropospheric O 3 Change (O3) in Dobson Units (DU) Since Preindustrial
Times, and the Accompanying Net (SW plus LW) Radiative Forcings
(WrrT2), After Accounting for Stratospheric Temperature Adjustment
(using the Fixed Dynamical Heating Method) 10-56
I-xix
-------
List of Tables
(cont'd)
Number Page
11-1 Average 24-h Ozone Concentrations Producing the Highest Frequency
of Cracks of a Certain Length in the Middle and Central Zones of the
Rubber Test Strips 11-3
11-2 Cuprammonium Fluidity of Moist Cotton Cloth Exposed to 20 to 60 ppb Ozone . . 11-4
11-3 Color Change After 12 Weeks of Exposure to a Mixture of
Photochemical Oxidants 11-11
I-xx
-------
List of Figures
Number Page
2-1 Schematic overview of O3 photochemistry in the stratosphere and troposphere 2-4
2-2a Surface weather chart showing sea level (MSL) pressure (kPa), and
surface fronts 2-9
2-2b Vertical cross section along dashed line (a-a') from northwest to the southeast 2-9
2-3 The diurnal evolution of the planetary boundary layer while high pressure
prevails over land 2-11
2-4 Locations of low level jet occurrences in decreasing order of prevalence
(most frequent, common, observed) 2-12
2-5 Conceptual two-reservoir model showing conditions in the PBL and in the
lower free troposphere during a multiday O3 episode 2-13
2-6 A scatter plot of daily maximum 8-h average O3 concentrations versus daily
maximum temperature for May through September 1994 to 2004 in the
Baltimore, MD Air Quality Forecast Area 2-14
2-7 A scatter plot of daily maximum 8-h average O3 concentrations versus daily
maximum temperature for May through September 1996 to 2004 at sites
downwind of Phoenix, AZ 2-15
2-8 Measured values of O3 and NOZ (NOy - NOX) during the afternoon at rural
sites in the eastern United States (grey circles) and in urban areas and urban
plumes associated with Nashville, TN (gray dashes); Paris, France (black
diamonds); and Los Angeles CA (Xs) 2-18
2-9 Main components of a comprehensive atmospheric chemistry modeling
system, such as Models-3 2-19
3-1 Countywide mean daily maximum 8-h O3 concentrations, May to
September 2000 to 2004 3-4
3-2 Countywide 95th percentile value of daily maximum 8-h O3 concentrations,
May to September 2000 to 2004 3-5
3-3 Box plots showing daily maximum 8-h O3 averaged by month over 1993 to
2002 in the five regions in the eastern United States derived by Lehman
et al. (2004) 3-7
I-xxi
-------
List of Figures
(cont'd)
Number Page
3-4a-c Hourly average O3 concentrations observed at selected (a) rural-agricultural
(b) rural-forested, and (c) rural-residential or commercial sites for 2004 3-8
3-5a-d Daily 8-h maximum O3 concentrations observed at selected national park sites .... 3-10
3-6 Vertical profile of O3 obtained over low vegetation 3-16
3-7 Vertical profile of O3 obtained in a spruce forest 3-17
3-8 Composite, nationwide diurnal variability in hourly averaged O3 in urban areas . . . 3-18
3-9 Composite, nationwide diurnal variability in 8-h average O3 in urban areas 3-19
3-10a-f Diurnal variability in hourly averaged O3 in selected urban areas 3-20
3-10g-l Diurnal variability in hourly averaged O3 in selected urban areas 3-21
3-1 la-f Diurnal variability in 8-h O3 in selected urban areas 3-23
3-1 lg-1 Diurnal variability in 8-h O3 in selected urban areas 3-24
3-12a-d Diurnal variations in hourly averaged O3 on weekdays and weekends in
four cities 3-26
3-12e-h Diurnal variations in hourly averaged O3 on weekdays and weekends in
four cities 3-27
3-13a-d Diurnal variations in 8-h average O3 on weekdays and weekends in four cities .... 3-28
3-13e-h Diurnal variations in 8-h average O3 on weekdays and weekends in four cities .... 3-29
3-14a-f Diurnal variability in 8-h average O3 in selected urban areas 3-30
3-14g-l Diurnal variability in 8 hour averaged O3 in selected urban areas 3-31
3-15 Composite diurnal variability in hourly O3 concentrations observed at
CASTNET sites 3-33
3-16 Composite diurnal variability in 8-h O3 concentrations observed at
CASTNET sites 3-33
I-xxii
-------
List of Figures
(cont'd)
Number Page
3-17 Year-to-year variability in nationwide mean daily maximum 8-h O3
concentrations 3-34
3-18 Year-to-year variability in nationwide 95th percentile value of the daily
maximum 8-h O3 concentrations 3-35
3-19a-h Year-to-year variability in mean daily maximum 8-h O3 concentrations at
selected national park (NP), national wildlife refuge (NWR), and national
monument (NM) sites 3-36
3-20a-h Year-to-year variability in 95th percentile of daily maximum 8-h O3
concentrations at selected national park (NP), national wildlife refuge
(NWR), and national monument (NM) sites 3-37
3-21 Binned mean PM2 5 concentrations versus binned mean O3 concentrations
observed at Fort Meade, MD from July 1999 to July 2001 3-40
3-22 The co-occurrence pattern for O3 and nitrogen dioxide using 2001 data
from the AQS 3-43
3-23 The co-occurrence pattern for O3 and sulfur dioxide using 2001 data from AQS .... 3-44
3-24 The co-occurrence pattern for O3 and PM25 using 2001 data from AQS 3-45
3-25a Monthly maximum hourly average O3 concentrations at Yellowstone
National Park (WY) in 1998, 1999, 2000, and 2001 3-46
3-25b Hourly average O3 concentrations at Yellowstone National Park (WY)
for the period January to December 2001 3-46
3-26 Estimates of background contribution to surface afternoon (13 to 17 LT)
O3 concentrations in the United States as a function of local O3 concentration,
site altitude, and season 3-50
3-27 Time-series of hourly average O3 concentrations observed at five national
parks: Denali (AK), Voyageur (MN), Olympic (WA), Glacier (MT), and
Yellowstone (WY) 3-54
3-28 Hypothetical exposure time profile: pollutant exposure as a function of time
showing how the average exposure, integrated exposure, and peak exposure
relate to the instantaneous exposure 3-56
I-xxiii
-------
List of Figures
(cont'd)
Number Page
3-29 Conceptual overview of an exposure model 3-60
4-1 Structure of lower airways with progression from the large airways to
the alveolus 4-3
4-2 Ozone uptake fraction as a function of volumetric penetration (VP) in a
representative subject 4-5
4-3 Ozone uptake efficiency as a function of breathing frequency at a minute
ventilation of 30 L/min 4-9
5-1 Schematic overview of ozone interaction with epithelial lining fluid and
lung cells 5-3
5-2 The major cellular targets and proposed mechanisms of ozone toxicity in
the lung 5-6
5-3 Mechanisms of ozone toxicity 5-9
5-4 Mouse chromosomes on which genes or gene loci have been identified that
modulate responses to O3 5-33
5-5 Schematic comparison of the duration-response profiles for epithelial
hyperplasia, bronchoalveolar exudation, and interstitial fibrosis in the
centriacinar region of lung exposed to a constant low concentration of ozone 5-36
6-1 Ozone-induced changes in FEV] (top panel) and O3 concentration profiles
(bottom panel) as a function of exposure duration 6-9
6-2 Recovery of FEV] responses following a 2 h exposure to 0.4 ppm O3 with IE 6-12
6-3 Predicted O3-induced decrements in FEV] as a function of exposure duration
and level of IE (line labels are VE levels) in young healthy adults (20 yrs of age)
exposed to 0.3 ppm O3 6-26
6-4 Time course of acute responses seen in humans exposed to O3 6-35
7-1 Percent change (95% CI) in morning PEF in children per standardized
increment 7-41
7-2 Percent change (95% CI) in afternoon PEF in children per standardized
increment 7-42
I-xxiv
-------
List of Figures
(cont'd)
Number Page
7-3 Percent changes in PEF per 30 ppb increase in 8-h avg O3 in urban children 7-44
7-4 Percent change in PEF per 30 ppb increase in 8-h avg O3 with a cumulative
lag of 1 to 5 days 7-45
7-5 Odds ratios for the incidence of cough among asthmatic children per
standardized increment 7-49
7-6 Odds ratios for extra medication use among asthmatic children per
standardized increment 7-50
7-7 Odds ratio for the incidence of symptoms per 30 ppb increase in 8-h avg O3
with a cumulative lag of 1 to 4 days 7-52
7-8 Ozone-associated percent change (95% CI) in emergency department visits
for asthma per standardized increment 7-68
7-9 Ozone-associated percent change (95% CI) in total respiratory hospitalizations
for all-year analyses per standardized increment 7-73
7-10 Ozone-associated percent change (95% CI) in total respiratory hospitalizations
by season per standardized increment 7-74
7-11 Percent changes in total respiratory hospitalizations per 40 ppb increase
in 1-h max O3 in children less than two years of age during the summer
(May to August) 7-78
7-12 Ozone-associated percent change (95% CI) in total respiratory hospitalizations
with adjustment for PM indices per standardized increment 7-80
7-13 Ozone-associated percent change (95% CI) in total cardiovascular
hospitalizations per standardized increment 7-81
7-14 All cause (nonaccidental) O3 excess mortality risk estimates (95% CI) for
all-year analyses per standardized increment 7-86
7-15 All cause (nonaccidental) O3 excess mortality risk estimates (95% CI) for
all-year analyses per standardized increment 7-87
7-16 Median 24-h avg O3 concentrations (10th percentile to 90th percentile range)
for 95 U.S. communities (NMMAPS) from 1987 to 2000, arranged by
O3 concentration 7-89
I-xxv
-------
List of Figures
(cont'd)
Number Page
8-lA,B Frequency distributions of FEV] changes following 6.6-h exposures to a
constant concentration of O3 or filtered air 8-17
8-2 Frequency distributions of FEV] changes following 6.6-h exposures to a
constant concentration of O3 or filtered air 8-19
8-3 Resolution time-line for the respiratory, physiological, and biochemical
parameters are derived from studies reported in Chapter 6 and Chapter 6 Annex . . . 8-30
8-4 Acute (1-8 h) O3 exposure-induced cellular and molecular changes and
timelines for their resolution depicted here are derived from the data
reported in Leikauf et al. (1995) and Mudway and Kelly (2000) 8-31
9-1 Common anthropogenic stressors and the essential ecological attributes
they affect 9-20
10-1 Complexity of factors that determine human exposure to UV radiation 10-3
10-2 Comparison of solar flux above the atmosphere with flux at the Earth's surface . . . 10-8
10-3 Ozone column abundances from the years 1990 to 1992 for 0, 40, and 80° N
as well as 80° S 10-9
10-4 Monthly averaged vertical O3 profiles (partial pressure in mPa) as a function
of atmospheric pressure for Trinidad Head, CA; Boulder, CO; Huntsville, AL
and Wallops Island, VA 10-11
10-5 The sensitivity of ground-level UV flux to a 1 DU change in total column O3,
under clear sky conditions, as a function of solar zenith angle (SZA) 10-13
10-6 Estimated global mean radiative forcing exerted by gas and various particle
phase species for the year 2000, relative to 1750 10-47
10-7 Mid-tropospheric O3 abundance (ppb) in northern midlatitudes (36 °N-59 °N)
for the years 1970 to 1996 10-52
11-1 In-service fading of nylon 6 yarn inside house 11-6
11-2 In-service fading of nylon 6 yarn outside house 11-7
I-xxvi
-------
List of Figures
(cont'd)
Number Page
11-3 Observed color changes for natural colorant-on-paper systems during exposure
to 0.40 ppm O3 at 25 °C ± 1 °C, 50% RH, in the absence of light 11-9
11-4 Observed color changes for natural colorant-on-site during exposure to
0.40 ppm O3 at 25 °C ± 1 °C, 50% RH, in the absence of light 11-10
I-xxvii
-------
Authors, Contributors, and Reviewers
CHAPTER 1. INTRODUCTION
Principal Author
Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
CHAPTER 2 - PHYSICS AND CHEMISTR Y OF OZONE IN THE A TMOSPHERE
Principal Authors
Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Russell Dickerson—Department of Atmospheric and Oceanic Sciences, University of
Maryland, College Park, MD
Contributing Authors
Dr. Brooke Hemming—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Daniel Jacob—Department of Earth and Planetary Sciences, Harvard University,
Cambridge, MA
Dr. William Keene—Department of Environmental Sciences, University of Virginia,
Charlottesville, VA
Dr. Tadeusz Kleindienst—National Exposure Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennie Moody—Department of Environmental Sciences, University of Virginia,
Charlottesville, VA
Mr. Charles Piety—Department of Atmospheric and Oceanic Sciences, University of Maryland,
College Park, MD
Dr. Sandy Sillman—Department of Atmospheric, Oceanic, and Space Sciences, University of
Michigan, Ann Arbor, MI
I-xxviii
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributing Authors
(cont'd)
Dr. Jeffrey Stehr—Department of Atmospheric and Oceanic Sciences, University of Maryland,
College Park, MD
Dr. Bret Taubman—Department of Atmospheric Sciences, Pennsylvania State University,
State College, PA
Contributors and Reviewers
Dr. Christoph Bruhl—Max Planck Institute for Atmospheric Chemistry, Mainz, Germany
Dr. Mohammed Elshahawy—Department of Meteorology and Astronomy, Cairo University,
Giza, Egypt.
Dr. Arlene Fiore—National Oceanic and Atmospheric Administration/Geophysical Fluid Dynamics
Laboratory, Princeton, NJ
Mr. Chris Geron—National Risk Management Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. David Golden—Department of Chemistry, Stanford University, Palo Alto, CA
Dr. John Merrill—Graduate School of Oceanography, University of Rhode Island, Kingston, RI
Dr. Sam Oltmans—National Oceanic and Atmospheric Administration/Climate Monitoring and
Diagnostic Laboratory, Boulder, CO
Dr. David Parrish—National Oceanic and Atmospheric Administration/Aeronomy Laboratory,
Boulder, CO
Dr. Perry Samson—Department of Atmospheric, Ocean, and Space Sciences, University of
Michigan, Ann Arbor, MI
Dr. Sandy Sillman—Department of Atmospheric, Ocean, and Space Sciences, University of
Michigan, Ann Arbor, MI
Dr. Melvin Shapiro—National Center for Atmospheric Research, Boulder, CO
I-xxix
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 3 - ENVIRONMENTAL CONCENTRATIONS, PATTERNS,
AND EXPOSURE ESTIMA TES
Principal Authors
Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly Comfort—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Arlene Fiore—National Oceanic and Atmospheric Administration/Geophysical Fluid Dynamics
Laboratory, Princeton, NJ
Dr. Daniel Jacob—Department of Earth and Planetary Sciences, Harvard University,
Cambridge, MA
Dr. Alan S. Lefohn—ASL &Associates, Helena, MT
Dr. Clifford Weisel—Environmental and Occupational Health Sciences Institute, Rutgers
University, New Brunswick, NJ
Contributing Authors
Dr. Jee Young Kim National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis Kotchmar—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Timothy Lewis—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Thomas McCurdy—National Exposure Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Contributors and Reviewers
Dr. Christoph Bruehl—Max Planck Institute for Atmospheric Chemistry, Mainz, Germany
Dr. Russell Dickerson—Department of Atmospheric and Oceanic Sciences, University of
Maryland, College Park, MD
I-xxx
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Judith Graham—American Chemistry Council, Washington, B.C.
Dr. Laszlo Horvath—Hungarian Meteorological Service, Budapest, Hungary
Dr. Ted Johnson—TRJ Associates, Durham, NC
Dr. John Merrill—Graduate School of Oceanography, University of Rhode Island, Kingston, RI
Dr. Jennie Moody—Department of Environmental Sciences, University of Virginia,
Charlottesville, VA
Dr. Sam Oltmans—National Oceanic and Atmospheric Administration/Climate Monitoring and
Diagnostic Laboratory, Boulder, CO
Dr. Michiel G.M. Roemer, TNO, The Netherlands
Dr. Sandy Sillman—Department of Atmospheric, Ocean, and Space Sciences, University of
Michigan, Ann Arbor, MI
Dr. Tamas Weidinger—Department of Meteorology, University of Budapest, Budapest, Hungary
CHAPTER 4 - DOSIMETRY, SPECIESHOMOLOGY, SENSITIVITY,
AND EXTRAPOLA TION
Principal Authors
Dr. John Overton—U.S. Environmental Protection Agency, National Health and Environmental
Effects Research Laboratory-Research Triangle Park, NC 27711 (retired)
Dr. James S. Brown—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lori White—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
I-xxxi
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
Dr. Gary Hatch—U.S. Environmental Protection Agency, National Health and Environmental
Effects Research Laboratory, NC
CHAPTER 5 - TOXICOLOGICAL EFFECTS IN LABORATORY ANIMALS AND
IN VITRO TEST SYSTEMS
Principal Authors
Dr. Lori White—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. James Raub—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 (retired)
Dr. Deepak Bhalla—Department of Occupational and Environmental Health Sciences, Wayne State
University, Detroit, MI
Dr. Carroll Cross—School of Medicine, University of California, Davis, CA
Dr. Mitch Cohen—NYU School of Medicine, New York University, New York, NY
Contributors and Reviewers
Dr. Steven Kleeberger—National Institute of Environmental Health Sciences, Research
Triangle Park, NC 27711
Dr. George Liekauf—Department of Environmental Health, University of Cincinnati,
Cincinnati, OH
Dr. David Basset—Department of Occupational and Environmental Health Sciences, Wayne State
University, Detroit, MI
Dr. E.M. Postlethwait—Department of Environmental Health Sciences, University of Texas
Medical Branch, Galveston, TX
Dr. Kent Pinkerton—Center for Health and the Environment, University of California, Davis, CA
I-xxxii
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Edward Schelegle—Department of Anatomy, Physiology, and Cell Biology, University of
California, Davis, CA
Dr. Judith Graham—American Chemical Council, Arlington, VA
Dr. Paul Reinhart—National Center for Environmental Assessment (B243-03), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
CHAPTER 6 - CONTROLLED HUMAN EXPOSURE STUDIES
Principal Authors
Dr. James S. Brown—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. James Raub—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711 (retired)
Dr. William C. Adams—Human Performance Laboratory, University of California, Davis, CA
(retired)
Dr. Milan J. Hazucha—Center for Environmental Medicine, Asthma, and Lung Biology,
University of North Carolina, Chapel Hill, NC
Dr. E. William Spannhake—Department of Environmental Health Sciences, Johns Hopkins
University, Baltimore, MD
Contributors and Reviewers
Dr. Edward Avol—Department of Preventive Medicine, University of Southern California,
Los Angeles, CA
Dr. Jane Q. Koenig—Department of Environmental and Occupational Health, University of
Washington, Seattle, WA
Dr. Michael Madden—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Chapel Hill, NC
I-xxxiii
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr.William McDonnell—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Chapel Hill, NC
CHAPTER 7 - EPIDEMIOLOGICAL STUDIES OF HUMAN HEALTH EFFECTS
Principal Authors
Dr. Dennis Kotchmar—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jee Young Kim—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David Svendsgaard—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Kazuhiko Ito—New York University School of Medicine, Nelson Institute of
Environmental Medicine, Tuxedo, NY
Dr. Patrick Kinney—Columbia University, Mailman School of Public Health, New York, NY
Reviewers
Dr. Richard Burnett—Health Canada, Ottawa, CN
Dr. Vic Hasselblad—Duke University, Durham, NC
Dr. Lucas Neas—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Chapel Hill, NC
I-xxxiv
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 8 - INTEGRATIVE SYNTHESIS: EXPOSURE AND HEALTH EFFECTS
Principal Authors
Dr. Srikanth Nadadur—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lester Grant—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Contributing Authors
Dr. Jee Young Kim—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Mary Ross—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Lori White—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. James S. Brown—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Reviewers
Dr. John Vandenberg—National Center for Environmental Assessment, Washington, DC
Dr. Daniel Costa—National Program Director for Air, Office of Research and Development,
Research Triangle Park, NC 27711
Dr. Paul Reinhart—National Center for Environmental Assessment (B243-03), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
I-xxxv
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 9 - ENVIRONMENTAL EFFECTS ON VEGETATION AND ECOSYSTEMS
Principal Authors
Dr. Jay Garner—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711 (retired)
Dr. Timothy Lewis—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William Hogsett—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Corvallis, OR
Dr. Christian Andersen—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Corvallis, OR
Dr. Allen Lefohn—ASL and Associates, Helena, MT
Dr. David Karnosky—Forest Resources and Environmental Sciences, Michigan Technological
University, Houghton, MI
Dr. Michael Nannini—Center for Aquatic Ecology, Illinois Natural History Survey, Kinmundy, IL
Dr. Nancy Grulke—Pacific Southwest Research Station Forest Fire Laboratory, USDA Forest
Service, Riverside, CA
Dr. Richard Adams—Department of Agriculture and Resource Economics, Oregon State
University., Corvallis, OR
Dr. Robert Heath—Department of Botany and Plant Sciences, University of California,
Riverside, CA
Dr. Victor Runeckle—Biology Department, University of British Columbia, Vancouver, B.C., CN
(retired)
Dr. Arthur Chappelka—Auburn University, School of Forestry, Auburn, AL
Dr. William Massman—USDA Forest Service, Ft. Collins, CO
Dr. Robert Musselman—USDA Forest Service, Fort Collins, CO
Dr. Peter Woodbury—Cornell University, Ithaca, NY (former USDA Forest Service)
I-xxxvi
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
Dr. Fitzgerald Booker—USDA-ARS Plant Science Research Unit, 3908 Inwood Rd., Raleigh, NC
27603
Dr. Boris Chevone—Department of Plant Pathology, Virginia Technological University,
Blacksburg, VA 24061
Dr. Alan Davison—School of Biology, Newcastle University, Newcastle on Tyne,
United Kingdom, NE1 7RU
Dr. Bruce L. Dixon—Department of Agricultural Economics, University of Arkansas,
Fayetteville, AR 72701
Dr. David Grantz—Kearney Agricultural Center, University of California at Riverside, Parlier,
CA 93648
Dr. Allen S. Heagle—1216 Scott PL, Raleigh, NC 27511
Dr. Robert Horst, Jr.—121 Thorwald Dr., Plainsboro, NJ 08536
Dr. John Innes—Forest Sciences Centre, Department of Forest Resources, University of British
Columbia, Vancouver, BC, Canada V6T 1Z4
Dr. Hans-Jiirgen Jager—Heinrich-Buff-Ring 26-32, Institute of Plant Ecology, Justus-Leibig
University, Gessen, Germany D35392
Dr. Robert Kohut— Tower Road, Boyce Thompson Institute, Rm 131,Cornell University,
Ithaca, NY 14853
Dr. Sagar Krupa—1519 Gortner Ave., Department of Plant Pathology, University of Minnesota,
St. Paul, MN 55108
Dr. William Manning—203 Morrill, Department of Microbiology, University of Massachusetts,
Amherst, MA 01003
Dr. Howard Neufeld—Rankin Science Bldg., Appalachian State University, Boone, NC 28608
Dr. Paul Miller—USDA Forest Service, Pacific Southwest Research Station, 4955 Canyon Crest
Drive, Riverside CA 92507
Dr. Maria-Jose Sanz—Fundacion CEAM, c/Charles Darein, 14-Parque Te, Valencia, Spain
I-xxxvii
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. James Shortle—Department of Ag Econ, Armsby, Pennsylvania State University,
University Park, PA 16802
Dr. John Skelly—Department of Plant Pathology, Pennsylvania State University, University
Park, PA 16803
CHAPTER 10 - TROPOSPHERIC OZONE EFFECTS ON UV-B FLUX
AND CLIMATE CHANGE
Principal Authors
Dr. Brooke Hemming—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jee Young Kim—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributors and Reviewers
Dr. Sasha Madronich—Atmospheric Chemistry Division. National Center for Atmospheric
Research (NCAR), Boulder, CO 80307
Dr. Daniel J. Jacob—Atmospheric Chemistry and Environmental Engineering, Division of
Engineering & Applied Science, and Department of Earth & Planetary Sciences, Harvard
University, Cambridge, MA 02138
CHAPTER 11 - EFFECTS OF OZONE ON MAN-MADE MATERIALS
Principal Author
Mr. Bill Ewald—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711 (retired)
I-xxxviii
-------
U.S. Environmental Protection Agency Project Team
for Development of Air Quality Criteria for Ozone
and Related Photochemical Oxidants
Executive Direction
Dr. Lester D. Grant (Director)—National Center for Environmental Assessment-RTF Division,
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Scientific Staff
Dr. Lori White(Ozone Team Leader)—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly Comfort—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Brooke Hemming—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. James S. Brown—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis Kotchmar—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jee Young Kim—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David Svendsgaard—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Srikanth Nadadur—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Timothy Lewis—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jay Garner—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 (retired)
I-xxxix
-------
U.S. Environmental Protection Agency Project Team
for Development of Air Quality Criteria for Ozone
and Related Photochemical Oxidants
(cont'd)
Scientific Staff
(cont'd)
Dr. William Hogsett—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Corvallis, OR
Dr. Christian Andersen—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Corvallis, OR
Mr. Bill Ewald—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 (retired)
Mr. James Raub—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 (retired)
Technical Support Staff
Ms. Nancy Broom—Information Technology Manager, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Douglas B. Fennell—Technical Information Specialist, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Diane H. Ray—Program Specialist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Donna Wicker—Administrative Officer, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 (retired)
Mr. Richard Wilson—Clerk, National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
I-xl
-------
U.S. Environmental Protection Agency Project Team
for Development of Air Quality Criteria for Ozone
and Related Photochemical Oxidants
(cont'd)
Document Production Staff
Ms. Carolyn T. Perry—Manager, Computer Sciences Corporation, 2803 Slater Road, Suite 220,
Morrisville, NC 27560
Mr. John A. Bennett—Technical Information Specialist, Library Associates of Maryland,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
Mr. William Ellis—Records Management Technician, InfoPro, Inc., 8200 Greensboro Drive,
Suite 1450, McLean, VA 22102
Ms. Sandra L. Hughey—Technical Information Specialist, Library Associates of Maryland,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
Mr. Matthew Kirk—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road, Suite 220,
Morrisville, NC 27560
Dr. Barbara Liljequist—Technical Editor, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
Ms. Rosemary Procko—Senior Word Processor, TekSystems, 1201 Edwards Mill Road, Suite 201,
Raleigh, NC 27607
Ms. Faye Silliman—Publication/Graphics Specialist, InfoPro, Inc., 8200 Greensboro Drive,
Suite 1450, McLean, VA 22102
Mr. Carlton Witherspoon—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
I-xli
-------
U.S. Environmental Protection Agency Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC)
Ozone Review Panel
Chair
Dr. Rogene Henderson*, Scientist Emeritus, Lovelace Respiratory Research Institute, 2425
Ridgecrest Drive SE, Albuquerque, NM, 87108, Phone: 505-348-9464, Fax: 505-348-8541,
(rhenders@lrri.org) (FedEx: Dr. Rogene Henderson, Lovelace Respiratory Research Institute,
2425 Ridgecrest Drive SE, Albuquerque, NM, 87108, Phone: 505-348-9464)
Members
Dr. John Balmes, Professor, Department of Medicine, University of California San Francisco,
University of California - San Francisco, San Francisco, California, 94143, Phone: 415-206-8953,
Fax: 415-206-8949, Gbalmes@itsa.ucsf.edu)
Dr. Ellis Cowling*, University Distinguished Professor-at-Large, North Carolina State University,
Colleges of Natural Resources and Agriculture and Life Sciences, North Carolina State University,
1509 Varsity Drive, Raleigh, NC, 27695-7632, Phone: 919-515-7564 , Fax: 919-515-1700,
(ellis_cowling@ncsu. edu)
Dr. James D. Crapo*, Professor, Department of Medicine, National Jewish Medical and Research
Center. 1400 Jackson Street, Denver, CO, 80206, Phone: 303-398-1436, Fax: 303- 270-2243,
(crapoj@njc.org)
Dr. William (Jim) Gauderman, Associate Professor, Preventive Medicine, University of
Southern! California, 1540 Alcazar #220, Los Angeles, CA, 91016, Phone: 323-442-1567,
Fax: 323-442-2349, (jimg@usc.edu)
Dr. Henry Gong, Professor of Medicine and Preventive Medicine, Medicine and Preventive
Medicine, Keck School of Medicine, University of Southern California, Environmental Health
Service, MSB 51, Rancho Los Amigos NRC, 7601 East Imperial Highway, Downey, CA, 90242,
Phone: 562-401-7561, Fax: 562-803-6883, (hgong@ladhs.org)
Dr. Paul J. Hanson, Senior Research and Development Scientist, Environmental Sciences Division,
Oak Ridge National Laboratory (ORNL), Bethel Valley Road, Building 1062, Oak Ridge, TN,
37831-6422, Phone: 865-574-5361, Fax: 865-576-9939, (hansonpz@comcast.net)
Dr. Jack Harkema, Professor, Department of Pathobiology, College of Veterinary Medicine,
Michigan State University, 212 Food Safety & Toxicology Center, East Lansing, MI, 48824,
Phone: 517-353-8627, Fax: 517-353-9902, (harkemaj@msu.edu)
I-xlii
-------
U.S. Environmental Protection Agency Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC)
Ozone Review Panel
(cont'd)
Members
(cont'd)
Dr. Philip Hopke, Bayard D. Clarkson Distinguished Professor, Department of Chemical
Engineering, Clarkson University, Box 5708, Potsdam, NY, 13699-5708, Phone: 315-268-3861,
Fax: 315-268-4410, (hopkepk@clarkson.edu) (FedEx: 8 Clarkson Avenue, Potsdam, NY
136995708)
Dr. Michael T. Kleinman, Professor, Department of Community & Environmental Medicine,
100 FRF, University of California - Irvine, Irvine, CA, 92697-1825, Phone: 949-824-4765, Fax:
949-824-2070, (mtkleinm@uci.edu)
Dr. Allan Legge, President, Biosphere Solutions, 1601 11th Avenue NW, Calgary, Alberta,
CANADA, T2N 1H1, Phone: 403-282-4479, Fax: 403-282-4479, (allan.legge@shaw.ca)
Dr. Morton Lippmann, Professor, Nelson Institute of Environmental Medicine, New York
University School of Medicine, 57 Old Forge Road, Tuxedo, NY, 10987, Phone: 845-731-3558,
Fax: 845-351-5472, (lippmann@env.med.nyu.edu)
Dr. Frederick J. Miller*, Consultant, 911 Queensferry Road, Gary, NC, 27511, Phone:
919-467-3194, (fjmiller@nc.rr.com)
Dr. Maria Morandi, Assistant Professor of Environmental Science & Occupational Health,
Department of Environmental Sciences, School of Public Health, University of Texas - Houston
Health Science Center, 1200 Herman Pressler Street, Houston, TX, 77030, Phone: 713-500-9288,
Fax: 713-500-9249, (mmorandi@sph.uth.tmc.edu) (FedEx: 1200 Herman Pressler, Suite 624)
Dr. Charles Plopper, Professor, Department of Anatomy, Physiology and Cell Biology, School of
Veterinary Medicine, University of California - Davis, Davis, California, 95616, Phone:
530-752-7065, (cgplopper@ucdavis.edu)
Mr. Richard L. Poirot*, Environmental Analyst, Air Pollution Control Division, Department of
Environmental Conservation, Vermont Agency of Natural Resources, Bldg. 3 South, 103 South
Main Street, Waterbury, VT, 05671-0402, Phone: 802-241-3807, Fax: 802-241-2590,
(rich.poirot@state.vt.us)
Dr. Armistead (Ted) Russell, Georgia Power Distinguished Professor of Environmental
Engineering, Environmental Engineering Group, School of Civil and Environmental Engineering,
Georgia Institute of Technology, 311 Ferst Drive, Room 3310, Atlanta, GA, 30332-0512, Phone:
404-894-3079, Fax: 404-894-8266, (trussell@ce.gatech.edu)
I-xliii
-------
U.S. Environmental Protection Agency Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC)
Ozone Review Panel
(cont'd)
Members
(cont'd)
Dr. Elizabeth A. (Lianne) Sheppard, Research Associate Professor, Biostatistics and Environmental
& Occupational Health Sciences, Public Health and Community Medicine, University of
Washington, Box 357232, Seattle, WA, 98195-7232, Phone: 206-616-2722, Fax: 206 616-2724,
(sheppard@u. Washington, edu)
Dr. Frank Speizer*, Edward Kass Professor of Medicine, Channing Laboratory, Harvard Medical
School, 181 Longwood Avenue, Boston, MA, 02115-5804, Phone: 617-525-2275, Fax:
617-525-2066, (frank.speizer@channing.harvard.edu)
Dr. James Ultman, Professor, Chemical Engineering, Bioengineering program, Pennsylvania State
University, 106 Fenske Lab, University Park, PA, 16802, Phone: 814-863-4802, Fax:
814-865-7846, G'su@psu.edu)
Dr. Sverre Vedal, Professor of Medicine, Department of Environmental and Occupational Health
Sciences, School of Public Health and Community Medicine, University of Washington, 4225
Roosevelt Way NE, Suite 100, Seattle, WA, 98105-6099, Phone: 206-616-8285, Fax:
206-685-4696, (svedal@u.washington.edu)
Dr. James (Jim) Zidek, Professor, Statistics, Science, University of British Columbia, 6856
Agriculture Rd., Vancouver, BC, Canada, V6T 1Z2, Phone: 604-822-4302, Fax: 604-822-6960,
(jim@stat.ubc.ca)
Dr. Barbara Zielinska*, Research Professor , Division of Atmospheric Science, Desert Research
Institute, 2215 Raggio Parkway, Reno, NV, 89512-1095, Phone: 775-674-7066, Fax:
775-674-7008, (barbz@dri.edu)
Science Advisory Board Staff
Mr. Fred Butterfield, CASAC Designated Federal Officer, 1200 Pennsylvania Avenue, N.W.,
Washington, DC, 20460, Phone: 202-343-9994, Fax: 202-233-0643 (butterfield.fred@epa.gov)
(Physical/Courier/FedEx Address: Fred A. Butterfield, III, EPA Science Advisory Board Staff
Office (Mail Code 1400F), Woodies Building, 1025 F Street, N.W., Room 3604, Washington,
DC 20004, Telephone: 202-343-9994)
*Members of the statutory Clean Air Scientific Advisory Committee (CASAC) appointed by
the EPA Administrator
I-xliv
-------
ABBREVIATIONS AND ACRONYMS
cc
AA
ACh
ADSS
AER
AEROCE
AHR
AHSMOG
AIRPEX
AIRQUIS
AIRS
AM
ANF
AOP2
AOT40
APEX
APHEA
AQCD
AQS
ARIC
ATS
A/V
P
BAL
BALF
BC
BLD
alpha; probability value
ascorbic acid
acetylcholine
aged and diluted cigarette smoke
air exchange rate
Atmospheric/Ocean Chemistry Experiment
airway hyperreactivity
Adventist Health Study on Smog
Air Pollution Exposure (model)
Air Quality Information System (model)
Aerometric Information Retrieval System
alveolar macrophage
atrial natriuretic factor
antioxidant protein 2
seasonal sum of the difference between an hourly concentration at the
threshold value of 40 ppb, minus the threshold value of 40 ppb
Air Pollution Exposure (model)
Air Pollution on Health: European Approach (study)
Air Quality Criteria Document
Air Quality System
Atherosclerosis Risk in Communities (study)
American Thoracic Society
surface-to-volume ratio
beta-coefficient; slope of an equation
bronchioalveolar lavage
bronchioalveolar lavage fluid
black carbon
below limit of detection
I-xlv
-------
BS
BSA
BSA
BMZ
BP
C
C3a
CAA
CADS
CAPs
CAR
CASAC
CASTNet, CASTNET
CC16
CCSP
Cdyn, Cdyn
CDT
CE
CFCs
CFD
CFR
CH4
C2H5-H
CHAD
CH3-CC13
CH3-CHO
CH3-CO
CHO
black smoke
body surface area
bovine serum albumin
basement membrane zone
blood pressure
concentration
complement protein fragment
Clean Air Act
Cincinnati Activity Diary Study
concentrated ambient particles
centriacinar region
Clean Air Scientific Advisory Committee
Clean Air Status and Trends Network
Clara cell secretory protein
Clara cell secretory protein
dynamic lung compliance
Central Daylight Time
continuous exercise
chlorofiuorocarbons
computational fluid dynamics
Code of Federal Regulations
methane
ethane
isoprene
terpene
Consolidated Human Activities Database
methyl chloroform
acetaldehyde
acetyl
Chinese hamster ovary (cells)
I-xlvi
-------
CH3OOH
CI
CIE
CINC
CLM
CMAQ
CO
CO2
COD
COP
COPD
CRP
CTM
A
3-D
DHBA
DNA
DOAS
DPPC
DTPA
DU
e
ECG
EDU
EEC
ELF
ENA-78
ENSO
EPA
EST
acetic acid
confidence interval
Commission Internationale de 1'Eclaiarage (International Commission on
Illumination)
cytokine-induced neutrophil chemoattractant
chemiluminescence method
Community Model for Air Quality
carbon monoxide
carbon dioxide
coefficient of divergence
Conference of Parties
chronic obstructive pulmonary disease
C-reactive protein
chemistry transport model
delta; change in a variable
three-dimensional
2,3-dihydroxybenzoic acid
deoxyribonucleic acid
differential optical absorption spectroscopy
dipalmitoylglycero-3-phosphocholine
diethylenetriaminepentaacetic acid
Dobson units
epsilon; convergence precision
electrocardiographic; electrocardiogram
ethylenediurea
electroencephalographic
epithelial lining fluid
epithelial cell-derived neutrophil-activating peptide 78
El Nino-Southern Oscillation
U.S. Environmental Protection Agency
Eastern Standard Time
I-xlvii
-------
ETS
EVR
F
F344
FA
FACE
fB
FEF
FEF25.75
FEFX
FEVj
FGF
FGFR
FN
FR
FRM
FVC
GAM
GCM
GEE
GEOS-CHEM
GHG
GLM
GLRAG
GM-CSF
G6PD
GPx
GR
environmental tobacco smoke
equivalent ventilation rate
female
Fisher 344 (rat)
filtered air
free-air carbon dioxide exposure
breathing frequency
forced expiratory flow
forced expiratory flow between 25 and 75% of vital capacity; forced
expiratory flow at 25 to 75% of vital capacity
forced expiratory flow after X% vital capacity (e.g., after 25, 50, or
75% vital capacity)
forced expiratory volume in 1 second
fibroblast growth factor
fibroblast growth factor receptor
fibronectin
Federal Register
Federal Reference Method
forced vital capacity
Generalized Additive Model
general circulation model
Generalized Estimating Equation
three-dimensional model of atmospheric composition driven by
assimilated Goddard Earth Orbiting System observations
greenhouse gas
Generalized Linear Model
Great Lakes Regional Assessment Group
granulocyte-macrophage colony stimulating factor
glucose-6-phosphate dehydrogenase
glutathione peroxidase
glutathione reductase
I-xlviii
-------
GSH
GSHPx
GSTM1
GSTMlnull
H+
HCFCs
H2CO, HCHO
HDMA
HF
MFCs
HLA
FINE
HNO2, MONO
HNO3
HNO4
HO2
H2O2
HOX
HR
HRP
HRV
H2SO4
IARC
1C
ICAM
ICNIRP
IE
IFN
Ig
IL
glutathione; reduced glutathione
glutathione peroxidase
glutathione S-transferase u-1 (genotype)
glutathione S-transferase u-1 null (genotype)
hydrogen ion
hydrochlorofluorocarbons
formaldehyde
house dust mite allergen
hydrofluoride
hydro fluorocarbons
human leukocyte antigen
4-hydroxynonenal
nitrous acid
nitric acid
pernitric acid
hydroperoxyl; hydroperoxy
hydrogen peroxide
hydrogen oxides
heart rate
horseradish peroxidase
heart rate variability
sulfuric acid
International Agency for Research on Cancer
inspiratory capacity
intracellular adhesion molecule
International Commission on Non-Ionizing Radiation Protection
intermittent exercise
interferon
immunoglobulin (e.g., IgA, IgE, IgG, IgM)
interleukin (e.g., IL-1, ILIL-6, IL-8)
I-xlix
-------
iNOS
i.p.
IPCC
IQR
IR
LDH
LIDAR
LIS
LLJ
LOEL
LOESS
LOP
LPS
LRT
LT
LT
LTcc
M
M
MAP
MARAT
MCh
MCP
MED
MENTOR
MET
inducible nitric oxide synthase; NOS-2
intraperitoneal
Intergovernmental Panel on Climate Change
interquartile range
infrared
intrinsic mass transfer coefficient/parameter
mass transfer coefficient for gas phase
mass transfer coefficient for liquid phase
reaction rate constant
lactic acid dehydrogenase
Light Detection And Ranging
lateral intercellular space
low-level jet
lowest-observed-effect level
locally estimated smoothing splines
lipid ozonzation products
lipopolysaccharide
lower respiratory tract; lower airways
leukotriene (e.g., LTB4, LTC4, LTD4, LTE4)
local time
lympho toxin-cc
male
maximum number of iterations
mean arterial pressure
Mid-Atlantic Regional Assessment Team
methacholine
monocyte chemotactic protein
minimal erythema dose
Modeling Environment for Total Risk Studies
metabolic equivalent
1-1
-------
MI
MIP
MMEF
MONICA
MPAN
MPO
mRNA
MSA
MSL
MT
n,N
NAAQS
NAD(P)H
NAMS/SLAMS
NARSTO
NAS
NCEA-RTP
NCICAS
NCLAN
ND
NEM
NERAG
NF-KB
NH3
NHAPS
NH4HSO4
(NH4)2HSO4
NIH
myocardial infarction
macrophage inflammatory protein
maximal midexpiratory flow
Monitoring Trend and Determinants in Cardiovascular Disease
(registry)
peroxymethacryloyl nitrate; peroxy-methacrylic nitric anhydride
myeloperoxidase
messenger ribonucleic acid
metropolitan statistical area
mean sea level
metallothionein
number
National Ambient Air Quality Standards
reduced nicotinamide adenine dinucleotide phosphate
National Ambient Monitoring Stations and State and Local Air
Monitoring Stations
North American Regional Strategy for Atmospheric Ozone
Normative Aging Study
National Center for Environmental Assessment Division in Research
Triangle Park, NC
National Cooperative Inner-City Asthma Study
National Crop Loss Assessment Network
not detectable; not detected
National Ambient Air Quality Standards Exposure Model
New England Regional Assessment Group
nuclear factor kappa B
ammonia
National Human Activity Pattern Survey
ammonium bisulfate
ammonium sulfate
National Institutes of Health
I-li
-------
NIST
NK
NL
NM
NMHCs
NMMAPS
NO
NO2
N2O
NO3
NA
NOS
NOS-1
NOS-2
NOS-3
NOX
N0y
NOZ
NP
NPP
NQO1
NQOlwt
NRC
NS
NTP
NTS
NWR
O2
03
National Institute of Standards and Technology
natural killer (cells)
nasal lavage
national monument
nonmethane hydrocarbons
National Morbidity, Mortality and Air Pollution Study
nitric oxide
nitrogen dioxide
nitrous oxide
nitrate
nitrogen dioxide
nitric oxide synthase
neuronal nitric oxide synthase
inducible nitric oxide synthase; iNOS
endothelial nitric oxide synthase
nitrogen oxides
reactive nitrogen system components; sum of NOX andNOz; odd
nitrogen species
difference between NOy and NOX; reservoir and termination nitrogen
species
national park
net primary productivity
NAD(P)H-quinone oxidoreductase (genotype)
NAD(P)H-quinone oxidoreductase wild type (genotype)
National Research Council
national seashore
National Toxicology Program
nucleus tractus solitarius
national wildlife refuge
ground-state oxygen
ozone
I-lii
-------
o,*
1803
0('D)
OAQPS
OH
8-OHdG
0(3P)
OPE
OTC
OVA
OxComp
P
"90
PAF
PAN
Pa02
PAR
PBL
PBPK
PCI
PE
PEF
PEFR
PEM
"enh
PG
6PGD
PGP
PI
PM
electronically excited ozone
radiolabeled ozone
electronically excited oxygen atom
Office of Air Quality Planning and Standards
hydroxyl; hydroxy
8-hydroxy-2'-deoxyguanosine
ground-state oxygen atom
ozone production efficiency
open-top chamber
ovalbumin
oxidative capacity of the atmosphere
probability value
90th percentile
platelet-activating factor
peroxyacetyl nitrate; peroxyacetic nitric anhydride
partial pressure of arterial oxygen
proximal alveolar region
planetary boundary layer
physiologically based pharmacokinetic (approach)
picryl chloride
postexposure
peak expiratory flow
peak expiratory flow rate
personal exposure monitor
enhanced pause
prostaglandin (e.g., PGD2, PGE, PGE1? PGE2 PGFlK, PGF2K)
6-phosphogluconate dehydrogenase
protein gene product (e.g., PGP9.5)
probability interval
particulate matter
I-liii
-------
PM25
PM
10
PM10.2.5
PMNs
pNEM
polyADPR
ppb
ppbv
ppm
PPN
PRB
PSA
PSC
PUFA
PWM
QCE
R
r
R2
Raw
R'CO
R'C(O)-O2
RH
RL
RNA
RO2
ROOM
ROS
RR
RRMS
fine particulate matter (mass median aerodynamic diameter <2.5 jam)
combination of coarse and fine particulate matter
coarse particulate matter (mass median aerodynamic diameter between
10 and 2.5 urn)
polymorphonuclear leukocytes; neutrophils
Probabilistic National Ambient Air Quality Standard Exposure Model
poly(adenosinediphosphate-ribose)
parts per billion
parts per billion by volume
parts per million
peroxypropionyl nitrate; peroxypropionic nitric anhydride
policy relevant background
picryl sulfonic acid
polar stratospheric clouds
polyunsaturated fatty acid
pokeweed mitogen
quasi continuous exercise
intraclass correlation coefficient
correlation coefficient
multiple correlation coefficient
airway resistance
acyl
acyl peroxy
relative humidity
total pulmonary resistance
ribonucleic acid
organic peroxyl; organic peroxy
organic peroxides
reactive oxygen species
ribonucleotide reductase
relatively remote monitoring sites
I-liv
-------
RT
SAB
SAC
SAMD
SA
SBUV
SC
SD
SD
SES
sGAW
SHEDS
sICAM
SNAAQS
SNPs
SO2
S042
SOA
SOD
SOS
SP
SP
SPF
sRAW, SRa
SRES
STE
STRF
SUM06
SUM08
SZA
respiratory tract
Science Advisory Board
Staphylococcus aureus Cowan 1 strain
S-adenosyl methionine decarboxylase
oxygen saturation of arterial blood
solar backscattered ultraviolet radiation
stratum corneum
Sprague-Dawley (rat)
standard deviation
socioeconomic status
specific airways conductance
Simulation of Human Exposure and Dose System
soluble intracellular adhesion molecule
Secondary National Ambient Air Quality Standards
single nucleotide polymorphisms
sulfur dioxide
sulfate
secondary organic aerosol
superoxide dismutase
Southern Oxidant Study
substance P
surfactant protein (e.g., SP-A, SP-D)
specific-pathogen free
specific airways resistance
Special Report on Emissions Scenarios
stratospheric-tropospheric exchange
Spatio-Temporal Random Field
seasonal sum of all hourly average concentrations >0.06 ppm
seasonal sum of all hourly average concentrations >0.08 ppm
solar zenith angle
I-lv
-------
T
t
T3
T4
TAR
TAR WGI
TB
TBARS
99mTc-DTPA
TCO
T
J^CTL
TiO2
TLC
TLR
TNF
TNFR
TOMS
TRIM
TRIM.Expo
TSP
TWA
UA
UNEP
UNFCCC
URT
USGCRP
UV
UV-A
UV-B
UV-C
time; duration of exposure
£-test statistical value; t statistic
triiodothyronine
thyroxine
Third Assessment Report
Third Assessment Report of Working Group 1
terminal bronchioles
thiobarbituric acid reactive substances
radiolabeled diethylenetriaminepentaacetic acid
core temperature
cytotoxic T-lymphocytes
titanium dioxide
total lung capacity
Toll-like receptor
tumor necrosis factor
tumor necrosis factor receptor
Total Ozone Mapping/Monitoring Satellite; total ozone mapping
spectrometer
Total Risk Integrated Methodology (model)
Total Risk Integrated Methodology Exposure Event (model)
total suspended particulate
time-weighted average
uric acid
United Nations Environmental Program
United Nations Framework Convention on Climate Change
upper respiratory tract; upper airways
U.S. Global Change Research Program
ultraviolet
ultraviolet radiation of wavelengths 320 to 400 nm
ultraviolet radiation of wavelengths 280 to 320 nm
ultraviolet radiation of wavelengths 200 to 280 nm
I-lvi
-------
VC vital capacity
VD volume of the anatomic dead space
VE minute ventilation; expired volume per minute
VO2max maximal oxygen uptake (maximal aerobic capacity)
VOC volatile organic compound
VP volumetric penetration
VP50«/0 volume at which 50% of an inhaled bolus is absorbed
VT tidal volume
VTB terminal bronchiole region volume
VUA volume of the upper airways
W126 cumulative integrated exposure index with a sigmoidal weighting function
WMO World Meteorological Organization
WT wild type
I-lvii
-------
EXECUTIVE SUMMARY
E.I INTRODUCTION
Tropospheric or "surface-level" ozone (O3) is one of six major air pollutants regulated by
National Ambient Air Quality Standards (NAAQS) under the U.S. Clean Air Act. As mandated
by the Clean Air Act, the U.S. Environmental Protection Agency (EPA) must periodically
review the scientific bases (or "criteria") for the various NAAQS by assessing newly available
scientific information on a given criteria air pollutant. This document, Air Quality Criteria for
Ozone and Other Photochemical Oxidants, is an updated revision of the 1996 Ozone Air Quality
Criteria Document (O3 AQCD) that provided scientific bases for the current O3 NAAQS set in
1997.
E.I.I Clean Air Act Legal Requirements
Clean Air Act (CAA) Sections 108 and 109 govern establishment, review, and revision of
U.S. National Ambient Air Quality Standards (NAAQS).
• Section 108 directs the U.S. Environmental Protection Agency (EPA) Administrator to list
ubiquitous (widespread) air pollutants that may reasonably be anticipated to endanger public
health or welfare and to issue air quality criteria for them. The air quality criteria are to
reflect the latest scientific information useful in indicating the kind and extent of all
exposure-related effects on public health and welfare expected from the presence of the
pollutant in the ambient air.
• Section 109 directs the EPA Administrator to set and periodically revise, as appropriate, two
types of NAAQS: (a) primary NAAQS to protect against adverse health effects of listed
criteria pollutants among sensitive population groups, with an adequate margin of safety, and
(b) secondary NAAQS to protect against welfare effects (e.g., impacts on vegetation, crops,
ecosystems, visibility, climate, man-made materials, etc.). Section 109 also requires peer
review of the NAAQS and their underlying scientific bases by the Clean Air Scientific
Advisory Committee (CASAC), a committee of independent non-EPA experts.
E-l
-------
E.1.2 Chronology of Ozone NAAQS Revisions
In 1971, the U.S. EPA set primary and secondary standards for total photochemical
oxidants. Based on a criteria review completed in 1978, the original NAAQS set in 1971 were
revised in 1979 to focus on O3 as the indicator for new primary and secondary standards that
would be attained when the expected number of days per calender year with maximum 1-h
average O3 concentrations >0.12 ppm did not exceed one. The NAAQS for ambient O3 were
revised in 1997 by replacing the 1-h standards with an 8-h primary standard that is met when the
3-year average of the annual fourth highest daily maximum 8-h average concentration is
<0.08 ppm. The 1997 primary NAAQS was based on scientific data from controlled human
exposure, laboratory animal, and epidemiological studies and associated analyses presented in
the 1996 O3 AQCD and in the 1996 O3 Staff Paper.
• This revised O3 AQCD, prepared by EPA's National Center for Environmental Assessment
(NCEA), provides scientific bases to support the periodic review of O3 NAAQS. This
document assesses the latest available scientific information (published mainly through
December 2004) judged to be useful in deriving criteria as scientific bases for decisions on
possible revision of the current O3 NAAQS.
• A separate EPA O3 Staff Paper, prepared by EPA's Office of Air Quality Planning and
Standards (OAQPS), will draw upon key findings/conclusions from this document, together
with other analyses, to develop and present options for consideration by the EPA
Administrator regarding review and possible revision of the O3 NAAQS.
E.1.3 Document Organization and Structure
Volume I of this document consists of the present Executive Summary and eleven main
chapters of this revised O3 AQCD. Those main chapters focus primarily on interpretative
evaluation of key information, whereas more detailed descriptive summarization of pertinent
studies and/or supporting analyses are provided in accompanying annexes. Volume II contains
the annexes for Chapters 4 through 7, whereas Volume III contains the annex for Chapter 9.
E-2
-------
Topics covered in the main chapters of the present AQCD are as follows:
• This Executive Summary summarizes key findings and conclusions from Chapters 1 through
11 of this revised O3 AQCD, as they pertain to background information on O3-related
atmospheric science and air quality, human exposure aspects, dosimetric considerations,
health effect issues, and environmental effect issues.
• Chapter 1 provides a general introduction, including an overview of legal requirements,
the chronology of past revisions of O3-related NAAQS, and orientation to the structure of
this document.
• Chapters 2 and 3 provide background information on atmospheric chemistry/physics of O3
formation, air quality, and exposure aspects to help to place ensuing discussions of O3 health
and welfare effects into perspective.
• Chapters 4 through 7 then assess dosimetry aspects, experimental (controlled human
exposure and laboratory animal) studies, and epidemiologic (field/panel; other observational)
studies. Chapter 8 then provides an integrative synthesis of key findings and conclusions
derived from the preceding chapters with regard to ambient O3 concentrations, human
exposures, dosimetry, and health effects.
• Chapter 9 deals with effects of O3 on vegetation, crops, and natural ecosystems, whereas
Chapter 10 evaluates tropospheric O3 relationships to alterations in surface-level UVB flux
and climate change and Chapter 11 assesses materials damage (these all being key types of
welfare effects of relevance to decisions regarding secondary O3 NAAQS review).
E.2 ATMOSPHERIC CHEMISTRY AND PHYSICS OF TROPOSPHERIC
OZONE FORMATION
Key findings/conclusions from Chapter 2 regarding the chemistry and physics of surface-
level O3 formation include the following:
E-3
-------
Ozone (O3) is a secondary pollutant formed by atmospheric reactions involving two classes
of precursor compounds, volatile organic compounds (VOCs) and nitrogen oxides (NOX).
Carbon monoxide also contributes to O3 formation.
• The formation of O3 and associated compounds is a complex, nonlinear function of many
factors, including the intensity and spectral distribution of sunlight; atmospheric mixing and
other atmospheric processes; and the concentrations of precursors in ambient air.
• The photochemical oxidation of almost all anthropogenic and biogenic VOCs is initiated by
reaction with hydroxyl (OH) radicals. At night, when they are most abundant, NO3 radicals
oxidize alkenes. In coastal and other select environments, Cl and Br radicals can also initiate
the oxidation of VOCs.
• In urban areas, basically all classes of VOCs (alkanes, alkenes, aromatic hydrocarbons,
carbonyl compounds, etc.) and CO are important for ozone formation. Although knowledge
of the oxidative mechanisms of VOCs has improved in recent years, gaps in knowledge
involving key classes, such as aromatic hydrocarbons, still remain. For example, only about
half of the carbon initially present in aromatic hydrocarbons in smog chamber studies form
compounds that have been identified.
• In addition to gas phase reactions, other reactions also occur on the surfaces of or within
cloud droplets and airborne particles. Most of the well-established multiphase reactions tend
to reduce the rate of O3 formation in polluted environments. Direct reactions of O3 and
atmospheric particles appear to be too slow to reduce O3 formation significantly at typical
ambient PM levels.
• Oxidants other than O3 are found in the gas phase and in particles. The chemistry occurring
in particle bound-water and, hence, the mechanisms leading to the formation of reactive
oxygen species in particles are largely unknown.
E-4
-------
• Organic hydroperoxides produced in the oxidation of monoterpenes by O3 could contribute
substantially to secondary organic aerosol formation.
• Our basic understanding of meteorological processes associated with summertime O3
episodes has not changed over recent years. However, the realization is growing that long-
range transport processes are important for determining O3 concentrations at the surface. In
addition to synoptic scale flow fields, nocturnal low-level jets can transport pollutants
hundreds of km from their sources in either the upper boundary layer or the lower free
troposphere. Turbulence then brings O3 and other pollutants to the surface.
• Even in the absence of photochemical reactions in the troposphere, some O3 would be found
near the earth's surface due to its downward transport from the stratosphere. Intrusions of
stratospheric O3 that reach the surface are rare. Much more common are intrusions that
penetrate to the middle and upper troposphere. However, O3 transported to the middle and
upper troposphere can still affect surface concentrations through various mechanisms that
mix air between the planetary boundary layer and the free troposphere above.
• Associations between daily maximum O3 concentration and temperature vary across the
United States and depend on location. In some areas (e.g., Baltimore, MD and surrounding
areas), there is a strong positive association. In other areas (e.g., Phoenix, AZ), there is little
association.
• Chemistry transport models are used to improve understanding of atmospheric chemical and
physical processes, as well as to develop air pollution control strategies. Model evaluation
does not merely involve a straightforward comparison between model predictions and
observed concentration fields of a pollutant of interest (e.g., O3). Such comparisons may not
be meaningful because it is difficult to determine if agreement between measurements and
model predictions truly represents an accurate treatment of physical and chemical processes
in the model or the effects of compensating errors in model routines.
E-5
-------
The main methods currently used for routine monitoring of ambient ozone are based on
chemiluminescence or UV absorption. Measurements at most ambient monitoring sites are
based on UV absorption. Both of these methods are subject to interference by other
atmospheric components. Studies conducted in Mexico City and in a smog chamber have
found positive interference, but studies conducted in urban plumes did not find evidence for
significant positive interference in the UV absorption technique.
E.3 ENVIRONMENTAL DISPERSAL, AMBIENT CONCENTRATIONS,
AND HUMAN EXPOSURE TO OZONE
Key findings/conclusions derived from Chapter 3 with regard to ambient O3 concentrations
and human exposure are as follows:
• Ozone is monitored in populated areas in the United States during "ozone seasons," which
vary in length depending on location. All monitors should be operational from May to
September. However, in many areas, O3 is monitored throughout the year.
• The median of the mean daily maximum 8-h average O3 concentration from May to
September 2000 to 2004 across the U.S. was 0.049 ppm on a countywide average basis.
Ninety five per cent of countywide mean daily maximum 8-h average O3 concentrations were
less than 0.057 ppm for the same period. Because most monitors are located in the East,
these values should not be taken to represent conditions across the country.
• The daily maximum 1-h O3 concentrations tend to be much higher in large urban areas or in
areas downwind of large urban areas. For example, daily maximum 1-h O3 concentrations in
Houston, TX approached 0.20 ppm during the same period.
• Daily maximum 8-h average O3 concentrations are lower than, but are highly correlated
with, 1-h daily maximum O3 concentrations. For example, in the Baltimore, MD area, the
correlation coefficient between the two quantities was 0.98 for data obtained from May to
September 1994 to 2004.
E-6
-------
• Within individual metropolitan statistical areas (MSAs), O3 tends to be well correlated across
monitoring sites. However, there can be substantial spatial variations in concentrations.
Ozone in city centers tends to be lower than in regions either upwind or downwind of the
center, because of titration by NO emitted by motor vehicles.
• Ozone concentrations tend to peak in early- to mid-afternoon in areas where there is strong
photochemical production and later in the day in areas where transport is more important in
determining O3 abundance.
• Summertime maxima in O3 concentrations occur in areas in the United States where there is
substantial photochemical activity involving O3 precursors emitted from human activities.
Maxima can occur anytime from June through August.
• Springtime maxima are observed in relatively remote sites in the western United States and
at various other relatively unpolluted sites throughout the Northern Hemisphere. Relatively
high O3 concentrations can also be found during winter in several cities throughout the
southern United States.
• Long-term trends in O3 concentrations reflect notable decreases over time throughout the
United States, with decreases nationwide of approximately 29% in 2nd highest 1-h O3
concentrations from 1980 to 2003 and of about 21% in 4th highest 8-h O3 concentrations
during the same time period.
• These trends include dramatic decreases from peak 1-h O3 levels of 0.4 to 0.6 ppm seen in
the Los Angeles area at times in the late 1950's to 1970's to current peak levels of 0.17 ppm
and 0.15 ppm (1-h and 8-h avg, respectively) seen in the Los Angeles basin during
2000-2003.
• Downward trends in the upper tail of the O3 concentration distribution do not necessarily
reflect trends for O3 values towards the center of the O3 concentration distribution at national
parks. Concentrations toward the center of the distribution have remained more or less
E-7
-------
constant, and O3 values in the lower tail of the distribution show some evidence of slight
increases on a nationwide basis.
• Policy relevant background (PRB) O3 concentrations are used for assessing risks to human
health associated with O3 produced from anthropogenic sources in the United States, Canada
and Mexico. Because of the nature of the definition of PRB concentrations, they cannot be
derived from observations directly, instead they must be derived from model estimates.
• Current model estimates indicate that PRB O3 concentrations in the United States surface air
are generally 0.015 ppm to 0.035 ppm. Such concentrations decline from spring to summer
and are generally <0.025 ppm under conditions conducive to high O3 episodes. PRB O3
concentrations may be higher, especially at high altitude sites during the spring, due to
enhanced contributions from (a) pollution sources inside and outside North America and
(b) stratospheric O3 exchange.
• Only one model (GEOS-Chem) is documented in the literature for calculating PRB O3
concentrations. Estimated PRB O3 values are likely 10 ppbv too high in the Southeast in
summer and are accurate within 5 ppbv in other regions and seasons.
• Sufficient data for other oxidants (e.g., H2O2, PAN) and oxidation products (e.g., HNO3,
H2SO4) in the atmosphere are not available for use in epidemiologic time series studies.
Limited data for oxidants besides O3 in the gas and particle phases suggest that their
combined concentrations are probably <10 % that of O3.
• Relationships between O3 and PM2 5 are complex, in part because PM is not a distinct
chemical species, but is a mix of primary and secondary species. For example, PM25
concentrations were positively correlated with O3 during summer, but negatively correlated
with O3 during the winter at Ft. Meade, MD. Similar relationships were found for PM10 and
O3 in data collected in a number of urban areas during the 1980s.
E-S
-------
• Humans are exposed to O3 either outdoors or in various microenvironments. Ozone in
indoor environments results mainly from infiltration from outdoors. Once indoors, O3 is
removed by deposition on and reaction with surfaces and reactions with other pollutants.
Hence, O3 levels indoors tend to be notably lower than outdoor O3 concentrations measured
at nearby monitoring sites, although the indoor and ambient O3 concentrations tend to vary
together (i.e., the higher the ambient, the higher the indoor O3 levels).
• Personal exposure to O3 tends to be positively associated with time spent outdoors.
Although O3 concentrations obtained at stationary monitoring sites may not explain the
variance in individual personal exposures, they appear to serve reasonably well as surrogate
measures for aggregate personal exposures.
• Atmospheric reactions between O3 and certain other ambient airborne contaminants, e.g.,
terpenes emitted by vegetation or wood products, contribute to generation of ultrafine
particles, with formation of such particles being observed in both urban and rural areas.
These reactions also occur in indoor environments and involve O3 infiltrating from outdoors
and terpenes emitted by household products (e.g., air fresheners). Gaseous products
resulting from such reactions may also be toxic.
E.4 OZONE DOSIMETRY AND HEALTH EFFECTS
This section summarizes the main conclusions derived from the integrated synthesis of
information regarding health effects associated with ambient O3 exposures. The conclusions are
based on O3 dosimetry evaluations and human clinical, animal toxicologic, and epidemiologic
studies which have evaluated health effects associated with short-term, repeated, and long-term
exposures to O3 alone or in combination with other ambient pollutants. The controlled human
exposure (or "clinical") studies provide the clearest and most compelling evidence for human
health effects directly attributable to acute exposures to O3 per se. The evidence from human
and animal toxicologic studies presented in Chapters 4, 5, and 6 are further useful in not only
providing insights into possible mechanisms of action underlying different types of O3-related
health effects but, also, in helping to provide biological plausibility for health effects observed in
E-9
-------
epidemiologic studies assessed in Chapter 7. The studies have also been useful in identifying
susceptible and vulnerable populations that are at potentially greater risk for effects of O3
exposure. Overall, the new findings generally support and build further upon key health-related
conclusions drawn in the previous 1996 AQCD, as summarized below.
1. Dosimetric Considerations
Chapter 4 discusses dosimetric issues, including factors that are important to consider in
attempting animal-to-human extrapolations of experimentally-induced O3 effects.
• Dosimetric studies seek to quantify dose and factors affecting the dose of O3 and/or its active
metabolites at specific lung regions, target tissues, or cells. In both humans and animals, the
efficiency of O3 uptake is greater in the nasal passages than the oral pathway. In the lower
respiratory tract, increasing tidal volume increases O3 uptake, whereas increasing flow or
breathing frequency decreases O3 uptake. However, O3-induced rapid shallow breathing
appears to protect the large conducting airways while producing a more even distribution of
injury to the terminal bronchioles.
• In adult human females relative to males, the smaller airways and associated larger surface-
to-volume ratio enhance local O3 uptake and cause somewhat reduced penetration of O3 into
the distal lung. However, it is not clear from these findings if the actual anatomical location
of O3 uptake differs between males and females.
• Similarly exposed individuals vary in the amount of actual dose received, but O3 uptake is
not predictive of intersubject variability in measures of pulmonary function.
• The efficiency of O3 uptake is chemical-reaction rate dependent and the reaction products
(hydrogen peroxide, aldehydes, and hydroxyhydroperoxides) created by ozonolysis of lipids
in epithelial lining fluid (ELF) and cell membranes appear to mediate O3 toxicity.
• Ozone uptake in humans is increased by exposure to NO2 and SO2 and decreased during
the O3 exposure. This suggests that an inflammatory response during exposure to NO2 and
E-10
-------
SO2 may elicit increased production of O3-reactive substrates in the epithelial lining fluid and
that these substrates are depleted by O3 exposure but not by NO2 and SO2 exposures.
• Prior modeling studies have suggested the proximal alveolar and centriacinar regions as
principal target sites of acute O3-induced cell injury. New experimental work in rats
suggests that the conducting airways are also a primary site of injury.
• In most clinical studies, humans are exposed to O3 during exercise. Under these conditions,
the switch from nasal to oral breathing, coupled with increases in respiratory flow (as occurs
during exercise), causes a shift in the O3 dose distribution, allowing O3 to penetrate deeper
into the lung and thereby increasing the potential for bronchiolar and alveolar damage.
• Comparisons of acute exposures in rats and humans suggest that, though both species have
similar qualitative responses to O3 exposure, there are interspecies mechanistic disparities
that necessitate careful comparisons of dose-response relationships. Currently available data
suggest that lowest observable effect levels in resting rats are approximately 4- to 5-fold
higher than for exercising humans for some toxicological endpoints, e.g., increases in
bronchoalveolar lavage (BAL) protein or neutrophil (PMN) levels (indicators of O3-induced
lung inflammation responses).
• Thus, a number of variables seem to affect O3 uptake, notably including route of breathing,
breathing pattern, gender, copollutants, and certain pre-exposure conditions. These
differences are important in order to interrelate experimentally-demonstrated
pathophysiological effects and epidemiologically-observed associations between ambient O3
concentrations and health risks among human population groups.
2. Health Effects of Short-term Exposures to Ozone
The 1996 O3 AQCD assessed a substantial body of evidence from toxicologic, human
clinical, and epidemiologic studies. That AQCD concluded that short-term ambient O3 exposure
resulted in various respiratory health effects, including lung function decrements and increased
respiratory symptoms in both healthy and asthmatic individuals exposed during moderate to
E-ll
-------
heavy exercise to O3 concentrations ranging down to the lowest levels (0.12 ppm for 1 h;
0.08 ppm for 6.6 to 8 h) tested in the available controlled human exposure studies. Such
experimentally demonstrated effects were consistent with and lent plausibility to epidemiologic
observations highlighted in the 1996 AQCD of increases in daily hospital admissions and ED
visits for respiratory causes. Epidemiologic evidence also provided suggestive evidence for an
association between short-term O3 exposure and mortality. However, there was essentially no
evidence available in the 1996 O3 AQCD regarding potential cardiovascular effects of short-term
O3 exposure. The newly-available evidence assessed in this revised O3 AQCD notably enhances
our understanding of short-term O3 exposure effects, as summarized below, first in relation to
respiratory morbidity endpoints and then cardiovascular effects and, lastly, mortality.
A. Respiratory Morbidity
Lung Function:
• Controlled exposure studies clearly demonstrate acute reversible decrements in lung function
in healthy adults exposed to >0.08 ppm O3 when minute ventilation and/or duration of
exposure are increased sufficiently. On average, spirometric responses to O3 exposure
appear to decline with increasing age starting at approximately 18 to 20 years of age.
• There is considerable variability in responses between similarly exposed individuals, such
that some may experience distinctly larger effects even when small group mean responses
are observed. For example, healthy adults exposed to 0.08 ppm O3 for 6.6 h with moderate
exercise exhibited a group mean O3-induced decrement in FEVj of about 6%, but a
decrement of >10% was seen in 23% of these individuals. Also, exposure to 0.06 ppm O3
caused >10% lung function decline in a small percentage (7%) of the subjects.
• Summer camp field studies conducted in southern Ontario, Canada, in the northeastern U.S.,
and in southern California have also reported lung function responses in pre-adolescent
children associated with ambient O3 levels.
• Repeated acute (1- to 6-h) O3 exposures at 0.12 to 0.45 ppm over several days in controlled
exposure studies typically find that FEVj response to O3 is enhanced on the second of several
E-12
-------
days of exposure, but spirometric responses become attenuated on subsequent days with
these repeated exposures. However, this tolerance is lost after about a week without
exposure.
• Animal toxicologic studies also provide extensive evidence that acute O3 exposures alter
breathing patterns so as to cause rapid shallow breathing (i.e., increased frequency and
decreased tidal volume), an effect which attenuates after several days of exposure.
• Results from controlled human exposure studies and animal toxicologic studies provide clear
evidence of causality for the associations observed between acute (<24 h) O3 exposure and
relatively small, but statistically significant declines in lung function observed in numerous
recent epidemiologic studies. Declines in lung function are particularly noted in children,
asthmatics, and adults who work or exercise outdoors.
Respiratory Symptoms:
• Young healthy adult subjects exposed in clinical studies to O3 concentrations >0.08 ppm for
6 to 8 h during moderate exercise exhibit symptoms of cough and pain on deep inspiration.
An increase in the incidence of cough has been found in clinical studies as low as 0.12 ppm
in healthy adults during 1 to 3 h with very heavy exercise and other respiratory symptoms,
such as pain on deep inspiration and shortness of breath, have been observed at 0.16 ppm to
0.18 ppm with heavy and very heavy exercise. These O3-induced respiratory symptoms
gradually decrease in adults with increasing age. With repeated O3 exposures over several
days, respiratory symptoms become attenuated, but this tolerance is lost after about a week
without exposure.
• The epidemiologic evidence shows significant associations between acute exposure to
ambient O3 and increases in a wide variety of respiratory symptoms (e.g., cough, wheeze,
production of phlegm, and shortness of breath) in asthmatic children. Epidemiologic studies
also indicate that acute O3 exposure is likely associated with increased asthma medication
use in asthmatic children.
E-13
-------
• On the other hand, an effect of acute O3 exposure on respiratory symptoms in healthy
children is not as clearly indicated by epidemiology studies, consistent with diminished
symptom responses seen in healthy children in human clinical studies.
Airway Inflammation:
• Inflammatory responses have been observed subsequent to 6.6 h O3 exposures to the lowest
tested level of 0.08 ppm in healthy human adults. Some studies suggest that inflammatory
responses may be detected in some individuals following O3 exposures even in the absence
of O3-induced pulmonary function decrements in those subjects.
• Repeated O3 exposures over several days leads to an attenuation of most inflammatory
markers. However, none of the several markers of lung injury and permeability evaluated
show attenuation, indicating continued lung tissue damage during repeated exposure.
• Animal toxicologic studies provide extensive evidence that acute (1 to 3 h) O3 exposures as
low as 0.1 to 0.5 ppm can cause (1) lung inflammatory responses (typified by increased
reactive oxygen species, inflammatory cytokines, influx of PMNs, and activation of alveolar
macrophages); (2) damage to epithelial airway tissues, (3) increases in permeability of both
lung endothelium and epithelium, and (4) increases in susceptibility to infectious diseases
due to modulation of lung host defenses.
• Consistent with these experimental findings, there is also limited epidemiologic evidence
showing an association between acute ambient O3 exposure and airway inflammation in
children acutely exposed to ambient O3 concentrations (1-h max O3 of approximately
0.1 ppm).
• The extensive human clinical and animal toxicological evidence, together with the limited
available epidemiologic evidence, is clearly indicative of a causal role for O3 in
inflammatory responses in the airways.
E-14
-------
Airway Responsiveness:
• Controlled human exposure studies have found that acute O3 exposure causes an increase in
nonspecific airway responsiveness, as indicated by reductions in concentrations of
methacholine or histamine required to produce a given decrease in FEVj or increase in SRaw.
• Acute (2- or 3-h) O3 exposure at 0.25 or 0.4 ppm of allergic asthmatic subjects, who
characteristically already have somewhat increased airway responsiveness at baseline, was
found to cause further increases in airway responsiveness in response to allergen challenges.
Also, repeated daily exposure to 0.125 ppm O3 for 4 days exacerbated lung function
decrements in response to bronchial allergen challenges among persons with preexisting
allergic airway disease, with or without asthma.
• Ozone-induced exacerbation of airway responsiveness persists longer and attenuates more
slowly than O3-induced pulmonary function decrements and respiratory symptom responses.
Heightened airway responsiveness (reactivity) has also been observed in several laboratory
animal species with acute exposures (1 to 3 h) to 0.5 to 1.0 ppm O3. Ozone increases airway
hyperreactivity to bronchoconstrictive agents (e.g., ovalbumin), and there is a temporal
relationship between inflammatory cell influx and O3-induced increases in airway reactivity.
Several studies of sensitized laboratory animals showing O3-induced increases in airway
hyperreactivity are consistent with O3 exacerbation of airway hyperresponsiveness reported in
atopic humans with asthma.
• Airway responsiveness has not been widely examined in epidemiologic studies. However, the
evidence from human clinical and animal toxicological studies clearly indicate that acute
exposure to O3 can induce airway hyperreactivity, thus likely placing atopic asthmatics at
greater risk for more prolonged bouts of breathing difficulties due to airway constriction in
response to various airborne allergens or other triggering stimuli.
Respiratory Hospital Admissions and Emergency Department Visits:
• Aggregate population time-series studies observed that ambient O3 concentrations are
positively and robustly associated with respiratory-related hospitalizations and asthma ED
E-15
-------
visits during the warm season. These observations are strongly supported by the human
clinical, animal toxicologic, and epidemiologic evidence for lung function decrements,
increased respiratory symptoms, airway inflammation, and airway hyperreactivity.
• Taken together, the overall evidence supports a causal relationship between acute ambient O3
exposures and increased respiratory morbidity outcomes resulting in increased ED visits and
hospitalizations during the warm season.
B. Cardiovascular Morbidity
At the time of the 1996 O3 AQCD, the possibility of O3-induced cardiovascular effects was
a largely unrecognized issue. Newly-available evidence has emerged since then which provides
considerable plausibility for how O3 exposure could exert cardiovascular impacts.
• Direct O3 effects such as O3-induced release from lung epithelial cells of platelet activating
factor (PAF) that may contribute to blood clot formation that would increase the risk of
serious cardiovascular outcomes (e..g, heart attack, stroke, mortality). Also, interactions of
O3 with surfactant components in epithelial lining fluid of the lung results in production of
oxysterols and reactive oxygen species that may exhibit PAF-like activity contributing to
clotting and/or exert cytotoxic effects on lung and heart cells.
• Indirect effects of O3 may involve O3-induced secretions of vasoconstrictive substances
and/or effects on neuronal reflexes that may result in increased arterial blood pressure and/or
altered electrophysiologic control of heart rate or rhythm. Some animal toxicological studies
have shown O3-induced decreases in heart rate, mean arterial pressure, and core temperature.
• Some field/panel studies that examined associations between O3 and various cardiac
physiologic endpoints have yielded limited epidemiologic evidence suggestive of a potential
association between acute O3 exposure and altered HRV, ventricular arrhythmias, and
incidence of MI.
• Highly suggestive evidence for O3-induced cardiovascular effects is provided by a few
population studies of cardiovascular hospital admissions which reported positive O3
E-16
-------
associations during the warm season between ambient O3 concentrations and cardiovascular
hospitalizations. Only one controlled human exposure study that evaluated effects of O3
exposure on cardiovascular health outcomes found no significant O3-induced differences in
ECG, heart rate, or blood pressure in healthy or hypertensive subjects, but did observe an
overall increase in myocardial work and impairment in pulmonary gas exchange.
• Overall, this generally limited body of evidence is highly suggestive that O3 directly and/or
indirectly contributes to cardiovascular-related morbidity, but much remains to be done to
more fully substantiate links between ambient O3 exposure and adverse cardiovascular
outcomes.
C. Mortality
Numerous recent epidemiologic studies conducted in the United States and abroad have
investigated the association between acute exposure to O3 and mortality. Results from several
large U.S. multicity studies as well as several single-city studies indicate a positive association
between increases in ambient O3 levels and excess risk of all-cause (nonaccidental) daily
mortality.
• Consistent with observed O3-related increases in respiratory- and cardiovascular-related
morbidity, several newer multicity studies, single-city studies, and several meta-analyses of
these studies have provided relatively strong epidemiologic evidence for associations
between short-term O3 exposure and all-cause mortality, even after adjustment for the
influence of season and PM.
• Determining cause-specific mortality is more difficult due to reduced statistical power by
which to examine cause-specific associations and the lack of clarifying information on
contributing causes of death. That is, attribution to one or the other of the more specific
cardiopulmonary causes may underplay contributions of chronic cardiovascular disease to
"respiratory" deaths (e.g., a heart attack victim succumbing to acute pneumonia) or vice
versa.
E-17
-------
• Consistently positive associations have been reported for O3-related cardiovascular mortality
across approximately 30 studies, with two well-conducted multicity studies in the United
States and Europe yielding small, but statistically significant positive associations.
• Both animal and human studies provide evidence suggestive of plausible pathways by which
risk of respiratory or cardiovascular morbidity and mortality could be increased by ambient
O3 either acting alone or in combination with copollutants in ambient air mixes.
• This overall body of evidence is highly suggestive that O3 directly or indirectly contributes to
non-accidental and cardiopulmonary-related mortality, but additional research is needed to
more fully establish underlying mechanisms by which such effects occur.
3. Health Effects of Long-term Exposures to Ozone
In the 1996 O3 AQCD, the available epidemiologic data provided only suggestive evidence
that respiratory health effects were associated with chronic O3 exposure. Animal toxicologic
studies indicated that chronic O3 exposure caused structural changes in the respiratory tract, and
simulated seasonal exposure studies in animals suggested that such exposures might have
cumulative impacts. As summarized below, recent studies are generally consistent with the
conclusions drawn in the previous 1996 AQCD.
A. Respiratory Morbidity
Lung Function:
• Recent epidemiologic studies observed that reduced lung function growth in children was
associated with seasonal exposure to O3; however, cohort studies investigating the effect of
annual or multiyear O3 exposure observed little clear evidence for impacts of longer-term,
relatively low-level O3 exposure on lung function development in children.
• The epidemiologic data, collectively, indicate that the current evidence is suggestive but
inconclusive for respiratory health effects from long-term O3 exposure.
E-18
-------
Morphological Changes:
• Animal toxicologic studies continue to show chronic O3-induced structural alterations in
several regions of the respiratory tract including the centracinar region. Morphologic
evidence from some recent studies using exposure regimens that mimic seasonal exposure
patterns report increased lung injury compared to conventional chronic stable exposures.
• Infant rhesus monkeys repeatedly exposed to 0.5 ppm 8h/day O3 for 11 episodes exhibited:
(1) remodeling of the distal airways; (2) abnormalities in tracheal basement membrane; (3)
eosinophil accumulation in conducting airways; and (4) decrements in airway innervation.
Long-term O3 exposure of rats to 0.5 or 1.0 ppm for 20 months resulted in upper respiratory
tract mucus metaplasia and hyperplasia in the nasal epithelium (0.25 or 0.5 ppm, 8h/day,
7days/wk for 13 weeks).
• The persistent nature of these cytological changes raise the possibility of long-lasting
alterations in human airways in response to chronic O3 exposure, but it is highly uncertain as
to what long-term patterns of exposure or O3 concentrations in humans may be requisite to
produce analogous morphological changes. Nor is it now possible to characterize the
possible magnitude or severity of any such effects occurring in humans in response to
ambient O3 exposures at levels observed in the United States.
Incidence of Lung Cancer:
• The weight of evidence from recent animal toxicological studies and a very limited number
of epidemiologic studies do not support ambient O3 as a pulmonary carcinogen.
B. Mortality
• Results from the few available epidemiologic studies are inconsistent regarding the
association between long-term exposure to O3 and mortality. There is little evidence to
suggest a causal relationship between chronic O3 exposure and increased risk for mortality in
humans.
E-19
-------
4. Health Effects of Ozone-Containing Pollutant Mixtures
The potential interaction of pollutant mixtures with O3 is poorly understood and the animal
studies reviewed in the 1996 O3 AQCD reported additive, synergistic or antagonistic effects
depending on the exposure regimen and the endpoint studied. A few new controlled human
exposure and animal toxicology studies reviewed in Chapters 4, 5, and 6 investigated health
effects associated with O3-containing pollutant mixtures of near ambient levels. As noted below,
recent studies, although generally consistent with conclusions drawn in the 1996 O3 AQCD, have
added some new information, particularly with regard to interactions between O3 and PM.
• Controlled human exposure studies indicate that continuous exposure of healthy human
adults to SO2 or NO2 increases bolus dose O3 absorption, suggesting that co-exposure to
other gaseous pollutants in the ambient air may enhance O3 absorption.
• Other controlled human exposure studies that evaluated response to allergens in asthmatics
(allergic and dust-mite sensitive) suggest that O3 enhances response to allergen challenge.
Consistent with these findings, animal toxicology studies also reported enhanced response to
allergen on exposure to O3.
• A few other animal toxicology studies that exclusively investigated the co-exposure of PM
and O3 reported increased response (lung tissue injury, inflammatory and phagocytosis) to
the mixture of PM + O3 compared to either PM or O3 alone.
• Recent investigations on the copollutant interactions using simulated urban photochemical
oxidant mixes suggest the need for similar studies in understanding the biological basis for
air pollutant mixture effects observed in epidemiologic studies.
5. Susceptibility or Vulnerability to Effects Associated with Exposure to Ozone
Various factors have been shown to influence individuals' responses to environmental air
pollutants. Factors that increase susceptibility to O3-related effects include innate factors, such
as genetic predisposition or developmental effects, or disease status. Other factors can lead to
enhanced vulnerability to O3-related effects, such as heightened exposures or activity patterns.
In the 1996 O3 AQCD, available evidence suggested that children, asthmatics, and outdoor
E-20
-------
workers were populations that may be more susceptible or vulnerable to effects of O3 exposure.
In addition, controlled human exposure studies also demonstrated a large variation in sensitivity
and responsiveness to O3 in studies of healthy subjects, but the specific factors that contributed
to this intersubject variability were yet to be identified. Recent studies have built upon the
evidence available in the previous review. Factors related to susceptibility or vulnerability to O3
exposure-related effects are briefly summarized below:
People with Preexisting Pulmonary Diseases:
• Ozone-induced differential responses in lung function and AHR in people with allergic
rhinitis suggest that asthmatics have potentially greater responses than healthy people with
exposure to O3. There is a tendency for slightly increased spirometric responses in mild
asthmatics and allergic rhinitics relative to healthy young adults. Spirometric responses in
asthmatics appear to be affected by baseline lung function, i.e., responses increase with
disease severity.
• Repeated O3 exposure over several days has been shown to increase responsiveness to
bronchial allergen challenge in subjects with preexisting allergic airway disease, with or
without asthma. Asthmatics also show a significantly greater neutrophil response (18 h
postexposure) than similarly-exposed healthy individuals.
• Epidemiologic studies have reported associations with a range of respiratory health outcomes
in asthmatics, from decreases in lung function to hospitalization or ED visits for asthma, thus
supporting this population group as being likely to experience increased risk for O3-induced
health effects.
• Controlled human exposure studies have not found evidence of larger spirometric changes in
people with COPD relative to healthy subjects, this may be due to the fact that most people
with COPD are older adults who would not be expected to have such changes based on their
age. However, new epidemiologic evidence indicates that people with COPD may be more
likely to experience other effects, including emergency room visits, hospital admissions, or
premature mortality.
E-21
-------
Age-related:
• Controlled human exposure studies have shown that lung function responses to O3 varies
with age, with responsiveness generally diminishing after about 18 to 20 years of age.
Children and older adults thus have lesser respiratory symptoms with O3 exposure than
young healthy adults. Potentially increased O3 doses can be received by individuals
experiencing less severe respiratory symptoms.
• Evidence from newer epidemiologic studies supports the 1996 O3 AQCD conclusions that
children are more likely at increased risk for O3-induced health effects. Notably,
epidemiologic studies have indicated adverse respiratory health outcomes associated with O3
exposure in children. In addition, recently published epidemiologic studies also suggest that
older adults (aged >65 years) appear to be at excess risk of O3-related mortality or
hospitalization.
Heightened vulnerability due to greater exposures:
• Epidemiologic studies have provided some evidence to indicate that outdoor workers are
more vulnerable to O3-related effects, which is likely related to their increased exposure to
ambient air pollution.
• Controlled human exposure studies clearly established differential biological response to O3
based on physical activity (exertion). Epidemiologic studies also suggest that exercising
(moderate to high physical exertion) children and adolescents appear to demonstrate
increased responsiveness to ambient concentrations of O3 and may be more likely to
experience O3-induced health effects. Animal studies show a similar impact of exercise on
responsiveness to O3.
Genetic susceptibility:
• Animal toxicologic studies provide supportive evidence to the observations of innate
susceptibility. Various strains of mice and rats have demonstrated the importance, in
general, of genetic background in O3 susceptibility. Moreover, genetic and molecular
E-22
-------
characterization studies in laboratory animals identified genetic loci responsible for both
sensitivity and resistance.
New human clinical and epidemiologic studies also have shown that genetic polymorphisms
for antioxidant enzymes and inflammatory genes (GSTM1, NQO1, and Tnf-a) may modulate
the effect of O3 exposure on pulmonary function and airway inflammation.
E.5 VEGETATION AND ECOLOGICAL EFFECTS
Data published since 1996, as assessed in Chapter 9 and associated annex materials,
continues to support and strengthen the conclusions of previous O3 AQCDs. The main
findings/conclusions derived from the current Chapter 9 assessment of O3 ecological effects are
as follows.
General
• The ecological effects of O3 appear to be widespread across the United States based on recent
biomonitoring studies using clover and other species grown in plots across the United States,
as well as regional forest health visible injury surveys.
• Some plant community compositions may be shifting based on recent studies of competition
among plants in managed pasture lands, as well in natural unmanaged lands where increased
O3 effects on sensitive species in the community can reduce their presence in the community.
• Research to date has focused at the species level, with very few studies at the ecosystem
level. The lack of data at this organizational level hampers the assessment of O3 risk to
ecosystem services, such as water quality and quantity, that contribute to human well-being.
Methodologies
• New methodologies coming into use since 1996 have not fundamentally altered our
understanding of O3 effects on plants or the conclusions of the 1996 O3 AQCD. Since 1996,
there has been a shift from chamber-based studies to the field-based approach, including plot
E-23
-------
and regional visible injury surveys and the use of the non-chambered free air CO2 exposure
(FACE) systems. The FACE system results support earlier observations of foliar injury and
reduced volume growth in aspen and indicate reduced yield in soybean cultivars similar to
earlier studies in open-top chamber systems (OTC).
• The use of biomonitoring since 1996 has advanced identification and symptom verification
of sensitive species and has been a useful tool for indicating the extent of O3 effects across
most of the eastern and southeastern United States and many parts of the West.
• The development and improvement of stomatal models for predicting O3 uptake in Europe
have fostered more universal measures of exposure response. These simulation tools may
provide a better means to relate ambient exposure to plant response in the future but
currently are insufficient for use across broad geographical areas of the United States.
• Since 1996, the use of passive samplers for monitoring O3 in rural and remote areas has
expanded, offering a potential for improved exposure data in areas not actively monitored.
The testing and development of these samplers will ultimately provide a strategy to expand
air quality monitoring into areas for which exposure characterization is currently done by
geospatial extrapolation techniques such as Kriging.
Mode of Action
• There are several steps in the process of O3 uptake and toxicity that are better understood
now than in 1996, based on new information gained in part by use of improved molecular
tools for following rapid changes that occur within the leaf. These advancements are
important for refining hypotheses on O3 uptake and improving understanding of exposure-
response relationships.
• Ozone entrance into the leaf through the stomata remains the critical step in O3 sensitivity.
Although the initial reactions within the leaf are still unclear, the involvement of H2O2 is
clearly indicated. The initial sites of membrane reactions seem to involve transport
E-24
-------
properties. The primary set of metabolic reactions that O3 triggers currently includes those
typical of "wounding" responses generated by leaf cutting or by insect attack.
• The alteration of normal metabolism due to wounding spreads outside the cytoplasm. One of
the secondary reactions is linked to a senescence response. The loss of photo synthetic
capacity is linked to lower productivity (although not fully elucidated) and to problems with
efficient translocation of carbon.
• Chronic O3 effects are linked to the senescence process or some physiological process
closely linked to senescence, e.g., translocation, re-absorption, allocation of nutrients and
carbon.
Modification of Growth Response
• Many biotic and abiotic factors, including insects, pathogens, root microbes and fungi,
temperature, water and nutrient availability, and other air pollutants, as well as elevated CO2,
influence or alter the plant's response to O3. A few studies published since 1996 have
improved our understanding of the role of these interactions in modifying O3-induced plant
responses.
• Biotic Interactions: Recent studies have supported earlier conclusions that O3 often increases
the likelihood and success of insect attacks, at least by chewing insects. Less is known
regarding sucking insects (e.g., aphids). It seems that some insect problems could be
exacerbated by increased O3 exposure, but predicting any particular O3-insect interaction is
not possible at this time.
• Biotic Interactions: More information is available regarding disease interactions.
Ozone exposure generally increases plant diseases associated with facultative necrotrophic
plant pathogens. Pathogens that benefit from damage to cells are enhanced by O3 stress to
their hosts, whereas pathogens that require healthy hosts are depressed by O3 stress.
E-25
-------
• Biotic Interactions: A few new studies have demonstrated O3 impacts on intraspecific plant
competition. In grass-legume pastures, the legume component is more O3-sensitive and is
reduced over time. Similarly, grass competition on pine seedlings can enhance O3 effects on
the seedlings, possibly through the grass's ability to outcompete the seedlings for water.
• Abiotic Interactions: New information on the role of abiotic or physical factors interacting
with O3 stress support 1996 O3 AQCD conclusions. Some studies have shown an increasing
effect of O3 with increasing temperature, but others have shown little effect of temperature.
Temperature is an important variable affecting plant O3 response in the presence of elevated
CO2 levels associated with climate change. It also appears that low temperatures are
important, in that O3 exposure sensitizes plants to low temperature stress.
• Abiotic Interactions: New information on the role of drought and water availability
published since 1996 confirms earlier conclusions regarding the increased effects of O3
where readily available soil moisture results in increased needle/leaf conductivity and thus
increased O3 uptake. Additional studies demonstrated again the partial "protection" against
adverse effects of O3 by drought. There was also evidence that O3 predisposes plants to
drought stress. The net results of these interactions are negative, at least in the short term,
although longer lived species like trees could benefit from increased water use efficiency.
Effects-Based Exposure Indices
• Exposure indices are metrics relating plant response (i.e., growth or yield) to monitored
ambient O3 concentrations over time to provide a consistent metric for reviewing and
comparing exposure-response effects obtained from various studies. Such metrics may also
provide a basis for developing air quality standards that are protective of ecological
resources. The 1996 O3 AQCD focused on research where a large number of indices were
developed that included various functional and statistical summaries of ambient hourly
concentrations over designated time periods. The development of those indices focused on
considering and including some, but not all, the factors that affect O3 uptake and expression
of effects.
E-26
-------
• Conclusions from the 1996 O3 AQCD regarding an ambient-exposure based index are still
valid. No information since 1996 significantly alters the basic conclusions, and most studies
in this interim have further supported them. The key conclusions are as follows:
- Ozone effects in plants are cumulative;
- Higher O3 concentrations appear to be more important than low concentrations in
eliciting a response;
- Plant sensitivity to O3 varies with time of day and plant development stage; and
- Exposure indices that accumulate the O3 hourly concentrations and preferentially weight
the higher concentrations have a better statistical fit to growth/yield response than do
mean or peak indices.
• Based on the current state of knowledge, exposure indices that differentially weight the
higher hourly average O3 concentrations but include the mid-level values represent the best
approach for relating vegetation effects to O3 exposure in the United States. A large database
for crops and tree seedlings exists and has been used for establishing exposure-response
relationships and predicting effects for a range of exposure concentrations. In 1996, EPA
considered three specific concentration-weighted indices for use as air quality indicators: the
cutoff concentration-weighted SUM06, the AOT60, and the sigmoid-weighted W126. All
three performed equally well based on goodness-of-fit tests. Since 1996, there have been no
published experimental studies that would alter the consideration of these concentration-
weighted cumulative indices.
• Studies available since 1996 strengthen earlier conclusions on the role of exposure
components such as duration, concentration, and temporal patterns in determining plant
growth response to O3 exposure. New studies since 1996 have shown experimentally the
disconnection of peak events and maximal stomatal conductance at a variety of sites. The
identification of sensitivity linked to time of day (i.e., period of maximum conductance) was
reported in the 1996 O3 AQCD. The new studies will offer future avenues for building this
temporal component into an exposure index. Similarly, recent reviews of the plant literature
have reported a large number of species with nighttime conductance capable of O3 uptake.
The designated time interval for cumulating exposure (i.e., 12 to 24 h) needs reconsideration.
E-27
-------
• Recent research in Europe has focused on a flux-based approach to improve upon the
ambient air concentration-based (i.e., exposure indices) approach to assess risk from O3
across different climate regions. However, such approaches need further development to
incorporate the necessary complexity across space and time to be non-site and non-species
specific. Also, at this time, the database is inadequate for linking O3 flux to growth
responses.
Ozone Exposure-Plant Response Relationships
• Data published since 1996 continues to support and strengthen the conclusions of previous
O3 AQCDs that there is strong evidence that current ambient O3 concentrations cause
(1) decreased growth and biomass accumulation in annual, perennial and woody plants,
including agronomic crops, annuals, shrubs, grasses, and trees; (2) decreased yield and/or
nutritive quality in a large number of agronomic and forage crops; and (3) impaired aesthetic
quality of many native plants and trees by increased foliar injury.
• Since 1996, the published studies have supported earlier conclusions on reduced growth and
yield in a number of crops and trees, and have used multiple approaches, including regional
visible injury surveys, measured growth responses across ambient exposure gradients,
empirical exposure studies in chambered and non-chambered systems, as well as process
model simulations.
• Studies of growth response using open-top chambers have provided useful data for assessing
O3 impact on common and economically valuable species, and developing functional
growth-response models that enable the prediction of O3 impact over a wide range of
ambient air exposures. The studies were designed to maximize statistical robustness by
replicating a number of treatments and, at the same time, considering issues of extrapolation
by conducting the studies across a wide range of crop-growing regions and forested sites in
the United States. Such designs allowed the studies to account for climate and growing
conditions, as well as regional crop growing practices.
E-28
-------
• Recent exposure studies in Illinois using the non-chambered FACE system with soybean
cultivars reported reductions in yield of soybeans in two year-long studies similar to the
reductions found in multiple soybean studies conducted in the 1980's using open-top
chamber systems. Multiple-year exposure studies using the FACE system (a) found foliar
injury and reduced volume growth in aspen and maple similar to results reported from earlier
open-top chamber studies and (b) highlight the importance of multiyear studies with
longer-growing species.
• Since 1996, additional studies have supported 1996 O3 AQCD conclusions that deciduous
trees are generally less O3 sensitive than most annual species or crop plants, with the
exception of a few very sensitive genera (e.g., sensitive clones or genotypes oiPopulus) and
sensitive species (e.g., black cherry). Coniferous species have a wide range of O3
sensitivities but, in general, are less sensitive than deciduous species. Among conifers, the
slower-growing species are less sensitive than faster-growing species. Data from a few
European studies support these conclusions.
• For all types of perennial vegetation, cumulative effects over more than one growing season
may be important; studies of one or a few seasons may under- or overestimate O3 impacts on
these species. Results from multiyear studies sometimes found a pattern of increased effects
in subsequent years, whereas other studies reported growth decreases due to O3 that become
less significant or disappear over time. It is difficult to conduct empirical experiments with
long-lived trees, because even multiyear exposures only account for a small fraction of the
tree's lifetime. Model simulations of growth have been a tractable approach to account for
time and changing climate in assessing the impact on long-lived trees.
Ecosystem Effects
• There is strong evidence that O3, in locations where ambient levels are relatively high, is an
important stressor of ecosystems, with documented impacts on the biotic condition,
ecological processes, and chemical/physical nature of natural ecosystems. Experimentally
documented effects on individual keystone species and their associated microflora and fauna
may cascade through the ecosystem to the landscape level, but this has not been quantified.
E-29
-------
Systematic injury surveys (e.g., USDA Forest Health Network; Europe's TCP Forests) show
that foliar injury and crown/canopy deformations occur in O3-sensitive species in many
regions of the United States and Europe. However, the lack of general correspondence
between foliar symptoms and growth effects means that other methods must be used to
estimate regional effects of O3 on tree growth. Regional studies of radial growth in mature
trees, combined with data from many controlled studies with seedlings and a few studies
with mature trees, suggest that ambient O3 may be reducing the growth of mature trees in
some U.S. locations.
The use of physiological-based process models to simulate tree growth, combined with
stand-level models predicting forest composition and productivity, is an approach being used
recently in assessing O3 impacts on forests. These tools suggest that modest O3 effects on
growth may accumulate over time and interact with effects of other naturally occurring
stresses (e.g., drought, nutrient availability). For mixed-species stands, the models predict
that overall stand growth may not be affected, but competitive interactions among species
may change the composition due to growth reductions in sensitive species.
The knowledge base for examining the range of ecological effects of O3 on natural
ecosystems is growing, but significant uncertainties remain regarding O3 effects at the
ecosystem level. A number of significant areas for investigation that would improve our
ability to assess O3 effects on ecosystems and the services they provide for human well-being
have been outlined and discussed (see Annex AX-9).
E.6 TROPOSPHERIC OZONE EFFECTS ON UV-B FLUX AND ITS
ROLE IN CLIMATE CHANGE
Molecular properties specific to O3 include a capacity for absorbing incoming ultraviolet
(UV) and infrared (IR) radiation, and both incoming solar and outgoing terrestrial IR radiation.
Consequently, O3 plays an essential role in shielding the earth's surface from harmful levels of
UV-B radiation, by way of the stratospheric O3 layer. Its effectiveness as a screen for the
residual UV-B flux that penetrates the stratosphere and passes into the troposphere and its role in
E-30
-------
reducing UV-induced human health effects are addressed in Chapter 10. The radiation-
absorbing properties of O3 also make it a greenhouse gas (GHG) having global and regional
consequences for climate, as also addressed in Chapter 10. Important conclusions from
Chapter 10 are summarized below.
• The distribution of (X within the atmosphere. Ozone is distributed very unevenly within the
atmosphere, with -90% of the total atmospheric burden present in the stratosphere. The
remaining -10% is distributed within the troposphere, with higher relative concentrations
near the source of its precursors at the surface. Concentrations of O3 at the mid- and upper-
troposphere vary, depending upon meteorological conditions.
• Multiple factors govern the flux of UV-B radiation at the Earth's surface. Latitude and
altitude are the two most important factors that define the residual UV-B flux at the surface.
Natural variation in the total column density of stratospheric O3 is also an important factor.
All of these factors are followed in importance by tropospheric clouds, particulate matter
(PM) and O3. The effect of natural stratospheric variation, clouds, PM and tropospheric O3
on UV fluxes within the troposphere and at the surface are each very difficult to predict.
• A UV-B "climatology" is needed to predict human exposure levels. A UV-B climatology,
representing patterns and trends in UV-B flux at the Earth's surface, must be based on
extended in situ observations in order to adequately capture natural variability and the effects
of human activities on atmospheric UV-B absorbers. At present, the body of UV-B
measurements cannot support the development of a climatology.
• Human exposure to UV-B radiation. Quantitative evaluation of human exposure to UV-B
radiation is necessary to perform health risk assessment for UV-B-related health effects.
Individuals who participate in outdoor sports and activities, work outdoors, live in
geographic areas with higher solar flux, and/or engage in high-risk behavior (e.g., extended
sun bathing) can reasonably be projected to be at increased risk for higher UV radiation
exposures. However, little is known about the impact of variability in these factors on
individual exposure to UV radiation.
E-31
-------
• Human health effects of UV-B radiation. Exposure to UV-B radiation is associated with
increased risk of erythema, nonmelanoma and melanoma skin cancers, ocular damage, and
immune system suppression. Some studies have attempted to estimate the potential effects
of changes in surface-level UV flux resulting from stratospheric O3 depletion on these health
outcomes; however, the numerous simplifying assumptions made in the assessments limit the
usefulness of the risk estimates. The effect of changes in surface-level O3 concentrations on
UV-induced health outcomes cannot yet be critically assessed within reasonable uncertainty.
• Vitamin D-related health benefits of UV-B radiation. A potential health benefit of increased
UV-B exposure relates to the production of vitamin D in humans. Several studies have
found that UV-B radiation, by increasing vitamin D production, is associated with reduced
risks of various cancers. However, as with other impacts of UV-B on human health, this
beneficial effect of UV-B has not been studied in sufficient detail to allow for a credible
health benefits assessment. No study has been done of the decreased risk of cancer resulting
from increased UV radiation attributable to decreased tropospheric O3 levels, but the change
in risk is expected to be unappreciable.
• Ozone is a potent GHG. Ozone traps incoming solar radiation at both ends of the spectrum,
as well as shortwave radiation that is scattered from high-albedo portions of the Earth's
surface. Outgoing terrestrial IR is absorbed by O3 within the range where water vapor does
not absorb, so that natural variability in humidity does not alter its radiative impact. These
effects directly force climate. By participating in the oxidative chemistry of the atmosphere,
O3 can indirectly and negatively force climate by the removal of other greenhouse gases.
• Multiple factors influence the forcing effect of tropospheric O3. Estimates of present-day
forcing by O3 depend upon currently available information on pre-industrial and current O3
concentrations. Both are limited and, therefore, very uncertain. Other factors, including the
albedo of underlying surface, altitude and co-occurrence of PM can also complicate the
calculation of globally-averaged forcing.
E-32
-------
• Globally-averaged direct forcing by (X. On the basis of the best available information, a
2001 Intergovernmental Panel on Climate Change (IPCC) report offered an estimated value
of 0.35 ± 0.15 WnT2 for the annual, globally-averaged direct forcing by tropospheric O3.
Another recent estimate places this value at 0.5 ± 0.2 WnT2.
• Projections of forcing by O3 into the future. A CTM-climate modeling intercomparison
study carried out as part of the third assessment by the IPCC yielded an estimated 0.4 to
0.78 WnT2 forcing by O3 by the year 2100. The authors of this study concluded that O3 can
be expected to be an important contributor to climate forcing into the future.
• Climate forcing by O? at the regional scale may be its most import impact on climate.
Satellites have detected high O3 concentrations localized at the regional scale that are
associated with large urban centers and extensive biomass burning. Climate forcing by these
high, regional-scale O3 concentrations have been estimated to be on the order of 1 WnT2 (a
substantial fraction of the direct, globally-averaged forcing due to well-mixed GHGs,
including CO2). The impact of climate forcing at this level depends upon the particular
characteristics of the region in which it occurs. At present, regional-scale modeling studies
are not available that provide estimates of these effects.
E.7 MATERIALS DAMAGE
The Chapter 11 discussion of O3 effects on man-made materials mainly summarizes key
information from the 1996 O3 AQCD, given that little new pertinent research information on O3-
related materials damage has been published since then. Key points include the following:
• Ozone and other photochemical oxidants react with many economically important man-made
materials, decreasing their useful life and aesthetic appearance. Materials damaged by O3
include elastomers; textiles and fibers; dyes, pigments, and inks; and paints and other surface
coatings.
E-33
-------
• Elastomeric compounds (natural rubber and synthetic polymers and copolymers of
butadiene, isoprene, and styrene) are highly susceptible, to even low O3 concentrations.
These compounds are damaged by O3 breaking molecular chains at the carbon-carbon double
bond and by adding a chain of three oxygen atoms directly across the double bond. This
structure change promotes characteristic cracking of stressed/stretched rubber called
"weathering." Tensile strain produces cracks on the surface of the rubber that increase in
size and number with increased stress/stretching. The rate of crack growth is dependent on
degree of stress, type of rubber compound, O3 concentration, duration of exposure, O3
velocity, and temperature. After initial cracking, further O3 penetration results in additional
cracking and, eventually, mechanical weakening.
• Ozone can damage textiles and fabrics by mechanisms similar to those associated with
elastomers. Generally, synthetic fibers are less affected by O3 than natural fibers. Overall,
O3 contribution to degradation of textiles and fabrics is not considered significant.
• Ozone fading of textile dyes is a diffusion-controlled process, with the rate of fading being
controlled by diffusion of the dye to the fiber surface. Many textile dyes react with O3. The
rate and severity of the O3 attack is influenced by the chemical nature of the textile fiber.
• Paints applied to exterior surfaces of buildings and other structures (e.g., bridges), as well as
several artists' pigments, are also sensitive to fading and oxidation by O3 at concentrations
found in U.S. urban areas.
E-34
-------
1. INTRODUCTION
This is an update revision of the document, "Air Quality Criteria for Ozone and Related
Photochemical Oxidants" published by the U.S. Environmental Protection Agency (EPA) in
1996 (U.S. Environmental Protection Agency, 1996a). That 1996 Ozone Air Quality Criteria
Document (O3 AQCD) provided scientific bases for Congressionally-mandated periodic review
by the EPA of the Ozone National Ambient Air Quality Standards (O3 NAAQS), which led to
promulgation of new O3 NAAQS by EPA in 1997 (Federal Register, 1997).
The present document critically assesses the latest scientific information relative to
characterizing health and welfare effects associated with the presence of various concentrations
of O3 and related oxidants in ambient air. It builds upon the previous 1996 EPA O3 AQCD,
by focusing on evaluation and integration of scientific information relevant to O3 NAAQS
criteria development that has become available since that covered by the 1996 criteria review;
and it will provide scientific bases for the current periodic review of the O3 NAAQS.
This introductory chapter of the revised O3 AQCD presents: (a) background information
on legislative requirements, the criteria and NAAQS review process, and the history of O3
NAAQS reviews (including a chronology of changes in key elements of the O3 standards);
(b) an overview of the current O3 criteria review process and associated key milestones; and
(c) an orientation to the general organizational structure and content of the document.
1.1 LEGAL AND HISTORICAL BACKGROUND
1.1.1 Legislative Requirements
Two sections of the U.S. Clean Air Act (CAA) govern establishment, review, and revision
of National Ambient Air Quality Standards (NAAQS). Section 108 of the CAA (42 U.S.C.
7408) directs the Administrator of the U.S. Environmental Protection Agency (EPA) to identify
ambient air pollutants that may be reasonably anticipated to endanger public health or welfare
and to issue air quality criteria for them. The air quality criteria are to reflect the latest scientific
information useful in indicating the kind and extent of all identifiable effects on public health or
welfare that may be expected from the presence of a given pollutant in ambient air.
1-1
-------
Section 109(a) of the CAA (42 U.S.C. 7409) directs the Administrator of EPA to propose
and promulgate primary and secondary NAAQS for pollutants identified under Section 108.
Section 109(b)(l) defines a primary standard as one that, in the judgment of the Administrator, is
requisite to protect the public health (see inset below) based on the criteria and allowing for an
adequate margin of safety. The secondary standard, as defined in Section 109(b)(2), must
specify a level of air quality that, in the judgment of the Administrator, is requisite to protect the
public welfare (see inset below) from any known or anticipated adverse effects associated with
the presence of the pollutant in ambient air, based on the criteria.
EXAMPLES OF
PUBLIC HEALTH EFFECTS
i D Effects on the health of the general population,
or identifiable groups within the population,
who are exposed to pollutants in ambient air
i D Effects on mortality
i D Effects on morbidity
i D Effects on other health conditions including
indicators of:
• pre-morbid processes,
• risk factors, and
• disease
EXAMPLES OF
PUBLIC WELFARE EFFECTS
i D Effects on personal comfort and well-being
i D Effects on economic values
i D Deterioration of property
i D Hazards to transportation
i D Effects on the environment, including:
animals
climate
crops
materials
soils
1 vegetation
1 visibility
1 water
1 weather
1 wildlife
Section 109(d) of the CAA (42 U.S.C. 7409) requires periodic review and, if appropriate,
revision of existing criteria and standards. If, in the Administrator's judgment, the Agency's
review and revision of criteria make appropriate the proposal of new or revised standards, such
standards are to be revised and promulgated in accordance with CAA Section 109(b). Or, the
Administrator may find that revision of the standards is not appropriate and conclude the review
by leaving existing standards unchanged. Section 109(d)(2) of the CAA also requires that an
independent scientific review committee be established to advise the EPA Administrator on
NAAQS matters, including the scientific soundness of criteria (scientific bases) supporting
NAAQS decisions. This role is fulfilled by the Clean Air Scientific Advisory Committee
(CASAC), which is administratively supported by EPA's Science Advisory Board (SAB).
1-2
-------
1.1.2 Criteria and NAAQS Review Process
Periodic reviews by EPA of criteria and NAAQS for a given criteria air pollutant progress
through a number of steps, beginning with preparation of an air quality criteria document
(AQCD) by EPA's National Center for Environmental Assessment Division in Research
Triangle Park, NC (NCEA-RTP). The AQCD provides a critical assessment of the latest
available scientific information upon which the NAAQS are to be based. Drawing upon the
AQCD, staff of EPA's Office of Air Quality Planning and Standards (OAQPS) prepare a Staff
Paper that evaluates policy implications of the key studies and scientific information contained
in the AQCD and presents EPA staff conclusions and recommendations for standard-setting
options for the EPA Administrator to consider. The Staff Paper is intended to help "bridge the
gap" between the scientific assessment contained in the AQCD and the judgments required of
the Administrator in determining whether it is appropriate to retain or to revise the NAAQS.
Iterative drafts of the AQCD and the Staff Paper (as well as other analyses, such as exposure
and/or risk assessments, supporting the Staff Paper) are made available for public comment and
CAS AC review. Final versions of the AQCD and Staff Paper incorporate changes made in
response to CASAC and public review. Based on the information in these documents, the EPA
Administrator proposes decisions on whether to retain or revise the NAAQS, taking into account
public comments and CASAC advice and recommendations. The Administrator's proposed
decisions are published in the Federal Register, with a preamble that presents the rationale for
the decisions and solicits public comment. After considering comments received on the
proposed decisions, the Administrator then makes final decisions on retaining or revising the
NAAQS, which are promulgated in & Federal Register notice that addresses significant
comments received on the proposal.
NAAQS decisions involve consideration of the four basic elements of a standard:
indicator, averaging time, form, and level. The indicator defines the pollutant to be measured in
the ambient air for the purpose of determining compliance with the standard. The averaging
time defines the time period over which air quality measurements are to be obtained and
averaged, considering evidence of effects associated with various time periods of exposure.
The form of a standard defines the air quality statistic that is to be compared to the level of the
standard (i.e., an ambient concentration of the indicator pollutant) in determining whether an
area attains the standard. The form of the standard specifies the air quality measurements that
1-2
-------
of
measurements
are to be used for compliance purposes (e.g., the 98th percentile of an annual distribution
daily concentrations; the annual arithmetic average), the monitors from which the measur
are to be obtained (e.g., one or more population-oriented monitors in an area), and whether the
statistic is to be averaged across multiple years. These basic elements of a standard are the
primary focus of the staff conclusions and recommendations in the Staff Paper and in the
subsequent rulemaking, building upon the policy-relevant scientific information assessed in the
AQCD and on the policy analyses contained in the Staff Paper. These four elements taken
together determine the degree of public health and welfare protection afforded by the NAAQS.
1.1.3 Regulatory Chronology1
On April 30, 1971, primary and secondary NAAQS for photochemical oxidants were
promulgated by EPA under Section 109 of the CAA (36 FR 8186). These NAAQS were set at
an hourly average of 0.08 ppm total photochemical oxidants, not to be exceeded more than 1 h
per year. On April 20, 1977, the EPA announced (42 FR 20493) the first review and updating of
the 1970 Air Quality Criteria Document for Photochemical Oxidants in accordance with Section
109(d) of the CAA. In preparing that criteria document, EPA made two external review drafts of
the document available for public comment, and these drafts were peer reviewed by the
Subcommittee on Scientific Criteria for Photochemical Oxidants of EPA's Science Advisory
Board (SAB). A final revised AQCD for Ozone and Other Photochemical Oxidants was then
published on June 22, 1978.
Based on the revised 1978 AQCD and taking into account the advice and recommendations
of the SAB Subcommittee and public comments, the EPA announced (44 FR 8202) a final
decision to revise the NAAQS for photochemical oxidants on February 8, 1979. That final
rulemaking revised the primary standard from 0.08 ppm to 0.12 ppm, set the secondary standard
to be the same as the primary standard, changed the chemical designation of the standards from
photochemical oxidants to O3, and revised the definition of the point at which the standard is
attained as indicated in Table 1-1.
'This following text is excerpted and adapted from the "Proposed Decision on the National Ambient
Air Quality Standards for Ozone," 57 FR 35542, 35542-35557 (August, 10, 1992) and the "National Ambient Air
Quality Standards for Ozone; Final Rule," 62 FR 38856, 83356-38896 (July 18, 1997).
1-4
-------
Table 1-1. National Ambient Air Quality Standards (NAAQS) for Ozone
Date of Promulgation Primary and Secondary NAAQS Averaging Time
February 8, 1979 0.12 ppnf (23 5 ug/m3) 1 hb
July 18, 1997 0.08 ppma (157 ug/m3) 8 hc
al ppm = 1962 ug/m3, 1 ug/m3 = 5.097 x 1Q-4 ppm @ 25 °C, 760 mm Hg.
bThe standard is attained when the expected number of days per calendar year with a maximum hourly average
concentration above 235 ug/m3 (0.12 ppm) is equal to or less than one.
"Based on the 3-year average of the annual fourth-highest daily maximum 8-h average concentration measured
at each monitor within an area.
Source: Federal Register (1979, 1997).
On March 17, 1982, in response to requirements of Section 109(d) of the CAA, the EPA
announced (47 FR 11561) that it planned to revise the existing 1978 AQCD for Ozone and Other
Photochemical Oxidants; and, on August 22, 1983, it announced (48 FR 38009) that review of
the primary and secondary NAAQS for O3 had been initiated. The EPA provided a number of
opportunities for expert review and public comment on revised chapters of the AQCD, including
two public peer-review workshops in December 1982 and November 1983. Comments made at
both workshops were considered by EPA in preparing the First External Review Draft that was
made available (49 FR 29845) on July 24, 1984, for public review. On February 13, 1985
(50 FR 6049) and then on April 2, 1986 (51 FR 11339), EPA announced two public CASAC
meetings, which were held on March 4-6, 1985 and April 21-22, 1986, respectively. At these
meetings, the CASAC reviewed external review drafts of the revised AQCD for O3 and Other
Photochemical Oxidants. After these two reviews, CASAC's consensus views were summarized
by the CASAC Chair in an October 1986 letter to the EPA Administrator, which stated that the
document "represents a scientifically balanced and defensible summary of the extensive
scientific literature." Taking into account public and CASAC comments on the two external
review drafts, revisions were made by EPA and the final document was released by EPA in
August 1986.
The first draft of the Staff Paper "Review of the National Ambient Air Quality Standards
for Ozone: Assessment of Scientific and Technical Information " drew upon key findings and
conclusions from the AQCD and was reviewed by CASAC at the April 21-22, 1986 public
1-5
-------
meeting. At that meeting, the CASAC recommended that new information on prolonged O3
exposure effects be considered in a second draft of the Staff Paper. The CASAC reviewed the
resulting second draft and also heard a presentation of new and emerging information on the
health and welfare effects of O3 at a December 14-15, 1987 public review meeting. The CASAC
concluded that sufficient new information existed to recommend incorporation of relevant
new data into a supplement to the 1986 AQCD (O3 Supplement) and into a third draft of the
Staff Paper.
A draft O3 Supplement, entitled "Summary of Selected New Information on Effects of
Ozone on Health and Vegetation: Draft Supplement to Air Quality Criteria for Ozone and Other
Photochemical Oxidants, " and the revised Staff Paper were made available to CASAC and to
the public in November 1988. The O3 Supplement assessed selected literature concerning
exposure- and concentration-response relationships observed for health effects in humans and
experimental animals and for vegetation effects that appeared in papers published or in-press
from 1986 through early 1989. On December 14-15, 1988, the CASAC held a public meeting to
review these documents and then sent the EPA Administrator a letter (dated May 1, 1989),
which stated that the draft O3 Supplement, the 1986 AQCD, and the draft Staff Paper "provide
an adequate scientific basis for the EPA to retain or revise the primary and secondary standards
of ozone." The CASAC concluded (a) that it would be some time before sufficient new
information on the health effects of multihour and chronic exposure to O3 would be published in
scientific journals to receive full peer review and, thus, be suitable for inclusion in a criteria
document and (b) that such information could be considered in the next review of the O3
NAAQS. A final version of the O3 Supplement was published in 1992 (U.S. Environmental
Protection Agency, 1992).
On October 22, 1991, the American Lung Association and other plaintiffs filed suit to
compel the Agency to complete the review of the criteria and standards for O3 in accordance
with the CAA. The U.S. District Court for the Eastern District of New York subsequently issued
an order requiring the EPA to announce its proposed decision on whether to revise the standards
for O3 by August 1, 1992 and to announce its final decision by March 1, 1993.
The proposed decision on O3, which appeared in the Federal Register on August 10, 1992
(57 FR 35542), indicated that revision of the existing 1-h O3 NAAQS was not appropriate at that
time. A public hearing on this decision was held in Washington, DC on September 1, 1992; and
1-6
-------
public comments were received through October 9, 1992. The final decision not to revise the
1-h NAAQS was published in the Federal Register on March 9, 1993 (58 FR 13008). However,
that decision did not take into consideration a number of more recent studies on the health
and welfare effects of O3 that had been published since the last of the literature assessed in
the O3 Supplement (i.e., studies available through 1985 and into early 1986).
The Agency initiated consideration of such studies as part of the next congress!onally-
mandated periodic review of O3 criteria and NAAQS. The new studies were assessed in revised
draft O3 AQCD chapters that were peer reviewed in July and September 1993 workshops,
followed by public release of the O3 AQCD First External Review Draft in February 1994 and
CAS AC review on July 20-21, 1994. Further drafts of the O3 AQCD, revised in response to
public comments and CAS AC review, were reviewed by CAS AC on March 21-25, 1995, and at
a final CAS AC review meeting on September 19-20, 1995. The scientific soundness of the
revised O3 AQCD was recognized by CAS AC in a November 28, 1995 letter to the EPA
Administrator; and the final O3 AQCD was published in July 1996.
The first draft of the associated Staff Paper, "Review of the National Ambient Air Quality
Standards for Ozone: Assessment of Scientific and Technical Information, " was also reviewed
by CAS AC at the March 21-22, 1995 public meeting. CAS AC also reviewed subsequent drafts
of the Staff Paper at public meetings on September 19-20, 1995 and March 21, 1996, with
completion of CAS AC review of the primary and secondary standard portions of the draft
Staff Paper being communicated in letters to the EPA Administrator dated November 30, 1995
and April 4, 1996, respectively. The final O3 Staff Paper was published in June 1996 (U.S.
Environmental Protection Agency, 1996b).
On December 13, 1996, EPA published its proposed decision to revise the O3 NAAQS
(61 FR 65716). Extensive opportunities for public comment on the proposed decision, including
several public hearings and two national satellite telecasts, were then provided by EPA; and
EPA's final decision to promulgate a new 8-h O3 NAAQS (see Table 1-1) was published on July
18, 1997 (62 FR 38856).
Following promulgation of the new standards, numerous petitions for review of the
standards were filed in the U.S. Court of Appeals for the District of Columbia Circuit (D.C.
Circuit)2. On May 14, 1999, the Court remanded the O3 NAAQS to EPA, finding that
:''American Trucking Associations v'. EPA, No. 97-1441.
1-7
-------
Section 109 of the CAA, as interpreted by EPA, effected an unconstitutional delegation of
legislative authority3. In addition, the Court directed that, in responding to the remand, EPA
should consider the potential beneficial health effects of O3 pollution in shielding the public from
the effects of solar ultraviolet (UV) radiation. On January 27, 2000, EPA petitioned the U.S.
Supreme Court for certiorari on the constitutional issue (and two other issues), but did not
request review of the D.C. Circuit ruling regarding the potential beneficial health effects of O3.
On February 27, 2001, the U.S. Supreme Court unanimously reversed the judgment of the D.C.
Circuit on the constitutional issue (holding that section 109 of the CAA does not delegate
legislative power to the EPA in contravention of the Constitution) and remanded the case to the
D.C. Circuit to consider challenges to the O3 NAAQS that had not been addressed by that Court's
earlier decisions4. On March 26, 2002, the D.C. Circuit issued its final decision, finding that the
1997 O3 NAAQS were "neither arbitrary nor capricious," and denied the remaining petitions
for review5.
On November 14, 2001, EPA proposed to respond to the Court's remand to consider the
potential beneficial health effects of O3 pollution in shielding the public from the effects of solar
UV radiation by leaving the 1997 8-h NAAQS unchanged. Following a review of information in
the record and the substantive comments received on the proposed response, EPA issued a final
response to the remand, reaffirming the 8 h O3 NAAQS (68 FR 614, January 6, 2003).
1.2 CURRENT OZONE CRITERIA AND NAAQS REVIEW
1.2.1 Key Milestones and Procedures for Document Preparation
It is important to note at the outset that development of the present O3 AQCD included
substantial external expert review and opportunities for public input through (a) public
workshops involving the general scientific community, (b) iterative reviews of successive
AQCD drafts by CASAC, and (c) comments from the public on successive drafts. Extensive
external inputs received through such reviews help to ensure that the current periodic review of
3 American Trucking Associations v. EPA, 175 F.3d 1027 (D.C. Cir, 1999).
^Whitman v. American Trucking Associations, 531 U.S. 457 (2001).
^American Trucking Associations v. EPA, 283 F.3d 355, (D.C. Cir. 2002).
1-8
-------
the O3 standards is based on critical assessment of the latest available pertinent science as
presented in this document and drawn upon in the associated Ozone Staff Paper.
The procedures for developing this revised O3 AQCD build on experience derived from the
other recent criteria document preparation efforts, with key milestones for development of
this O3 AQCD being listed in Table 1-2. Briefly, respective responsibilities for production of the
document and key milestones are as follows. An NCEA-RTP Ozone Team created and
implemented a project plan for developing the O3 AQCD, taking into account input from
individuals in other EPA program and policy offices identified as part of the EPA Ozone Work
Group. The resulting plan, i.e., the "Project Work Plan for Revised Air Criteria for Ozone and
Related Photochemical Oxidants" (November 2002), was discussed with CAS AC in January
2003. Under the processes established in Sections 108 and 109 of the CAA, the EPA officially
initiated the current criteria and NAAQS review by announcing the commencement of the
review in the Federal Register (65 FR 57810, September, 2000) with a call for information. That
Federal Register notice included (1) a request asking for recently available research information
on O3 that may not yet have been published and (2) a request for individuals with the appropriate
type and level of expertise to contribute to the writing of O3 AQCD materials to identify
themselves. The specific authors of chapters or sections of the proposed document included both
EPA and non-EPA scientific experts, who were selected on the basis of their expertise on the
subject areas and their familiarity with the relevant literature. The project team defined critical
issues and topics to be addressed by the authors and provided direction in order to focus on
evaluation of those studies most clearly identified as important for standard setting. An ongoing
literature search that was underway prior to initiation of work on this document continued
throughout its preparation to identify pertinent O3 literature published since early 1996.
As with other NAAQS reviews, critical assessment of relevant scientific information is
presented in this updated O3 AQCD. The main focus of this document is the evaluation and
interpretation of pertinent atmospheric science information, air quality data, human exposure
information, and health and welfare effects information newly published since that assessed in
the 1996 O3 AQCD. Draft versions of AQCD chapter materials were evaluated via expert peer-
consultation workshop discussions (see Table 1-2) that focused on the selection of pertinent
studies to be included in the chapters, the potential need for additional information to be added to
the chapters, and the quality of the characterization and interpretation of the literature. The
1-9
-------
Table 1-2. Key Milestones for Development of Revised Ozone Air Quality
Criteria Document (O, AQCD)
Major Milestones
1. Literature Search
2. Federal Register Call for Information
3. Draft Project Plan Available for Public Comment
4. Revised Draft Project Plan Released for CASAC Consultation
5. CASAC Consultation on Draft Project Work Plan
6. Peer-Consultation Workshop on Draft Ecological Effects Materials
7. Peer-Consultation Workshops on Draft Atmospheric
Science/Exposure and Dosimetry/Health Chapters
8. First External Review Draft of O3 AQCD
9. Public Comment Period (90 days)
10. CASAC Public Review Meeting (First External Review Draft)
11. Second External Review Draft of O3 AQCD
12. Public Comment Period (90 days)
13. CASAC Public Review Meeting (Second External Review Draft)
14. Final O3 AQCD
Dates
Ongoing
September 2000
Dec 2001-March 2002
December 2002
January 2003
April 2003
July 2004
January 2005
Feb - April 2005
May 4-5, 2005
August 2005
Sept - Nov 2005
December 6-8, 2005
February 28, 2006
authors of the draft chapters then revised them on the basis of the workshop and/or other expert
review comments6. These and other integrative materials were then incorporated into the First
External Review Draft (January 2005) of this O3 AQCD, which was made available for public
comment and CASAC review, as indicated in Table 1-2.
Following review of the First External Review Draft at a May 4-5, 2005 CASAC meeting,
EPA incorporated revisions into the draft O3 AQCD in response to comments from CASAC and
the public and made a Second External Review Draft (August, 2005) available for further public
comment and CASAC review as shown in Table 1-2. More specifically, the Second External
Review Draft underwent public comment during September-November, 2005, and was reviewed
by CASAC at a December 6-8, 2005 public meeting. This final O3 AQCD, completed by
6It should be noted that materials contributed by non-EPA authors were, at times, modified by EPA Ozone
Team staff in response to internal and/or external review comments and that EPA is responsible for the ultimate
content of this O3 AQCD.
1-10
-------
February 28, 2006, incorporates revisions made in response to public comments and CASAC
reviews of the earlier draft AQCD materials. An electronic version of this document can be
accessed via an EPA website at: www.epa.gov/ncea.
The EPA's Office of Air Quality Planning and Standards (OAQPS) staff is also preparing
further draft O3 Staff Paper materials which draw upon key information contained in this final O3
AQCD. After review of that draft O3 Staff Paper by the public and by CASAC, EPA will take
public and CASAC comments into account in producing a Final Ozone Staff Paper. That Staff
Paper, in final form, will present options for consideration by the Administrator of EPA
regarding whether to retain or, if appropriate, to revise the O3 NAAQS.
1.3 ORGANIZATIONAL STRUCTURE OF THE DOCUMENT
1.3.1 General Document Format
The general format used in preparing this O3 AQCD is to open each new section for the
updated document with concise summarization of key findings and conclusions from the
previous 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996a). After presentation of
such background information, the remainder of each section typically provides an updated
discussion of newer literature and resulting key conclusions. In some cases where no new
information is available, the summary of key findings and conclusions from the previous criteria
document must suffice as the basis for current key conclusions. Increased emphasis is placed in
the main chapters of this revised O3 AQCD on interpretative evaluation and integration of
evidence pertaining to a given topic than has been typical of previous EPA air quality criteria
documents, with more detailed descriptions of individual studies being provided in a series of
accompanying annexes.
A list of references published since completion of the 1996 criteria document was made
available to the authors. The references were selected from information data base searches
conducted by EPA. Additional references were added to the list (e.g., missed or recently
published papers or "in press" publications) as work proceeded in creating draft document
materials. As an aid in selecting pertinent new literature, the authors were also provided with a
summary of issues that needed to be addressed in this revised O3 AQCD. These issues were
identified by NCEA-RTP Ozone Team members, by the EPA Ozone Work Group, and by
1-11
-------
authors and reviewers of draft O3 AQCD materials, and they were further expanded, as
appropriate, based on public discussions, workshops, or other comments received by EPA in the
course of development of this document.
1.3.2 Organization and Content of the Document
This revised AQCD for Ozone and Related Photochemical Oxidants critically assesses
scientific information on the health and welfare effects associated with exposure to the
concentrations of these pollutants in ambient air. The document does not provide a detailed
literature review; but, rather, discusses cited references that reflect the current state of knowledge
on the most relevant issues pertinent to the derivation of NAAQS for O3 and/or related
photochemical oxidants. Although emphasis is placed on discussion of health and welfare
effects information, other scientific data are presented and evaluated in order to provide a better
understanding of the nature, sources, distribution, measurement, and concentrations of O3 and
related photochemical oxidants in ambient air, as well as the measurement of population
exposure to these pollutants.
The main focus of the scientific information discussed in the text comes from literature
published since completion of the 1996 O3 AQCD (U.S. Environmental Protection Agency,
1996a). Emphasis is placed on studies conducted at or near O3 concentrations found in ambient
air. Other studies are included if they contain unique data, such as the documentation of a
previously unreported effect or of a mechanism for an observed effect; or if they were multiple-
concentration studies designed to elucidate exposure-response relationships. Generally, this is
not an issue for human clinical or epidemiology studies. However, for animal toxicology
studies, consideration is given mainly to those studies conducted at less than 1 ppm O3.
Key information from studies assessed in the previous O3 AQCD and whose data impacted the
derivation of the current NAAQS are briefly summarized in the text, along with specific citations
to the previous document. Prior studies are also discussed if they (1) are open to reinterpretation
in light of newer data, or (2) are potentially useful in deriving revised standards for O3.
Generally, only information that has undergone scientific peer review and has been published
(or accepted for publication) through December 2004 is included in this draft document. A few
particularly pertinent and important new studies published or accepted for publication beyond
the end of 2004 are also considered.
1-12
-------
This revised O3 AQCD consists of three volumes. The first volume includes an Executive
Summary and Conclusions, as well as Chapters 1 through 11 of the document. This introductory
chapter (Chapter 1) presents background information on the purpose of the document, legislative
requirements, and the history of past O3 NAAQS regulatory actions, as well as an overview of
the organization and content of the document. Chapter 2 provides information on the physics
and chemistry of O3 and related photochemical oxidants in the atmosphere. Chapter 3 covers
tropospheric O3 environmental concentrations, patterns, and exposures. The accompanying
annexes to each of these background chapters are found in Volume II.
Health information pertinent to derivation of the primary O3 NAAQS is then mainly
covered in the next several chapters (Chapters 4 through 8). Chapter 4 discusses O3 dosimetry
aspects; and Chapters 5, 6, and 7 discuss animal toxicological studies, controlled-exposure
studies of human health effects, and epidemiologic studies of ambient air exposure effects on
human populations, respectively. Chapter 8 then provides an integrative and interpretive
evaluation of key information relevant to O3 exposure and health risks of most pertinence to the
review of primary O3 NAAQS. The annexes to these health-related chapters are found in
Volume II.
The remaining three chapters of the document assess welfare effects information pertinent
to the review of secondary O3 NAAQS. More specifically, Chapter 9 deals with ecological and
other environmental effects of O3 and related photochemical oxidants. Chapter 10 assesses
tropospheric O3 involvement in climate change processes, including impacts on solar UV flux in
Earth's lower atmosphere. Lastly, Chapter 11 discusses O3 effects on man-made materials as a
third type of welfare effect of potential concern. Annex materials related to welfare effects
(especially vegetation/ecological effects) are contained in Volume III.
1-13
-------
REFERENCES
Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30)
36: 8186-8201.
Federal Register. (1977) Review of the photochemical oxidant and hydrocarbon air quality standards. F. R.
(April 20) 42: 20493-20494.
Federal Register. (1979) National primary and secondary ambient air quality standards: revisions to the national
ambient air quality standards for photochemical oxidants. F. R. (February 8) 44: 8202-8237.
Federal Register. (1982) Air quality criteria document for ozone and other photochemical oxidants. F. R. (March 17)
47: 11561.
Federal Register. (1983) Review of the national ambient air quality standards for ozone. F. R. (August 22)
48: 38009.
Federal Register. (1984) Draft air quality criteria document for ozone and other photochemical oxidants. F. R.
(July 24) 49: 29845.
Federal Register. (1985) Science Advisory Board; Clean Air Scientific Advisory Committee; open meeting. F. R.
(February 13) 50: 6049.
Federal Register. (1986) Science Advisory Board; Clean Air Scientific Advisory Committee; open meeting. F. R.
(April 2) 51: 11339.
Federal Register. (1992) National ambient air quality standards for ozone; proposed decision. F. R. (August 10)
57: 35542-35557.
Federal Register. (1993) National ambient air quality standards for ozone - final decision. F. R. (March 9)
58: 13008-13019.
Federal Register. (1996) National ambient air quality standards for ozone: proposed decision. F. R. (December 13)
61:65,716-65,750.
Federal Register. (1997) National ambient air quality standards for ozone; final rule. F. R. (July 18)
62: 38856-38896.
Federal Register. (2000) Air Quality Criteria for Ozone and Related Photochemical Oxidants; notice; call for
information. F. R. (September 26) 65: 57810.
Federal Register. (2003) National ambient air quality standards for ozone: final response to remand; final rule. F. R.
(January 6) 68: 614-645.
U.S. Code. (2003a) Clean Air Act, §108, air quality criteria and control techniques.. U. S. C. 42: §7408.
U.S. Code. (2003b) Clean Air Act, §109, national ambient air quality standards. U. S. C. 42: §7409.
U.S. Court of Appeals for the District of Columbia. (1999a) American Trucking Associations, Inc. v. U.S.
Environmental Protection Agency. 195 F.3d 4 (D.C. Cir. 1999).
U.S. Court of Appeals for the District of Columbia. (1999b) American Trucking Associations, Inc. v. U.S.
Environmental Protection Agency. 175 F.3d 1027 (D.C. Cir. 1999).
U.S. Court of Appeals for the District of Columbia. (2002) American Trucking Associations, Inc. v. U.S.
Environmental Protection Agency. 283 F.3d 355, 378-79 (D.C. Cir. 2002).
U.S. Environmental Protection Agency. (1992) Summary of selected new information on effects of ozone on health
and vegetation: supplement to 1986 air quality criteria for ozone and other photochemical oxidants. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; report no. EPA/600/8-88/105F. Available from: NTIS, Springfield, VA; PB92-235670.
U.S. Environmental Protection Agency. (1996a) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
http ://cfpub2. epa. gov/ncea/.
U.S. Environmental Protection Agency. (1996b) Review of national ambient air quality standards for ozone:
assessment of scientific and technical information. OAQPS staff paper. Research Triangle Park, NC:
Office of Air Quality Planning and Standards; report no. EPA/452/R-96/007. Available from: NTIS,
Springfield, VA; PB96-203435. Available: http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_pr_sp.html
(29 September 2005).
U.S. Supreme Court. (2001) Whitman v. American Trucking Association. 531 U.S. 457 (nos. 99-1257 and 99-1426).
1-14
-------
2. PHYSICS AND CHEMISTRY OF OZONE
IN THE ATMOSPHERE
2.1 INTRODUCTION
Ozone (O3) and other oxidants, such as peroxacyl nitrates and hydrogen peroxide (H2O2)
form in polluted areas by atmospheric reactions involving two main classes of precursor
pollutants, volatile organic compounds (VOCs) and nitrogen oxides (NOX). Carbon monoxide
(CO) is also important for O3 formation in polluted areas. Ozone is thus a secondary pollutant.
The formation of O3, other oxidants and oxidation products from these precursors is a complex,
nonlinear function of many factors: the intensity and spectral distribution of sunlight;
atmospheric mixing and processing on cloud and aerosol particles; the concentrations of the
precursors in ambient air; and the rates of chemical reactions of the precursors. Information
contained in this chapter and in greater detail in Annex AX2 describes these processes,
numerical models that incorporate these processes to calculate O3 concentrations, and techniques
for measuring concentrations of ambient oxidants.
The atmosphere can be divided into several distinct vertical layers, based primarily on the
major mechanisms by which they are heated and cooled. The lowest major layer is the
troposphere, which extends from the earth's surface to about 8 km above polar regions and to
about 16 km above tropical regions. The planetary boundary layer (PEL) is the lower sublayer
of the troposphere, extending from the surface to about 1 or 2 km and is most strongly affected
by surface conditions. The stratosphere extends from the tropopause, or the top of the
troposphere, to about 50 km in altitude (Annex AX2.2.1). The emphasis in this chapter is placed
on chemical and physical processes occurring in the troposphere, in particular in the PEL. The
processes responsible for producing summertime O3 episodes are fairly well understood, as
discussed in the previous Air Quality Criteria Document for Ozone and Related Photochemical
Oxidants or 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996). This chapter
mainly considers topics for which there is substantial new information and on topics that form
the basis for discussions in later chapters.
2-1
-------
2.2 CHEMICAL PROCESSES INVOLVED IN OZONE FORMATION
AND DESTRUCTION
Ozone occurs not only in polluted urban atmospheres but throughout the troposphere, even
in remote areas of the globe. The same basic processes, involving sunlight-driven reactions
of NOX and VOCs contribute to O3 formation throughout the troposphere. These processes also
lead to the formation of other photochemical products, such as peroxyacetyl nitrate (PAN), nitric
acid (HNO3), and sulfuric acid (H2SO4), and to other compounds, such as formaldehyde (HCHO)
and other carbonyl compounds, such as aldehydes and ketones.
The photochemical formation of O3 in the troposphere proceeds through the oxidation of
nitric oxide (NO) to nitrogen dioxide (NO2) by organic (RO2) or hydro-peroxy (HO2) radicals.
The photolysis of NO2 yields nitric oxide (NO) and a ground-state oxygen atom, O(3P), which
then reacts with molecular oxygen to form O3. Free radicals oxidizing NO to NO2 are formed
during the oxidation of VOCs (Annex AX2.2.2).
The term VOC refers to all carbon-containing gas-phase compounds in the atmosphere,
both biogenic and anthropogenic in origin, excluding CO and CO2. Classes of organic
compounds important for the photochemical formation of O3 include alkanes, alkenes, aromatic
hydrocarbons, carbonyl compounds (e.g., aldehydes and ketones), alcohols, organic peroxides,
and halogenated organic compounds (e.g., alkyl halides). This array of compounds encompasses
a wide range of chemical properties and lifetimes: isoprene has an atmospheric lifetime of
approximately an hour, whereas methane has an atmospheric lifetime of about a decade.
In urban areas, compounds representing all classes of VOCs and CO are important for O3
formation. In nonurban vegetated areas, biogenic VOCs emitted from vegetation tend to be the
most important. In the remote troposphere, CH4 and CO are the main carbon-containing
precursors to O3 formation. CO also can play an important role in O3 formation in urban areas.
The oxidation of VOCs is initiated mainly by reaction with hydroxyl (OH) radicals. The primary
source of OH radicals in the atmosphere is the reaction of electronically excited O atoms, O(JD),
with water vapor. O(JD) is produced by the photolysis of O3 in the Hartley bands. In polluted
areas, the photolysis of aldehydes (e.g., HCHO), nitrous acid (HONO) and hydrogen
peroxide (H2O2) can also be significant sources of OH or HO2 radicals that can rapidly be
converted to OH (Eisele et al., 1997). Ozone can oxidize alkenes; and, at night, when they are
most abundant, NO3 radicals also oxidize alkenes. In coastal environments and other selected
2-2
-------
environments, atomic Cl and Br radicals can also initiate the oxidation of VOCs (Annex
AX2.2.3).
There are a large number of oxidized nitrogen containing compounds in the atmosphere
including NO, NO2, NO3, HNO2, HNO3, N2O5, HNO4, PAN and its homologues, other organic
nitrates and particulate nitrate. Collectively these species are referred to as NOy. Oxidized
nitrogen compounds are emitted to the atmosphere mainly as NO which rapidly interconverts
with NO2 and so NO and NO2 are often "lumped" together into their own group or family,
or NOX. NOX can be oxidized to reservoir and termination species (PAN and its homologues,
organic nitrates, HNO3, HNO4 and particulate nitrate). These reservoir and termination species
are referred to as NOZ. The major reactions involving interconversions of oxidized nitrogen
species are discussed in Annex AX2.2.4.
The photochemical cycles by which the oxidation of hydrocarbons leads to O3 production
are best understood by considering the oxidation of methane, structurally the simplest VOC.
The CH4 oxidation cycle serves as a model for the chemistry of the relatively clean or unpolluted
troposphere (although this is a simplification because vegetation releases large quantities of
complex VOCs, such as isoprene, into the atmosphere). In the polluted atmosphere, the
underlying chemical principles are the same, as discussed in Annex AX2.2.5. The conversion of
NO to NO2 occurring with the oxidation of VOCs is accompanied by the production of O3 and
the efficient regeneration of the OH radical, which in turn can react with other VOCs.
A schematic overview showing the major processes involved in O3 production and loss in the
troposphere and stratosphere is given in Figure 2-1.
The oxidation of alkanes and alkenes in the atmosphere has been treated in depth in
1996 O3 AQCD and is updated in Annexes AX2.2.6 and AX2.2.7. In contrast to simple
hydrocarbons containing one or two carbon atoms, detailed kinetic information about the gas
phase oxidation pathways of many anthropogenic hydrocarbons (e.g., aromatic compounds such
as benzene and toluene), biogenic hydrocarbons (e.g., isoprene, the monoterpenes), and their
intermediate oxidation products (e.g., epoxides, nitrates, and carbonyl compounds) is lacking.
Reaction with OH radicals represents the major loss process for alkanes. Reaction with chlorine
atoms is an additional sink for alkanes. Stable products of alkane photooxidation are known to
include carbonyl compounds, alkyl nitrates, and J-hydroxycarbonyls. Major uncertainties in the
atmospheric chemistry of the alkanes concern the chemistry of alkyl nitrate formation; these
-------
Stratosphere
Non.Po.ar
Figure 2-1. Schematic overview of O3 photochemistry in the stratosphere
and troposphere.
uncertainties affect the amount of NO-to-NO2 conversion occurring and, hence, the amounts
of O3 formed during photochemical degradation of the alkanes.
The reaction of OH radicals with aldehydes produced during the oxidation of alkanes
forms acyl (R'CO) radicals, and acyl peroxy radicals (R'C(O)-O2) are formed by the further
addition of O2. As an example, the oxidation of ethane (C2H5-H) yields acetaldehyde
(CH3-CHO). The reaction of CH3-CHO with OH radicals yields acetyl radicals (CH3-CO).
The acetyl radicals will then participate with O2 in a termolecular recombination reaction to form
acetyl peroxy radicals, which can then react with NO to form CH3 + CO2 or they can react with
2-4
-------
NO2 to form PAN. PAN acts as a temporary reservoir for NO2. Upon the thermal
decomposition of PAN, either locally or elsewhere, NO2 is released to participate in the O3
formation process again.
Alkenes react in ambient air with OH, NO3, and Cl radicals and with O3. All of these
reactions are important atmospheric transformation processes, and all proceed by initial addition
to the >C=C< bonds. Products of alkene photooxidation include carbonyl compounds,
hydroxynitrates and nitratocarbonyls, and decomposition products from the energy-rich
biradicals formed in alkene-O3 reactions. Major uncertainties in the atmospheric chemistry of
the alkenes concern the products and mechanisms of their reactions with O3, especially the yields
of free radicals that participate in O3 formation. Examples of oxidation mechanisms of complex
alkanes and alkenes can be found in comprehensive texts such as Seinfeld and Pandis (1998).
The oxidation of aromatic hydrocarbons constitutes an important component of the
chemistry of O3 formation in urban atmospheres (Annex AX2.2.8). Virtually all of the important
aromatic hydrocarbon precursors emitted in urban atmospheres are lost through reaction with the
hydroxyl radical. Loss rates for these compounds vary from slow (i.e., benzene) to moderate
(e.g., toluene), to very rapid (e.g., xylene and trimethylbenzene isomers). These loss rates are
very well understood at room temperature and atmospheric pressure, and numerous experiments
have been conducted that verify this. However, the mechanism for the oxidation of aromatic
hydrocarbons following reaction with OH is poorly understood, as evident from the poor mass
balance of the reaction products. The mechanism for the oxidation of toluene has been studied
most thoroughly, and there is general agreement on the initial steps in the mechanism. However,
at present there is no promising approach for resolving the remaining issues concerning the later
steps. The oxidation of aromatic hydrocarbons also leads to particle formation which could
remove gas-phase constituents that participate in O3 formation. What is known of the chemistry
of secondary organic aerosol formation from gaseous precursors was summarized in the latest
PM AQCD (U.S. Environmental Protection Agency, 2004).
The reactions of oxygenated VOCs are also important components of O3 formation
(Annex AX2.2.9). They may be produced either by the oxidation of hydrocarbons or they may
be present in ambient air as the result of direct emissions. For example, motor vehicles and
some industrial processes emit formaldehyde and vegetation emits methanol.
2-5
-------
As much as 30% of the carbon in hydrocarbons in many urban areas is in the form of
aromatic compounds. Yet, mass balance analyses performed on irradiated smog chamber
mixtures of aromatic hydrocarbons indicate that only about one-half of the carbon is in the form
of compounds that can be identified. The situation is not much better for some smaller
anthropogenic hydrocarbons. For example, only about 60% of the initial carbon can be
accounted for in the OH initiated oxidation of 1,3-butadiene. About two-thirds of the initial
carbon can be identified in product analyses of isoprene oxidation. Adequate analytical
techniques needed to identify and quantify key intermediate species are not available for many
compounds. In addition, methods to synthesize many of the suspected intermediate compounds
are not available so that laboratory studies of their reaction kinetics cannot be performed.
Similar considerations apply to the oxidation of biogenic hydrocarbons besides isoprene.
In addition to reactions occurring in the gas phase, reactions occurring on the surfaces of or
within cloud droplets and airborne particles also occur. Their collective surface area is huge,
implying that collisions with gas phase species occur on very short time scales. In addition to
hydrometeors (e.g., cloud and fog droplets and snow and ice crystals) there are also potential
reactions involving atmospheric particles of varying composition (e.g., wet [deliquesced]
inorganic particles, mineral dust, carbon chain agglomerates and organic carbon particles)
to consider. Most of the well-established multiphase reactions tend to reduce the rate of O3
formation in the polluted troposphere. Removal of HOX and NOX onto hydrated particles will
reduce the production of O3. However, the photolysis of HONO formed in reactions such as
these can increase the production of O3. The reactions of Br and Cl containing radicals
deplete O3 in selected environments such as the Arctic during spring, the tropical marine
boundary layer and inland salt flats and salt lakes. Direct reactions of O3 and atmospheric
particles appear to be too slow to reduce O3 formation significantly at typical ambient PM levels.
In addition, the oxidation of hydrocarbons by Cl radicals could lead to the rapid formation of
peroxy radicals and higher rates of O3 production in selected coastal environments. It should be
stressed that knowledge of multiphase processes is still evolving and there are still many
questions that remain to be answered as outlined in Annex AX2.2.10.
The oxidants, other than O3, that are formed from the chemistry described above could
exert effects on human health and perhaps also on vegetation. Gas phase oxidants include
PAN, H2O2 and CH3OOH and other organic hydroperoxides (Annex AX2.2). In addition to
2-6
-------
transfer from the gas phase, oxidants can be formed by photochemical reactions occurring in
particles (Annex 2.2.10.6). However, the pathways leading to the formation of oxidants in the
particle phase are not as well understood as they are in the gas phase. However, it is to be
expected that pathways leading to the formation of gas phase oxidants and secondary organic
aerosols are linked to some degree. In addition, the reaction of O3 with isoprene and other
biogenic hydrocarbons may also form oxidants in particles.
Reactions of O3 with monoterpenes have been shown to produce oxidants in the aerosol
phase. Docherty et al. (2005) found evidence for the substantial production of organic
hydroperoxides in secondary organic aerosol (SOA) resulting from the reaction of monoterpenes
with O3. Analysis of the SOA formed in their environmental chamber indicated that the SOA
was mainly organic hydroperoxides. In particular, they obtained yields of 47% and 85% of
organic peroxides from the oxidation of a- and p-pinene. The hydroperoxides then react with
aldehydes in particles to form peroxyhemiacetals, which can either rearrange to form other
compounds such as alcohols and acids or revert back to the hydroperoxides. The aldehydes are
also produced in large measure during the ozonolysis of the monoterpenes. Monoterpenes also
react with OH radicals resulting, however, in the production of more lower molecular weight
products than in their reaction with O3. Bonn et al. (2004) estimated that hydroperoxides lead to
63% of global SOA formation from the oxidation of terpenes. The oxidation of anthropogenic
aromatic hydrocarbons by OH radicals may also produce organic hydroperoxides in SOA
(Johnson et al., 2004). Although the results of chamber and modeling studies indicate
substantial production of organic hydroperoxides, it should be noted that data for organic
hydroperoxides in ambient aerosol samples are sparse.
Ozone chemical reactions that occur indoors are analogous to those occurring in ambient
air. In the indoor environment, O3 reacts with unsaturated VOCs, primarily terpenes or terpene-
related compounds from cleaning products, air fresheners, and wood products. The reactions are
dependent on the O3 indoor concentration, the indoor temperature and, in most cases, the air
exchange rate/ventilation rate. Some of the reaction products may more negatively impact
human health and artifacts in the indoor environment than their precursors (Wolkoff et al., 1999;
Wilkins et al., 2001; Weschler et al., 1992; Weschler and Shields, 1997; Rohr et al., 2002;
N0jgaard et al., 2005). Primary reaction products are Criegee biradicals, nitrate radicals,
and peroxyacetyl radicals. Secondary reaction products are hydroxy, alkyl, alkylperoxy,
2-7
-------
hydroperoxy, and alkoxy radicals. Reactions with alkenes can produce aldehydes, ketones,
and organic acids (Weschler and Shields, 2000; Weschler et al., 1992).
2.3 METEOROLOGICAL PROCESSES AFFECTING OZONE
Since the 1996 O3 AQCD, substantial new information about transport processes has
become available from numerical models, field experiments and satellite-based observations.
Ozone is produced naturally by photochemical reactions in the stratosphere, as shown in Figure
2-1. Some of this O3 is transported downward into the troposphere throughout the year, with
maximum contributions during late winter and early spring mainly in a process known as
tropopause folding. Figure 2-2a shows a synoptic situation associated with a tropopause folding
event. A vertical cross section taken through the atmosphere from a to a' is shown in Figure
2-2b. In this figure, the tropopause fold is shown folding downward above and slightly behind
the surface cold front, bringing stratospheric air with it. Although the tropopause is drawn with
a solid line, it should not be taken to mean that it is a material surface through which there is no
exchange. Rather these folds should be thought of as regions in which mixing of tropospheric
and stratospheric air is occurring (Shapiro, 1980). This imported stratospheric air contributes to
the natural background of O3 in the troposphere, especially in the free troposphere. It should be
noted that there is considerable uncertainty in the magnitude and distribution of this potentially
important source of tropospheric O3. Stratospheric intrusions that reach the surface are rare.
Much more common are intrusions which penetrate only to the middle and upper troposphere.
However, O3 transported to the upper and middle troposphere can still affect surface
concentrations through various exchange mechanisms that mix air from the free troposphere
with air in the planetary boundary layer. Substantial photochemical production of O3 in the
troposphere also begins in late winter and early spring; therefore, it cannot be assumed that O3
present at these times is only stratospheric in origin. The basic atmospheric dynamics and
thermodynamics of stratospheric-tropospheric exchange are outlined in Annex AX2.3.1.
Our understanding of the meterological processes associated with summertime O3 episodes
remains basically the same as outlined in the 1996 O3 AQCD. Major episodes of high O3
concentrations in the eastern United States and in Europe are associated with slow moving, high
pressure systems. High pressure systems during the warmer seasons are associated with the
2-8
-------
12UTC, 23Feb94
103.6kPa
Figure 2-2a. Surface weather chart showing sea level (MSL) pressure (kPa), and
surface fronts.
Source: Stall (2000).
CO
'3>
a CYYC
LBF
LCH a'
Figure 2-2b. Vertical cross section along dashed line (a-a') from northwest to the
southeast (CYYC = Calgary, Alberta; LBF = North Platte, NB; LCH = Lake
Charles, LA). The approximate location of the jet stream core is indicated
by the hatched area. The position of the surface front is indicated by the
cold-frontal symbols and the frontal inversion top by the dashed line.
Note: This is 12 h later than the situations shown in Figure 2-2a.
Source: Adapted from Stull (2000).
2-9
-------
sinking of air, resulting in warm, generally cloudless skies, with light winds. The sinking of air
results in the development of stable conditions near the surface which inhibit or reduce the
vertical mixing of O3 precursors. The combination of inhibited vertical mixing and light winds
minimizes the dispersal of pollutants emitted in urban areas, allowing their concentrations to
build up. Photochemical activity involving these precursors is enhanced because of higher
temperatures and the availability of sunlight. In the eastern United States, high O3
concentrations during a large scale episode can extend over hundreds of thousands of square
kilometers for several days. These conditions have been described in greater detail in the
1996 O3 AQCD. The transport of pollutants downwind of major urban centers is characterized
by the development of urban plumes. However, the presence of mountain barriers limits mixing
(as in Los Angeles and Mexico City) and results in a higher frequency and duration of days with
high O3 concentrations. Ozone concentrations in southern urban areas (such as Houston, TX and
Atlanta, GA) tend to decrease with increasing wind speed. In northern cities (such as Chicago,
IL; New York, NY; Boston, MA; and Portland, ME), the average O3 concentrations over the
metropolitan areas increase with wind speed, indicating that transport of O3 and its precursors
from upwind areas is important (Husar and Renard, 1998; Schichtel and Husar, 2001).
Ozone and other secondary pollutants are determined by meteorological and chemical
processes extending typically over spatial scales of several hundred kilometers (e.g., Civerolo
et al., 2003; Rao et al., 2003). An analysis of the output of regional model studies conducted by
Kasibhatla and Chameides (2000) suggests that O3 can be transported over a few thousand
kilometers in the upper boundary layer of the eastern half of the United States during some O3
episodes. Convection is capable of transporting O3 and its precursors vertically through the
troposphere, as shown in Annex AX2.3.2. Nocturnal low level jets (LLJs) can also transport
pollutants hundreds of kilometers (Annex AX2.3.3). Schematic diagrams showing the
atmospheric conditions during the formation of LLJs and the regions in which they are most
prevalent are given in Figures 2-3 and 2-4. Such LLJs have also been observed off the coast of
California. Turbulence associated with LLJs can bring these pollutants to the surface and result
in secondary O3 maxima during the early morning in many locations (Corsmeier et al., 1997).
Aircraft observations indicate that there can be substantial differences in mixing ratios of
key species between the surface and the atmosphere above (Fehsenfeld et al., 1996; Berkowitz
and Shaw, 1997). In particular, mixing ratios of O3 can be higher in the lower free troposphere
2-10
-------
2-
u>
•55 1 -
X
0-
Free Troposphere
Cloud Layer
Cloud Layer
CT Low Level Jet,__>
Stable Nocturnal Boundary Layer
Mixed
Layer
Surface Layer
\
Afternoon
I
Sunset
Midnight
Sunrise
Noon
Figure 2-3. The diurnal evolution of the planetary boundary layer (PEL) while high
pressure prevails over land. Three major layers exist (not including the
surface layer): a turbulent mixed layer; a less turbulent residual layer which
contains former mixed layer air; and a nocturnal, stable boundary layer
that is characterized by periods of sporadic turbulence.
Source: Adapted from Figures 1.7 and 1.12 of Stall (1999).
(aloft) than in the planetary boundary layer (PEL) during multiday O3 episodes (Taubmann et al.,
2004, 2005). These conditions are illustrated schematically in Figure 2-5. Convective processes
and small scale turbulence transport O3 and other pollutants both upward and downward
throughout the planetary boundary layer and the free troposphere. Ozone and its precursors can
be transported vertically by convection into the upper part of the mixed layer on one day, then
transported overnight as a layer of elevated mixing ratios, and then entrained into a growing
convective boundary layer downwind and brought back down to the surface. High O3
concentrations showing large diurnal variations at the surface in southern New England were
associated with the presence of such layers (Berkowitz et al., 1998). Because of wind shear,
winds several hundred meters above the ground can bring pollutants from the west, even though
surface winds are from the southwest during periods of high O3 in the eastern United States
(Blumenthal et al., 1997). These considerations suggest that, in many areas of the United States,
O3 formation involves processes occurring over hundreds if not thousands of square kilometers.
2-11
-------
Figure 2-4. Locations of low level jet occurrences in decreasing order of prevalence
(most frequent, common, observed). These locations are based on 2-years
radiosonde data obtained over limited areas. With better data coverage,
other low level jets may well be observed elsewhere in the United States.
Source: Bonner (1968).
Although the vast majority of measurements are made near the Earth's surface, there is
substantial photochemistry and transport of O3 occurring above the boundary layer in the free
troposphere. In the free troposphere, pollutants are chemically more stable and can be
transported over much longer distances and O3 is produced more efficiently than in the planetary
boundary layer. Results from the Atmosphere/Ocean Chemistry Experiment (AEROCE)
indicated that springtime maxima for surface O3 over the western North Atlantic Ocean result
from tropopause folding in close proximity to convective clouds (Annex AX2.3.4). The
convection lifts O3 and its precursors to the free troposphere where they mix with O3 from the
stratosphere and the mixture is transported eastward. Results from the North Atlantic Regional
Experiment (Annex AX2.3.4) indicate that summertime air is transported along the East Coast
northeastward and upward ahead of cold fronts. New England and the Maritime Provinces of
Canada receive substantial amounts of O3 and other pollutants through this mechanism.
2-12
-------
Two Reservoir Model
>
QJ
NO
SO2
O
CO
0
BC
0
0
> Actinic
Flux
a £1.9
.•' < Actinic
Flux
a a 1.9
0
VOCs Primary
Particles
Figure 2-5. Conceptual two-reservoir model showing conditions in the planetary
boundary layer (PEL) and in the lower free troposphere during a multiday O3
episode. The dotted line represents the top of PEL. Emissions occur in the
PEL, where small, unmixed black carbon, sulfate, and crustal particles in the
PM2 5 size range are also shown. Ozone concentrations as well as potential
temperature (0) and actinic flux are lower in the PEL than in the lower free
troposphere, while relative humidity and the Angstrom exponent for
aerosol scattering (a) are higher. Larger, internally mixed sulfate and
carbonaceous particles (still in the PM2 5 size range) and more O3 exist in the
lower free troposphere.
Source: Taubman et al. (2004, 2005).
Pollutants transported in this way can then be entrained in stronger and more stable westerly
winds aloft and can travel across the North Atlantic Ocean. The pollutants can then be brought
to the surface by subsidence in high pressure systems (typically behind the cold front in advance
of the one mentioned above). Thus, pollutants from North America can be brought down either
over the North Atlantic Ocean or in Europe. Pollutants can be transported across the North
Pacific Ocean from Asia to North America in a similar way. Behind an advancing cold front,
cold and dry stratospheric air is also being transported downward and southward. Stratospheric
2-13
-------
constituents and tropospheric constituents can then mix by small-scale turbulent exchange
processes. The results of these studies suggest that the mechanisms involved in the long-range
transport of O3 and its precursors are closely tied to the processes involved in stratospheric-
tropospheric exchange. Land-sea breezes affect the concentration and dispersal of pollutants in
coastal zone cities (Annex AX2.5).
The local rate of O3 formation depends on atmospheric conditions such as the availability
of solar ultraviolet radiation capable of initiating photolysis reactions, air temperatures and the
concentrations of chemical precursors (Annex AX2.3.6). The dependence of daily maximum
8-h O3 concentrations on daily maximum temperature is illustrated in Figure 2-6 for the
Baltimore, MD area. As can be seen, O3 concentrations tend to increase with temperature
(r = 0.74). However, this trend is absent in data from Phoenix, AZ as can be seen in Figure 2-7
(r = 0.14). These figures show that relations of O3 to precursor variables are location-specific
and relations observed in one area cannot be readily extrapolated to another. Factors that may be
responsible for the differences in O3 behavior in the two areas are discussed in Section AX2.3.6.
a
_Q.
•a
E
u
o
o
O
CO
E
nj 40-
s
>,
Q
n •
D
Q oy
D_n_ annj
n _°p By^E
° - D^BoBi|
a a BDOOQ al a
o o "uB"a n°|nBiDB
D OD|°aB na a||D
n D an D D aD°
B e°
i i
" B°no B
HnC !
BBn
ill I
H| p Q
|DHjpf|
°°° D°l
on
°°O°DBB D
nnc : nB||Dn
a ng u
°°J|P Bl|lnDn
°gg || |n|n 1 n
f ! llBO^"
B |oO° D
|lB|i
ID H
_ gnnD
a a
10
15
20
25
30
35
40
45
50
Figure 2-6. A scatter plot of daily maximum 8-h average O3 concentrations versus daily
maximum temperature for May through September 1994 to 2004 in the
Baltimore, MD Air Quality Forecast Area.
Source: Piety (2005).
2-14
-------
.a
a.
c
8
c
o
o
co
E
3
E
'x
ro
ra
O
160
140
120
100
80
60
40
20
10
15
20
25
30
35
40
45
50
Figure 2-7. A scatter plot of daily maximum 8-h average O3 concentrations versus daily
maximum temperature for May through September 1996 to 2004 at sites
downwind of Phoenix, AZ.
Source: Piety (2005).
2.4 RELATIONS OF OZONE TO ITS PRECURSORS
Rather than varying directly with emissions of its precursors, O3 changes in a nonlinear
fashion with the concentrations of its precursors (Annex AX2.4). At the low NOX concentrations
found in most environments, ranging from remote continental areas to rural and suburban areas
downwind of urban centers (low - NOX regime), the net production of O3 increases with
increasing NOX. At the high NOX concentrations found in downtown metropolitan areas,
especially near busy streets and roadways and in power plant plumes, there is scavenging
(titration) of O3 by reaction with NO (high - NOX regime). In between these two regimes, there is
a transition stage in which O3 shows only a weak dependence on NOX concentrations. In the
high - NOX regime, NO2 scavenges OH radicals which would otherwise oxidize VOCs to
produce peroxy radicals, which in turn would oxidize NO to NO2. In this regime, O3 production
is limited by the availability of free radicals. The production of free radicals is in turn limited by
the availability of solar UV radiation capable of photolyzing O3 (in the Hartley bands) or
2-15
-------
aldehydes and/or by the abundance of VOCs whose oxidation produce more radicals than they
consume. In the low-NOx regime, the overall effect of the oxidation of VOCs is to generate
(or at least not consume) free radicals, and O3 production varies directly with NOX. There are a
number of ways to refer to the chemistry in these two chemical regimes. Sometimes the terms
VOC-limited and NOx-limited are used. However, there are difficulties with this usage because
(1) VOC measurements are not as abundant as they are for nitrogen oxides, (2) rate coefficients
for reaction of individual VOCs with free radicals vary over an extremely wide range, and
(3) consideration is not given to CO nor to reactions that can produce free radicals without
involving VOCs. The terms NOx-limited and NOx-saturated (e.g., Jaegle et al., 2001) will be
used wherever possible to more adequately describe these two regimes. However, the
terminology used in original articles will also be used here.
The chemistry of OH radicals, which are responsible for initiating the oxidation of
hydrocarbons, shows behavior similar to that for O3 with respect to NOX concentrations
(Hameed et al., 1979; Pinto et al., 1993; Poppe et al., 1993; Zimmerman and Poppe, 1993).
These considerations introduce a high degree of uncertainty into attempts to relate changes in O3
concentrations to emissions of precursors. There are no definitive rules governing the levels
of NOX at which the transition from NOx-limited to NOx-saturated conditions occurs. The
transition between these two regimes is highly spatially and temporally dependent and depends
also on the nature and abundance of the hydrocarbons that are present.
Trainer et al. (1993) and Olszyna et al. (1994) have shown that O3 and NOy are highly
correlated in rural areas in the eastern United States. Trainer et al. (1993) also showed that O3
levels correlate even better with NOZ than with NOy, as may be expected because NOZ represents
the amount of NOX that has been oxidized, forming O3 in the process. NOZ is equal to the
difference between measured total reactive nitrogen (NOy) and NOX and represents the summed
products of the oxidation of NOX. NOZ is composed mainly of HNO3, PAN and other organic
nitrates, particulate nitrate, and HNO4.
Trainer et al. (1993) also suggested that the slope of the regression line between O3
and NOZ can be used to estimate the rate of O3 production per NOX oxidized (also known as
the O3 production efficiency, or OPE). Ryerson et al. (1998, 2001) used measured correlations
between O3 and NOZ to identify different rates of O3 production in plumes from large point
sources. A number of studies in the planetary boundary layer over the continental United States
2-16
-------
have found that the OPE ranges typically from one to nearly ten. However, it may be higher in
the upper troposhere and in certain areas, such as the Houston-Galveston area. Observations
indicate that the OPE depends mainly on the abundance of NOX.
Various techniques have been proposed to use ambient NOX and VOC measurements to
derive information about the dependence of O3 production on their concentrations. For example,
it has been suggested that O3 formation in individual urban areas could be understood in terms of
measurements of ambient NOX and VOC concentrations during the early morning (e.g., National
Research Council, 1991). In this approach, the ratio of summed (unweighted) VOC to NOX is
used to determine whether conditions were NOx-limited or VOC limited. This procedure is
inadequate because it omits many factors that are important for O3 production such as the impact
of biogenic VOCs (which are typically not present in urban centers during early morning);
important differences in the ability of individual VOCs to generate free radicals (rather than just
total VOC) and other differences in O3 forming potential for individual VOCs (Carter, 1995);
and changes in the VOC to NOX ratio due to photochemical reactions and deposition as air
moves downwind from urban areas (Milford et al., 1994).
Photochemical production of O3 generally occurs simultaneously with the production of
various other species such as nitric acid (HNO3), organic nitrates, and other oxidants such as
hydrogen peroxide. The relative rate of production of O3 and other species varies depending on
photochemical conditions, and can be used to provide information about O3-precursor
sensitivity. Sillman (1995) and Sillman and He (2002) identified several secondary reaction
products that show different correlation patterns for NOx-limited and NOx-saturated conditions.
The most important correlations are for O3 versus NOy, O3 versus NOZ, O3 versus HNO3,
and H2O2 versus HNO3. The correlations between O3 and NOy, and O3 and NOZ are especially
important because measurements of NOy and NOX are more widely available than for VOCs.
Measured O3 versus NOZ (Figure 2-8) shows distinctly different patterns in different locations.
In rural areas and in urban areas such as Nashville, TN, O3 is highly correlated with NOZ.
By contrast, in Los Angeles, CA, O3is not as highly correlated with NOZ, and the rate of increase
of O3 with NOZ is lower and the O3 concentrations for a given NOZ value are generally lower.
The different O3 versus NOZ relations in Nashville, TN and Los Angeles, CA reflects the
difference between NOx-limited conditions in Nashville versus an approach to NOX- saturated
conditions in Los Angeles.
2-17
-------
X X
10
20
30
40
NOZ (ppb)
Figure 2-8. Measured values of O3 and NOZ (NOy - NOX) during the afternoon at rural
sites in the eastern United States (grey circles) and in urban areas and urban
plumes associated with Nashville, TN (gray dashes); Paris, France (black
diamonds); and Los Angeles CA (Xs).
Sources: Trainer etal. (1993), Sillmanetal. (1997, 1998), SillmanandHe (2002).
The difference between NOx-limited and NOx-saturated regimes is also reflected in
measurements of hydrogen peroxide (H2O2). Hydrogen peroxide production is highly sensitive
to the abundance of free radicals and is thus favored in the NOx-limited regime. Measurements
in the rural eastern United States (Jacob et al., 1995), Nashville, TN (Sillman et al., 1998), and
Los Angeles, CA (Sakugawa and Kaplan, 1989), show large differences in H2O2 concentrations
between likely NOx-limited and NOx-saturated locations.
2.5 THE ROLE OF CHEMISTRY-TRANSPORT MODELS IN
UNDERSTANDING ATMOSPHERIC OZONE
Chemistry-transport models (CTMs) are used to improve understanding of atmospheric
chemical processes and to develop control strategies (Annex AX2.5). The main components of a
CTM are summarized in Figure 2-9. Models such as the CMAQ (Community Model for Air
Quality) system incorporate numerical algorithms describing the processes shown in Figure 2-9.
Also shown in Figure 2-9 is the meteorological model used to provide the inputs for calculating
2-18
-------
Initial/Boundary
Conditions and
Continuous Updates
of Met. Fields
from Observations
Meteorological
Model
Emissions
Model
Anthropogenic
(point, area sources)
Biogenic Emissions
>
' >
r
Gas-Phase
Chemistry //Aerosol \
Chemistry
Deposition \\ '/ and
Microphysics
Chemistry Transport Model
Visualization of Output
Process Analyses
Figure 2-9. Main components of a comprehensive atmospheric chemistry modeling
system, such as Models-3.
the transport of species in the CTM. Meteorological models, such as MM5, which supply these
inputs to the CTMs mentioned above, also provide daily weather forecasts. The domains of
these models extend typically over areas of millions of square kilometers.
Because these models are computationally intensive, it is often impractical to run them
over larger domains without sacrificing some features. For these reasons, both the
meteorological model and the CTM rely on boundary conditions that allow processes occurring
outside the model domain to influence their predictions. The entire system, consisting of
meteorological model, emissions processor, and output processors shown in Figure 2-9
constitutes the framework of EPA's Models-3.
2-19
-------
Because of the large number of chemical species and reactions that are involved in the
oxidation of realistic mixtures of anthropogenic and biogenic hydrocarbons, condensed
mechanisms must be used in atmospheric models. These mechanisms are tested by comparison
with smog chamber data. However, the existing chemical mechanisms often neglect many
important processes such as the formation and subsequent reactions of long-lived carbonyl
compounds, the incorporation of the most recent information about intermediate compounds, and
heterogeneous reactions involving cloud droplets and aerosol particles.
Emissions inventories are compiled for O3 precursors ( NOX, VOCs, and CO). Recent
estimates and more detailed discussions of the estimates are given in Annex AX2.5.2.
Anthropogenic NOX emissions are associated with combustion processes. Most emissions are in
the form of NO, which is formed at high combustion temperatures from atmospheric nitrogen
and oxygen and from fuel nitrogen. The two largest sources of NOX are electric power
generation plants and motor vehicles. Emissions of NOX therefore are highest in areas having a
high density of power plants and in urban regions having high traffic density. Natural NOX
sources include stratospheric intrusions, lightning, soils, and wildfires. Lightning, fertilized
soils, and wildfires are the major natural sources of NOX in the United States. Both nitrifying
and denitrifying organisms in the soil can produce NOX, mainly in the form of NO. Emission
rates depend mainly on fertilization levels and soil temperature and moisture. Spatial and
temporal variability in soil NOX emissions leads to considerable uncertainty in emissions
estimates. Nationwide, about 60% of lightning generated NOX occurs in the southern
United States and about 60% the total NOX emitted by soils occurs in the central corn belt of
the United States. The oxidation of NH3 emitted mainly by livestock and soils, leads to the
formation of a small amount of NO. Uncertainties in natural NOX inventories are much larger
than for anthropogenic NOX emissions.
Hundreds of VOCs, containing mainly two to about twelve carbon atoms, are emitted by
evaporation and combustion processes from a large number of anthropogenic sources. The two
largest anthropogenic source categories in the U.S. EPA's emissions inventories are industrial
processes and transportation. Emissions of VOCs from highway vehicles account for roughly
two-thirds of the transportation-related emissions.
The accuracy of VOC emission estimates is difficult to determine, both for stationary and
mobile sources. Evaporative emissions, which depend on temperature and other environmental
2-20
-------
factors, compound the difficulties of assigning accurate emission factors. In assigning VOC
emission estimates to the mobile source category, models are used that incorporate numerous
input parameters (e.g., type of fuel used, type of emission controls, age of vehicle), each of
which has some degree of uncertainty. Data for the ratio of CO to NOX and NMHC to NOX in
traffic tunnels (e.g., Pierson et al., 1990) indicated that emissions of NMHCs and CO from motor
vehicles have been underestimated by as much as a factor of two (based on the assumption that
emissions of NOX were reasonably well represented in the inventories). However, the results of
more recent studies have been mixed, with many studies showing agreement to within ±50%, as
summarized in Air Quality Criteria for Carbon Monoxide (U.S. Environmental Protection
Agency, 2000). Remote sensing data (Stedman et al., 1991) indicate that about 50% of NMHC
and CO emissions are produced by about 10% of the vehicles. These "super-emitters" are
typically poorly maintained. Vehicles of any age engaged in off-cycle operations (e.g., rapid
accelerations) emit much more than if operated in normal driving modes.
Vegetation emits significant quantities of VOCs, such as terpenoid compounds (isoprene,
2-methyl-3-buten-2-ol, monoterpenes), compounds in the hexanal family, alkenes, aldehydes,
organic acids, alcohols, ketones, and alkanes. The major chemicals emitted by plants are
isoprene (35%), 19 other terpenoid compounds and 17 non-terpenoid compounds including
oxygenated compounds (40%) (Guenther et al., 2000). Coniferous forests represent the largest
source on a nationwide basis, because of their extensive land coverage. Most biogenic emissions
occur during the summer, because of their dependence on temperature and incident sunlight.
Biogenic emissions are also higher in southern states than in northern states for these reasons and
because of species variations. The uncertainty in natural emissions is about 50% for isoprene
under midday summer conditions and could be as much as a factor often higher for some
compounds (Guenther et al., 2000). Uncertainties in both biogenic and anthropogenic VOC
emission inventories prevent determination of the relative contributions of these two categories
at least in many urban areas. On the regional and global scales, emissions of VOCs from
vegetation are much larger than those from anthropogenic sources.
The performance of CTMs must be evaluated by comparison with field data as part of a
cycle of model evaluations and subsequent improvements. Evaluations of the CMAQ are given
in Arnold et al. (2003) and Fuentes and Raftery (2005). Discrepancies between model
predictions and observations can be used to point out gaps in current understanding of
2-21
-------
atmospheric chemistry and to spur improvements in parameterizations of atmospheric chemical
and physical processes. Model evaluation does not merely involve a straightforward comparison
between model predictions and the concentration field of the pollutant of interest. Such
comparisons may not be meaningful because it is difficult to determine if agreement between
model predictions and observations truly represents an accurate treatment of physical and
chemical processes in the CTM or the effects of compensating errors in complex model routines.
Ideally, each of the model components (emissions inventories, chemical mechanism,
meteorological driver) should be evaluated individually. However, this is rarely done in
practice. A comparison between free radical concentrations predicted by parameterized
chemical mechanisms and observations suggests that radical concentrations were overestimated
by current chemical mechanisms for NOX concentrations <~5 ppb (Volz-Thomas et al., 2003).
In addition to comparisons between concentrations of calculated and measured species,
comparisons of correlations between measured primary VOCs and NOX and modeled VOCs
and NOX are especially useful for evaluating results from chemistry-transport models. Likewise,
comparisons of correlations between measured species and modeled species can be used to
provide information about the chemical state of the atmosphere and to evaluate model
representations (including O3 production per NOX, O3-NOX-VOC sensitivity, and the general
accuracy of photochemical representations). A CTM that demonstrates the accuracy of both its
computed VOC and NOX in comparison with ambient measurements and the spatial and temporal
relations among the critical secondary species associated with O3 has a higher probability of
representing O3-precursor relations correctly than one that does not.
2.6 TECHNIQUES FOR MEASURING OZONE AND ITS PRECURSORS
Several techniques have been developed for sampling and measurement of O3 in the
ambient atmosphere at ground level. Although the chemiluminescence method (CLM) using
ethylene is designated as the Federal Reference Method (FRM) for measuring O3, monitoring in
the NAMS/SLAMS networks is conducted mainly with UV absorption spectrometry using
commercial short path instruments. The primary reference standard instrument is a relatively
long-path UV absorption spectrometer maintained under carefully controlled conditions atNIST
(e.g., Fried and Hodgeson, 1982). Episodic measurements are made with a variety of other
2-22
-------
techniques based on the principles of chemiluminescence, electrochemistry, differential optical
absorption spectroscopy (DOAS), and LIDAR.
In principle, each of these methods is subject to interference. Kleindienst et al. (1993)
found that water vapor could cause a positive interference in the CLM with an average positive
deviation of 3% ozone/% water vapor at 25 °C. However, they also noted that water vapor could
cause positive interferences of up to 9% at high humidities (dew point of 24 °C). The UV
absorption spectrometers are subject to positive interference by atmospheric constituents, such as
certain aromatic aldehydes that absorb at the 253.7 nm Hg resonance line and are at least
partially removed by the MnO2 scrubber. Parrish and Fehsenfeld (2000) did not find any
evidence for significant interference (>1%) in flights through the Nashville urban plume. The
same group tested the air of Houston, El Paso, Nashville, Los Angeles, San Francisco and the
East Coast. They observed only one instance of substantive positive interference defined as the
UV absorption technique showing more than a few ppb more than the CLM. This occurred in
Laporte, TX under heavily polluted conditions and a low inversion, at night (Jobson et al., 2004).
Leston et al. (2005) observed interference of from 20 to 40 ppb in Mexico City and in a separate
smog chamber study. However, the concentrations of relevant compounds were many times
higher than found in U.S. urban areas. Thus, it is not likely that such interference could be more
than a few ppb under typical ambient conditions. However, Leston et al. (2005) suggested that
the use of other materials in the scrubber could have eliminated the interference seen in their
smog chamber study.
By far, most measurements of NO are made using the CLM, based on its reaction with O3.
Commercial instruments for measuring NO and NO2 are constructed with an internal converter
for reducing NO2 to NO and then measuring NO by the CLM. In principle, this technique yields
a measurement of NOX with NO2 found by difference between NOX and NO; but, these
converters also reduce NOZ compounds thereby introducing a positive interference in the
measurement of NO2. Other methods for measuring NO2 are available, such as photolytic
reduction followed by CLM, laser-induced fluorescence and DOAS. However, they require
further development before they can be used for routine monitoring in the NAMS/SLAMS
networks. More detailed descriptions of the issues and techniques discussed above and
techniques for measuring HNO3 and VOCs can be found in Annex AX2.6.
2-23
-------
2.7 SUMMARY
Ozone is formed by atmospheric reactions involving two classes of precursor compounds,
volatile organic compounds (VOCs) and nitrogen oxides (NOX). Ozone is thus a secondary
pollutant. Ozone is ubiquitous throughout the atmosphere; it is present even in remote areas of
the globe. The photochemical oxidation of almost all anthropogenic and biogenic VOCs is
initiated by reaction with hydroxyl (OH) radicals. At night, when they are most abundant, NO3
radicals also oxidize alkenes. In coastal and other select environments, Cl and Br radicals can
also initiate the oxidation of VOCs.
In urban areas, basically all classes of VOCs (alkanes, alkenes, aromatic hydrocarbons,
carbonyl compounds, etc.) and CO are important for O3 formation. Although knowledge of the
oxidative mechanisms of VOCs has improved over the past several years, gaps in knowledge
involving key classes, such as aromatic hydrocarbons, still remain. For example, only about half
of the carbon initially present in aromatic hydrocarbons in smog chamber studies form
compounds that have been identified.
In addition to gas phase reactions, reactions also occur on the surfaces of, or within cloud
droplets and airborne particles. Most of the well-established multiphase reactions tend to reduce
the rate of O3 formation in polluted environments. Reactions of Cl and Br containing radicals
deplete O3 in selected environments such as the Arctic during spring, the tropical marine
boundary layer and inland salt lakes. Direct reactions of O3 with atmospheric particles appear to
be too slow to reduce O3 formation significantly at typical ambient PM levels.
Our basic understanding of the meteorological processes associated with summertime O3
episodes has not changed over the past several years. However, the realization that long-range
transport processes are important for determining O3 concentrations at the surface is growing.
In addition to synoptic scale flow fields, nocturnal low-level jets are capable of transporting
pollutants hundreds of km from their sources in either the upper boundary layer or the lower free
troposphere. Turbulence then brings O3 and other pollutants to the surface. On larger scales,
important progress has been made in identifying the mechanisms of intercontinental transport
of O3 and other pollutants.
Some O3 would be found near the earth's surface as the result of its downward transport
from the stratosphere, even in the absence of photochemical reactions in the troposphere.
Intrusions of stratospheric O3 that reach the surface are rare. Much more common are intrusions
2-24
-------
that penetrate to the middle and upper troposphere. However, O3 transported to the middle and
upper troposphere can still affect surface concentrations through various mechanisms that mix
air between the planetary boundary layer and the free troposphere above.
The formation of O3 and associated compounds is a complex, nonlinear function of many
factors, including the intensity and spectral distribution of sunlight; atmospheric mixing and
other atmospheric processes; and the concentrations of the precursors in ambient air. At lower
NOX concentrations found in most environments, ranging from remote continental areas to rural
and suburban areas downwind of urban centers, the net production of O3 increases with
increasing NOX. At higher concentrations found in downtown metropolitan areas, especially
near busy streets and highways and in power plant plumes, there is net destruction of O3 by
reaction with NO. In between these two regimes, there is a transition stage in which O3
production shows only a weak dependence on NOX concentrations. The efficiency of O3
production per NOX oxidized is generally highest in areas where NOX concentrations are lowest
and decrease with increasing NOX concentration.
Chemistry transport models are used to improve understanding of atmospheric chemical
and physical processes as well as to develop air pollution control strategies. The performance of
these models must be evaluated by comparison with field data as part of a cycle of model
evaluations and subsequent improvements. Discrepancies between model predictions and
observations can be used to point out gaps in current understanding and thus to improve
parameterizations of atmospheric chemical and physical processes. Model evaluation does not
merely involve a straightforward comparison between model predictions and the concentration
fields of a pollutant of interest (e.g., O3). Such comparisons may not be meaningful because it is
difficult to determine if agreement between measurements and model predictions truly represents
an accurate treatment of physical and chemical processes in the model or the effects of
compensating errors in model routines.
The main methods currently in use for routine monitoring of ambient O3 are based on
chemiluminescence or UV absorption. Measurements at most ambient monitoring sites are
based on UV absorption. Both of these methods are subject to interference by other atmospheric
components. One study found large positive interference in Mexico City and in a smog
chamber, but a few studies conducted in urban plumes did not find significant positive
interference in the UV absorption technique. Sufficient new information is not available to
2-25
-------
amend estimates of the accuracy and precision of O3 monitors. Such a reevaluation requires
studies of the simultaneous effects of a number of potential interferants including water vapor,
organic compounds, and temperature on the UV photometric and chemiluminescent methods.
2-26
-------
REFERENCES
Arnold, J. R.; Dennis, R. L.; Tonnesen, G. S. (2003) Diagnostic evaluation of numerical air quality models with
specialized ambient observations: testing the Community Multiscale Air Quality modeling system (CMAQ) at
selected SOS 95 ground sites. Atmos. Environ. 37: 1185-1198.
Berkowitz, C. M; Shaw, W. J. (1997) Airborne measurements of boundary layer chemistry during the Southern
Oxidant Study: a case study. J. Geophys. Res. [Atmos.] 102: 12,795-12,804.
Berkowitz, C. M.; Fast, J. D.; Sprinston, S. R.; Larsen, R. J.; Spicer, C. W.; Doskey, P. V.; Hubbe, J. M.;
Plastridge, R. (1998) Formation mechanisms and chemical characteristics of elevated photochemical layers
over the northeast United States. J. Geophys. Res. [Atmos.] 103: 10,631-10,647.
Blumenthal, D. L.; Lurmann, F. W.; Kumar, N.; Dye, T. S.; Ray, S. E.; Korc, M. E.; Londergan, R.; Moore, G.
(1997) Transport and mixing phenomena related to ozone exceedances in the northeast U.S. (analysis based
on NARSTO-northeast data). Available: http://capita.wustl.edu/otag/reports/otagrept/otagrept.html
(30 October 2003).
Bonn, B.; Von Kuhlmann, R.; Lawrence, M. G. (2004) High contribution of biogenic hydroperoxides to secondary
organic aerosol formation. Geophys. Res. Lett. 31: L10108, 10.1029/2003GL019172.
Bonner, W. D. (1968) Climatology of the low level jet. Mon. Weather Rev. 96: 833-850.
Carter, W. P. L. (1995) Computer modeling of environmental chamber studies of maximum incremental reactivities
of volatile organic compounds. Atmos. Environ. 29: 2513.
Civerolo, K. L.; Mao, H. T.; Rao, S. T. (2003) The airshed for ozone and fine paniculate pollution in the eastern
United States. Pure Appl. Geophys. 160: 81-105.
Corsmeier, U.; Kalthhoff, N.; Kolle, O.; Motzian, M.; Fiedler, F. (1997) Ozone concentration jump in the stable
nocturnal boundary layer during a LLJ-event. Atmos. Environ. 31: 1977-1989.
Docherty, K. S.; Wu, W.; Lim, Y. B.; Ziemann, P. J. (2005) Contributions of organic peroxides to secondary aerosol
formed from reactions of monoterpenes with O3. Environ. Sci. Technol. 39: 4049-4059.
Eisele, F. L.; Mount, G. H.; Tanner, D.; Jefferson, A.; Shelter, R.; Harder, J. W.; Williams, E. J. (1997)
Understanding the production and interconversion of the hydroxyl radical during the tropospheric OH
photochemistry experiment. J. Geophys. Res. 102: 6457-6465.
Fehsenfeld, F. C.; Trainer, M.; Parrish, D. D.; Volz-Thomas, A.; Penkett, S. (1996) North Atlantic Regional
Experiment (NARE) 1993 summer intensive: foreword. J. Geophys. Res. [Atmos.] 101: 28,869-28,875.
Fried, A.; Hodgeson, J. (1982) Laser photoacoustic detection of nitrogen dioxide in the gas-phase titration of nitric
oxide with ozone. Anal. Chem. 54: 278-282.
Fuentes, M.; Raftery, A. E. (2005) Model evaluation and spatial interpolation by Bayesian combination of
observations with outputs from numerical models. Biometrics 61: 36-45.
Guenther, A.; Geron, C.; Pierce, T.; Lamb, B.; Harley, P.; Fall, R. (2000) Natural emissions of non-methane volatile
organic compounds, carbon monoxide, and oxides of nitrogen from North America. Atmos. Environ.
34: 2205-2230.
Hameed, S.; Pinto, J. P.; Stewart, R. W. (1979) Sensitivity of the predicted CO-OH-CH4 perturbation to tropospheric
NO-j concentrations. J. Geophys. Res. C: Oceans Atmos. 84: 763-768.
Husar, R. B.; Renard, W. P. (1998) Ozone as a function of local wind speed and direction: Evidence of local and
regional transport. Presented at: 91st annual meeting and exhibition of the Air & Waste Management
Association; June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association; online paper no.
98-A922. Available: http://capita.wustl.edu/capita/CapitaReports/REPORTSl.HTML (13 November 2003).
Jacob, D. J.; Horowitz, L. W.; Munger, J. W.; Heikes, B. G.; Dickerson, R. R.; Artz, R. S.; Keene, W. C. (1995)
Seasonal transition from NOX- to hydrocarbon-limited conditions for ozone production over the eastern
United States in September. J. Geophys. Res. 100: 9315-9324.
Jaegle, L.; Jacob, D. J.; Brune, W. H.; Wennberg, P. O. (2001) Chemistry of HOx radicals in the upper troposphere.
Atmos. Environ. 35: 469-489.
Jobson, B. T.; Berkowitz, C. M.; Kuster, W. C.; Goldan, P. D.; Williams, E. J.; Fehsenfeld, F. C.; Apel, E. C.;
Karl, T.; Lonneman, W. A.; Riemer, D. (2004) Hydrocarbon source signatures in Houston, Texas: influence
of the petrochemical industry. J. Geophys. Res. 109: D24305: 10.1029/2004 JD004887.
Johnson, D.; Jenkin, M. E.; Wirtz, K.; Martin-Riviejo, M. (2004) Simulating the formation of secondary organic
aerosol from the photooxidation of toluene. Environ. Chem. 1: 150-165.
Kasibhatla, P.; Chameides, W. L. (2000) Seasonal modeling of regional ozone pollution in the eastern United States.
Geophys. Res. Lett. 27: 1415-1418.
2-27
-------
Kleindienst, T. E.; Hudgens, E. E.; Smith, D. F.; McElroy, F. F.; Bufalini, J. J. (1993) Comparison of
chemiluminescence and ultraviolet ozone monitor responses in the presence of humidity and photochemical
pollutants. Air Waste 43: 213-222.
Leston, A. R.; Ollinson, W. M; Spicer, C. W.; Satola, J. (2005) Potential interference bias in ozone standard
compliance monitoring. J. Air Waste Manage. Assoc. 55: 1464-1472.
Milford, J. B.; Gao, D.; Sillman, S.; Blossey, P.; Russell, A. G. (1994) Total reactive nitrogen (NOy) as an indicator
of the sensitivity of ozone to reductions in hydrocarbon and NOX emissions. J. Geophys. Res. 99: 3533-3542.
National Research Council. (1991) Rethinking the ozone problem in urban and regional air pollution. Washington,
DC: National Academy Press. Available: http://www.nap.edu/books/0309046319/html/ [26 March, 2004].
Nejgaard, J. K.; Christensen, K. B.; Wolkoff, P. (2005) The effect on human eye blink frequency of exposure to
limonene oxidation products and methacrolein. Toxicol. Lett. 156: 241-251.
Olszyna, K. J.; Bailey, E. M.; Simonaitis, R.; Meagher, J. F. (1994) O3 and NOy relationships at a rural site. J.
Geophys. Res. [Atmos.] 99: 14,557-14,563.
Parrish, D. D.; Fehsenfeld, F. C. (2000) Methods for gas-phase measurements of ozone, ozone precursors and
aerosol precursors. Atmos. Environ. 34: 1921-1957.
Pierson, W. R.; Gertler, A. W.; Bradow, R. L. (1990) Comparison of the SCAQS tunnel study with historical data.
Presented at: 83rd annual meeting & exhibition of the Air and Waste Management Association; June;
Pittsburgh, PA. Pittsburgh, PA: Air and Waste Management Association; paper no. 90-175.3.
Piety, C. A. (2005) The relation between daily maximum ozone and daily maximum temperature [memorandum to
Dr. Joseph Pinto]. Research Triangle Park, NC: U.S. Environmental Protection Agency; July 18.
Pinto, J. P.; Bruhl, C.; Thompson, A. M. (1993) The current and future envirionmental role of atmospheric methane.
In: Khalil, M. A. K., ed. Atmospheric methane sources, sinks, and role in global change, p. 514-531. (NATO
ASI Series, v. 113).
Poppe, D.; Wallasch, M.; Zimmermann, J. (1993) The dependence of the concentration of OH on its precursors
under moderately polluted conditions: a model study. J. Atmos. Chem. 16: 61-78.
Rao, S. T.; Ku, J.-Y.; Berman, S.; Zhang, K.; Mao, H. (2003) Summertime characteristics of the atmospheric
boundary layer and relationships to ozone levels over the eastern United States. Pure Appl. Geophys.
160: 21-55.
Rohr, A. C.; Wilkins, C. K.; Clausen, P. A.; Hammer, M.; Nielsen, G. D.; Wolkoff, P.; Spengler, J. D. (2002) Upper
airway and pulmonary effects of oxidation products of (+)-a-pinene, J-limonene, and isoprene in B ALB/c
mice. Inhalation Toxicol. 14: 663-684.
Ryerson, T. B.; Buhr, M. P.; Frost, G. J.; Goldan, P. D.; Holloway, J. S.; Htibler, G.; Jobson, B. T.; Kuster, W. C.;
McKeen, S. A.; Parrish, D. D.; Roberts, J. M.; Sueper, D. T.; Trainer, M.; Williams, J.; Fehsenfeld, F. C.
(1998) Emissions lifetimes and ozone formation in power plant plumes. J. Geophys. Res.
103(D17): 22,569-22,583.
Ryerson, T. B.; Trainer, M.; Holloway, J. S.; Parrish, D. D.; Huey, L. G.; Sueper, D. T.; Frost, G. J.; Donnelly, S.
G.; Schauffler, S.; Atlas, E. L.; Kuster, W. C.; Goldan, P. D.; Htibler, G.; Meagher, J. F.; Fehsenfeld, F. C.
(2001) Observations of ozone formation in power plant plumes and implications for ozone control strategies.
Science (Washington, DC) 292: 719-723.
Sakugawa, H.; Kaplan, I. R. (1989) H2O2 and O3 in the atmosphere of Los Angeles and its vicinity: factors
controlling their formation and their role as oxidants of SO2. J. Geophys. Res. [Atmos.] 94: 12,957-12,973.
Schichtel, B. A.; Husar, R. B. (2001) Eastern North American transport climatology during high- and low-ozone
days. Atmos. Environ. 35: 1029-1038.
Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change. New
York, NY: John Wiley & Sons, Inc.
Shapiro, M. A. (1980) Turbulent mixing within tropopause folds as a mechanism for the exchange of chemical
constituents between the stratosphere and troposphere. J. Atmos. Sci. 37: 994-1004.
Sillman, S. (1995) The use of NOy, H2O2 and HNO3 as indicators for ozone-NOx-hydrocarbon sensitivity in urban
locations. J. Geophys. Res. 100: 14,175-14,188.
Sillman, S.; He, D.-Y. (2002) Some theoretical results concerning O3-NOX-VOC chemistry and NOX-VOC
indicators. J. Geophys. Res. (Atmos.) 107: 10.1029/2001JDOO1123.
Sillman, S.; He, D.; Cardelino, C.; Imhoff, R. E. (1997) The use of photochemical indicators to evaluate
ozone-NOx-hydrocarbon sensitivity: case studies from Atlanta, New York, and Los Angeles. J. Air Waste
Manage. Assoc. 47: 1030-1040.
2-28
-------
Sillman, S.; He, D.; Pippin, M. R.; Daum, P. H.; Imre, D. G.; Kleimnan, L. I.; Lee, J. H.; Weinstein-Lloyd, J. (1998)
Model correlations for ozone, reactive nitrogen, and peroxides for Nashville in comparison with
measurements: implications for O3-NOx-hydrocarbon chemistry. J. Geophys. Res. [Atmos.]
103: 22,629-22,644.
Stedman, D. H.; Bishop, G.; Peterson, J. E.; Guenther, P. L. (1991) On-road CO remote sensing in the Los Angeles
Basin: final report. Sacramento, CA: California Air Resources Board, ARE Contract No. A932-189.
Stall, R. B. (1999) An introduction to boundary layer meteorology. Dordrecht, The Netherlands: Kluwer Academic
Publishers; pp. 9-16, 499-506.
Stall, R. B. (2000) Meteorology for scientists and engineers: a technical companion book with Ahrens' Meteorology
Today. 2nd ed. Pacific Grove, CA: Brooks/Cole.
Taubman, B. F.; Marufu, L. T.; Piety, C. A.; Doddridge, E.G.; Stehr, J. W.; Dickerson, R. R. (2004) Airborne
characterization of the chemical, optical, and meteorological properties, and origins of a combined
ozone-haze episode over the eastern United States. J. Atmos. Sci. 61: 1781-1793.
Taubman, B. F.; Mains, J. C.; Thompson, A. M.; Marufu, L. T.; Doddridge, B. G.; Stehr, J. W.; Piety, C. A.;
Dickerson, R. R. (2005) Aircraft vertical profiles of trace gas and aerosol pollution over the mid-Atlantic
U.S.: statistics and meteorological cluster analysis. J. Geophys. Res. (Atmos.): submitted.
Trainer, M.; Parrish, D. D.; Buhr, M. P.; Norton, R. B.; Fehsenfeld, F. C.; Anlauf, K. G.; Bottenheim, J. W.; Tang,
Y. Z.; Wiebe, H. A.; Roberts, J. M.; Tanner, R. L.; Newman, L.; Bowersox, V. C.; Meagher, J. F.; Olszyna,
K. J.; Rodgers, M. O.; Wang, T.; Berresheim, H.; Demerjian, K. L.; Roychowdhury, U. K. (1993) Correlation
of ozone with NOy in photochemically aged air. J. Geophys. Res. [Atmos.] 98: 2917-2925.
U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
http://cfpub2.epa.gov/ncea/.
U.S. Environmental Protection Agency. (2000) Air quality criteria for carbon monoxide. Research Triangle Park,
NC: National Center for Environmental Assessment; report no. EPA/600/P-99/00 IF. Available:
http://www.epa.gov/ncea/pdfs/coaqcd.pdf (7 May 2003).
U.S. Environmental Protection Agency. (2004) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
http://cfpub.epa.gov/ncea/ [9 November, 2004].
Volz-Thomas, A.; Geiss, H.; Hofzumahaus, A.; Becker, K.-H. (2003) Introduction to special section: photochemistry
experiment in BERLIOZ. J. Geophys. Res. [Atmos.] 108(D4): 10.1029/JD002029.
Weschler, C. J.; Shields, H. C. (1997) Potential reactions among indoor pollutants. Atmos. Environ. 31: 3487-3495.
Weschler, C. J.; Shields, H. C. (2000) The influence of ventilation on reactions among indoor pollutants: modeling
and experimental observations. Indoor Air. 10: 92-100.
Weschler, C. J.; Hodgson, A. T.; Wooley, J. D. (1992) Indoor chemistry: ozone, volatile organic compounds, and
carpets. Environ. Sci. Technol. 26: 2371-2377.
Wilkins, C. K.; Clausen, P. A.; Wolkoff, P.; Larsen, S. T.; Hammer, M.; Larsen, K.; Hansen, V.; Nielsen, G. D.
(2001) Formation of strong irritants in mixtures of isoprene/ozone and isoprene/ozone/nitrogen dioxide.
Environ. Health Perspect. 109: 937-941.
Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Hougaard, K. S.; Nielsen, G. D. (1999) Formation of strong airway
irritants in a model mixture of (+)-a-pinene/ozone. Atmos. Environ. 33: 693-698.
Zimmermann, J.; Poppe, D. (1993) Nonlinear chemical couplings in the tropospheric NOX-HOX gas phase chemistry.
J. Atmos. Chem. 17: 141-155.
2-29
-------
3. ENVIRONMENTAL CONCENTRATIONS,
PATTERNS, AND EXPOSURE ESTIMATES
3.1 INTRODUCTION
Identification and Use of Existing Air Quality Data
Topics discussed in this chapter include the characterization of ambient air quality data for
ozone (O3), the uses of these data in assessing the exposure of vegetation to O3, concentrations
of O3 in microenvironments, and a discussion of the currently available human exposure data and
exposure model development. The information contained in this chapter pertaining to ambient
concentrations is taken primarily from the U.S. Environmental Protection Agency (EPA)
Air Quality System (AQS; formerly the AIRS database). The AQS contains readily accessible
detailed, hourly data that has been subject to EPA quality control and assurance procedures.
Data available in AQS were collected from 1979 to 2001. As discussed in the 1996 O3 Air
Quality Criteria Document or AQCD (U.S. Environmental Protection Agency, 1996), the data
available prior to 1979 may be unreliable due to calibration problems and other uncertainties.
As noted in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996), O3 is the
only photochemical oxidant other than nitrogen dioxide (NO2) that is routinely monitored and
for which a comprehensive database exists. Data for peroxyacetyl nitrate (PAN), hydrogen
peroxide (H2O2), and other oxidants either in the gas phase or particle phase typically have been
obtained only as part of special field studies. Consequently, no data on nationwide patterns of
occurrence are available for these non-O3 oxidants; nor are extensive data available on the
relationships of levels and patterns of these oxidants to those of O3. However, available data for
gas phase and particle phase oxidants are discussed in this chapter.
Characterizing Ambient Ozone Concentrations
The "concentration" of a specific air pollutant is typically defined as the amount (mass)
of that material per unit volume of air. However, most of the data presented in this chapter are
expressed as "mixing ratios" in terms of a volume-to-volume ratio (parts per million [ppm] or
parts per billion [ppb]). Data expressed this way are often referred to as concentrations, both in
5-1
-------
the literature and in the text, following common usage. Human exposures are expressed in units
of mixing ratio times time.
Several different air quality metrics have been suggested for evaluating exposures of
vegetation to O3. The peak-weighted, cumulative exposure indicators used in this chapter for
characterizing vegetation exposures are SUM06 and SUM08 (the sums of all hourly average
concentrations >0.06 and 0.08 ppm, respectively) and W126 (the sum of the hourly average
concentrations that have been weighted according to a sigmoid function that is based on a
hypothetical vegetation response [see Lefohn and Runeckles, 1987]). Further discussion of these
exposure indices is presented in Chapter 9.
The EPA has established "ozone seasons" during which measurement of ambient O3
concentrations for different locations within the United States and the U.S. territories is required
(CFR, 2000). Table AX3-1 shows the O3 seasons during which continuous, hourly averaged O3
concentrations must be monitored. Monitoring is optional outside of these O3 seasons and
indeed is conducted during the winter in a number of areas.
Data for O3 in ambient air across the United States are summarized in Section 3.2. The
data are summarized for urban, rural, and relatively remote sites. Relatively remote monitoring
sites (RRMS) are sites that are not strongly influenced by nearby pollution sources and are
located mainly in national parks in the West. However, this does not mean that they are free of
the effects of regional or local pollution, especially during tourist seasons. Data for the spatial
variability of O3 within urban areas are summarized in Section 3.3. Data for the diurnal and
seasonal variability of O3 concentrations are given in Section 3.4. The long term temporal
variability of O3 concentrations is discussed in Section 3.5. Relationships among O3 and other
species are discussed in Section 3.6. Information about the occurrence of other oxidants and
their relationship to O3 is given in this section. A discussion of Policy Relevant Background
(PRB) O3 concentrations is presented in Section 3.7. PRB O3 concentrations are background O3
concentrations that would be observed in the U.S. in the absence of anthropogenic emissions
of O3 precursors in the U.S., Canada, and Mexico. Background levels so defined help facilitate
the distinction between pollution levels that can be controlled by U.S. regulations (or through
international agreements with neighboring countries) from levels that are generally
uncontrollable by the U.S. Indoor sources and emissions of O3 are discussed in Section 3.8.
5-2
-------
Issues related to evaluating human exposure to O3 are summarized in Section 3.9. Finally,
a summary of key points in Chapter 3 is given in Section 3.10.
3.2 AMBIENT AIR QUALITY DATA FOR OZONE
Ozone Air Quality at Urban, Suburban, andNonurban Sites
Figure 3-1 shows the mean daily maximum 8-h O3 concentrations, and Figure 3-2 shows
the 95th percentile values of the daily maximum 8-h O3 concentrations, based on county wide
site-wise averages across the United States from May to September 2000 to 2004. The period
from May to September was chosen because, although O3 is monitored for different lengths of
time across the country, all O3 monitors should be operational during these months. Data
flagged because of quality control issues were removed with concurrence of the local monitoring
agency. Only days with data for 18 of 24 hours were kept, and a minimum of 115 of 153 days
were required in each year. Cut points for the tertile distributions on each map were chosen at
the median and 95th percentile values. These cut points were chosen because they represent
standard metrics for characterizing important aspects of human exposure used by the EPA. Any
other percentiles or statistics that are believed to be helpful for characterizing human exposures
could also be used. Blank areas on the maps indicate no data coverage. It should be noted that
county areas can be much larger in the West than in the East, but monitors are not spread evenly
within a county. As a result, the assigned concentration range might not represent conditions
throughout a particular county and, so, large areas in western counties where there are no
monitors were blanked out.
As shown in Figure 3-1, the median of the county wide, mean daily maximum 8-h O3
concentration across the United States is 49 ppb, and 5% of these site-wise means exceeded
57 ppb. Though the median and 95th percentile values are fairly close, these results cannot be
taken to imply that average O3 concentrations lie in a relatively narrow range throughout the
United States, because data coverage is not as complete in the West as it is in the East. High
mean daily maximum 8-h O3 concentrations are found in California and states in the Southwest
as well as in several counties in the East. As shown in Figure 3-2, the nationwide median of the
countywide, 95th percentile value of the daily maximum 8-h O3 concentration is 73 ppb and 5%
of these values are above 85 ppb. High values for the 95th percentiles are found in California,
-------
Seasonal (May-September) Mean of Daily Maximum 8-Hour Values, 2002-2004
Concentration PPM
X < 0.049
0.049 < X < 0.057
0.057 < X
Figure 3-1. Countywide mean daily maximum 8-h O3 concentrations, May to September
2000 to 2004.
Source: Fitz-Simons et al. (2005).
Texas, and some counties in the East, but not necessarily in the same counties in the East as
shown for the mean daily maximum 8-h concentrations in Figure 3-1.
Although mean O3 concentrations in Houston, TX were below the nationwide median, its
95th percentile value ranks in the highest 5% nationwide. Conversely, mean O3 concentrations
in southwestern states are among the highest in the United States, but values at the upper end of
the distribution (e.g., the 95th percentile value) in these states are not among the highest peak
values in the United States. In other areas where the highest mean O3 concentrations occurred
(e.g., California; Dallas-Fort Worth, TX; and the Northeast Corridor), the highest peak values
were also observed.
5-4
-------
Seasonal (May-September) 95th Percentile of Daily Maximum 8-Hour Values, 2002-2004
Concentration PPM
X < 0.073
0.073 < X < 0.085
0.085 < X
Figure 3-2. Countywide 95th percentile value of daily maximum 8-h O3 concentrations,
May to September 2000 to 2004.
Source: Fitz-Simons et al. (2005).
Although countywide averages are shown, it should be noted that considerable spatial
variability can exist within a county, especially within urban areas as described in Section 3.3.
In addition, there can also be differences in the diurnal profile of O3 among monitors within
counties.
Box plots showing the distribution of nationwide O3 concentrations for different averaging
periods (1-h daily maximum, 8-h daily maximum, and 24-h daily average) are given in Figures
AX3-4 to AX3-6 and numerical values are given in Table AX3-2. The differences between the
50th and 95th percentile values indicate the range of O3 levels between "typical" O3 days and
"high" O3 days. These differences are approximately 40, 30, and 25 ppb for the daily 1-h and
5-5
-------
8-h daily maxima and 24-h average O3 concentrations, respectively. As might be expected, the
daily maximum 1-h and 8-h O3 concentrations are highly correlated.
Lehman et al. (2004) divided the eastern United States into five regions, each of which
exhibit relatively distinct spatial and temporal patterns of O3 concentrations at nonurban sites.
Only sites classified as being rural or suburban and with land usage of forest, agriculture, or
residential were included in the analyses. These criteria were chosen to avoid sites where O3 is
scavenged by NO, which can be found in high concentrations near major sources such as traffic
in urban cores. The five regions, shown in Figure 3-3, are characterized by different patterns
of O3 properties such as temporal persistence and seasonal variability. Figure 3-3 shows
nonurban, monthly average, daily maximum 8-h O3 concentrations in the five regions in the
eastern United States from April to October 1993 to 2002.
Regional differences are immediately apparent. Highest concentrations among all the
regions are generally found in the Mid-Atlantic region (mean of 52 ppb), with highest values
throughout the O3 concentration distribution except for the overall maximum. Lowest mean
concentrations (42 ppb) are found in Florida. In the northern regions (the Northeast, Great
Lakes) and the Mid-Atlantic region, highest median and peak concentrations are found in July,
whereas in the Southwest region, highest median concentrations are found in August, with
highest peaks in June and September, i.e., outside the warmest summer months. In Florida,
highest monthly averaged median and peak concentrations are found during the spring. High O3
concentrations tend to be most persistent (3 to 4 days of persistence) in the southern regions, less
persistent in the Mid-Atlantic region (2 to 3 days) and least persistent in the northern regions
(1 or 2 days). Such analyses could not be made for the western United States, in part because of
the difficulty in finding regions with relatively coherent O3 properties as noted above for the
eastern United States.
Box plots showing the distributions of hourly average O3 concentrations for different types
of rural sites for 2004 are given in Figures 3-4a (rural-agricultural), 3-4b (rural-forest), and 3-4c
(rural-residential or commercial). Some associated metrics for vegetation exposures are given in
Figures AX3-8 to AX3-10. Note that high O3 concentrations are found at sites that are classified
as rural, such as Anne Arundel Co., MD; Yosemite NP, CA; and Crestline, CA. Land use
designations do not usually give an accurate picture of exposure regimes in rural areas, because
the land use characterization of "rural" does not imply that a specific location is isolated from
5-6
-------
"
Q.
ri
~
(8
C
O
180-
160-
140-
120'
100-
80-
60.
40-
20-
0-
Great Lakes Region
R
APR
MAY
JUN JUL
Month
AUG
SEP
OCT
180-
160
140-
S 120-
Q.
rt
— 100-
a>
C 80-
| 60.
40-
20-
0-
R
APR
Northeast Region
MAY
JUN JUL
Month
AUG
SEP
OCT
180 :
160-
140.
S 120.
Q.
n
•s 100-
a>
C 80.
O 60.
40-
20.
Southwest Region
R
JUN JUL
Month
AUG
SEP
OCT
0)
C
180-
160-
140-
120-
100-
80.
60.
40-
20.
0-
Mid-Atlantic Region
JUN JUL
Month
AUG
SEP
OCT
180-
160-
140.
2" 120-
o.
•2= 100-
0>
80-
60-
40.
20.
0-
Florida Region
JUN JUL
Month
AUG
SEP
OCT
Figure 3-3. Box plots showing daily maximum 8-h O3 averaged by month over 1993 to
2002 in the five regions in the eastern United States derived by Lehman
et al.(2004). The boxes define the interquartile range and the whiskers, the
extreme values.
Source: Lehman et al. (2004).
5-7
-------
Q.
•3
c
o
c
09
O
o
o
01
c
o
N
O
Q.
•S
c
o
c
0)
o
c
o
o
<1>
o
N
o
0.18-
0.16 -
0.14 -:
0,12-
0.10-
0.08 -:
0.06 •:
0.04 -:
0.02 T
0.00 -
0.18-
0.16-i
0.14-i
0.12-
0.10-i
0,08^
0.06^
0.04-i
0.02 ~i
0,00-
(a) Rural Agriculture
3618
3652
3643
3493
3621
i
3519
3486
T
Effingham, Indianapolis, AnneArundel Omaha. NE
IL IN Co., MD
Beaver
Co., PA
Fauquier Dodge
Co., VA Co., Wl
(b) Rural Forest
3420
2676
3419
3312
Yosernite NP, CA
Whiteface Mtn., NY Smoky Mtns. NP, TN Shenandoah NP, VA
(c) Rural Other (Rural Residential or Rural Commercial)
~ 0.18-
P :
a. 0.16-
Concentration (p
o o o o o
b b '-^ -* ^
Oi 00 O hO 4^
OJ 0.04 -
S 0.02 -
o.oo-
3490 3572 3460 3419
5^
i
T | B |
-A- ! JL
ft
• — *—
i * _•_
i *
i k i
Crestline, CA
Sandoval Co., NM
Guilford Co., NC
Farmville, NC
Figure 3-4a-c. Hourly average O3 concentrations observed at selected (a) rural-
agricultural (b) rural-forested, and (c) rural-residential or commercial
sites for 2004. The whiskers on the box plot represent the 10th and
90th percentile concentrations. The "X"s above and below the whiskers
are the values that fall below and above the 10th and 90th percentile
concentrations. The dots inside the box represent the mean, for the
statistic, at all sites. The number of observations is shown above each
box plot.
Source: Fitz-Simons et al. (2005).
-------
anthropogenic influences. Rather, the characterization refers only to the current use of the land,
not to the presence of sources. Since O3 produced from emissions in urban areas is transported
to more rural downwind locations, elevated O3 concentrations can occur at considerable
distances from urban centers. In addition, major sources of O3 precursors such as power plants
and highways are located in nonurban areas and also produce O3 in these areas. Due to lower
chemical scavenging in nonurban areas, O3 tends to persist longer in nonurban than in urban
areas which tends to lead to higher exposures in nonurban areas influenced by anthropogenic
precursor emissions.
Ozone Air Quality Data at Relatively Remote Monitoring Sites (RRMS)
Relatively Remote Monitoring Sites (RRMS) are sites located in national parks that tend to
be less affected by obvious pollution sources than other sites. However, this does not mean that
they are completely unaffected by local pollution, because of the large number of visitors to the
national parks.
Box plots showing the distribution of annual hourly averaged O3 concentrations at four
RRMS are given in Figures 3-5a-d. It is important to characterize hourly average O3
concentrations at RRMS so that assessments of the possible effects of O3 on human health and
vegetation use concentration ranges that span the range of O3 concentrations found in the U.S.
In many controlled exposure studies examining vegetation, O3 is filtered out of ambient air
before it is admitted into the exposure chambers. As a result, O3 levels of only a few ppb are
used as controls.
As can be seen from Figures 3-5a-d, annual mean values of the daily maximum 8-h O3
concentration have not changed much over the past 10 years of available data. Mean values
typically range from about 0.020 ppm to about 0.040 ppm. Concentrations only rarely exceed
0.080 ppm, in contrast to observations at other "rural" sites shown in Figures 3-4a-c.
It is unlikely that distributions found at sites with low maximum hourly average
concentrations in the western United States could represent those at sites in the eastern and
midwestern United States because of regional differences in sources of precursors and transport
patterns. Given the high density of sources in the eastern and midwestern United States, it is
unclear whether a site could be found in either of these regions that would not be influenced by
the transport of O3 from nearby urban areas. Thus, with the exception of the Voyageurs NP site
5-9
-------
a. Theodore Roosevelt National Park
b. Yellowstone National Park
016- 3525 3162 3629 1808 3649 3651 3652 3672 3199 3667
0.12
0.10
0.00-
I I I I I I I I I I
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
0.16- 3276 3263 3408 3444 3385 3333 3478 3311 3343 3402
0.12-
E
£ 010
0.06-
0.02-
I I I I I I I I I I
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
c. Glacier National Park
0.16- 3365 3317 3169 3277 3232 3230 3320 3311 3157 3333
0.12-
0.04-
0.02-
\ \ \ \ \ \ \ \ \ \
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
d. Olympic National Park
0.16- 3321 3438 3095 3416 3347 3351 3304 3361 3361 3276
g 0.08
0.02-
I I I I I I I I I I
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Figure 3-5a-d. Daily 8-h maximum O3 concentrations observed at selected national park
sites. The whiskers on the box plot represent the 10th and 90th percentile
concentrations. The "X"s above and below the whiskers are the values that
fall below and above the 10th and 90th percentile concentrations. The dots
inside the box represent the mean. The number of observations is shown
above each box plot.
Source: Fitz-Simons et al. (2005).
3-10
-------
in Minnesota, observations at RRMS are limited to those obtained in the western United States.
However, not all national park sites in the West can be considered to be free of strong regional
pollution influences, e.g., Yosemite NP (CA) as shown in Figure 3-4b. Maps showing the
nationwide distribution of various metrics for vegetation exposures are given in Section AX3.2,
Figures AX3-13 to AX3-27.
3.3 SPATIAL VARIABILITY OF OZONE IN URBAN AREAS
The spatial variability in O3 concentrations in 24 MSAs across the United States was
examined. These MSAs were selected to provide (1) information helpful for risk assessments,
(2) a general overview of the spatial variability of O3 in different regions of the country, and
(3) insight into the spatial distribution of O3 in cities where health outcome studies have been
conducted. Statistical analyses of the human health effects of airborne pollutants based on
aggregate population time-series data have often relied on ambient concentrations of pollutants
measured at one or more central sites in a given metropolitan area. In the particular case of
ground-level O3 pollution, central-site monitoring has been justified as a regional measure of
exposure mainly on the grounds that correlations between concentrations at neighboring sites
measured over time are usually high. In MSAs with multiple monitoring sites, averages over the
monitors have often been used to characterize population exposures. However, substantial
differences in concentrations between monitors can exist even though concentrations measured
at the monitoring sites are highly correlated, thus leading to the potential for exposure
misclassification error.
Metrics for characterizing spatial variability include the use of Pearson correlation
coefficients (r), values of the 90th percentile (P90) of the absolute difference in O3 concentrations,
and coefficients of divergence (COD)1. These methods of analysis follow those used for
characterizing PM25 and PM10_25 concentrations in Pinto et al. (2004) and in the latest edition of
1 The COD is defined as follows:
PI,. -. ^2
CODit=l-^ (% *ik
where xv and xik represent the 24-h average PM2 5 concentration for day /' at site y and site k andp is the number of
observations.
3-11
-------
the Particulate Matter Air Quality Criteria Document (PM AQCD) (U.S. Environmental Agency,
2004a). However, the calculations were performed on an hourly basis rather than on a 24-h
basis. Data were aggregated over local O3 seasons, the lengths of which vary from state to state.
In several southwestern states, it lasts all year long. However, it typically last for 6 months
in other areas, such as in New England, the mid-Atlantic states, the Midwest, and the Northwest
(see Table AX3-1).
Table 3-1 shows the chosen urban areas, the range of 24-h average O3 concentrations over
the local O3 season from 1999 to 2001, the range of intersite correlation coefficients, the range of
P90 differences in O3 concentrations between site pairs, and the range in COD values. A COD of
zero implies that values in both data sets are identical, and a COD of one indicates that the two
data sets are uncorrelated, with no matching values from either data set. In general, statistics
were calculated for partial MSAs. This was done so as to obtain reasonable lower estimates of
the spatial variability that is present, as opposed to examining the consolidated MSAs. However,
this could not be readily done for Boston, MA and New York, NY, so statistics were calculated
for those consolidated MSAs. More detailed calculations for a subset of nine MSAs are given in
Figures AX3-28 through AX3-36 in Section AX3.3.
As can be seen from Table 3-1, no clearly discernible regional differences were found
in the ranges of parameters analyzed. Additional urban areas would need to be examined to
discern broadscale patterns. The data indicate considerable variability in the concentration
fields. Mean O3 concentrations vary within individual urban areas by factors of 1.4 to 4 in
Table 3-1. Intersite correlation coefficients show mixed patterns (i.e., in some urban areas all
pairs of sites are moderately to highly correlated, while other areas show a larger range of
correlations). As may be expected, those areas showing a smaller range of seasonal mean
concentrations also show a smaller range of intersite correlation coefficients. However, there are
a number of cases where sites in an urban area may be moderately to highly correlated, but show
substantial differences in absolute concentrations. In many cases, P90 values can equal or exceed
seasonal mean O3 concentrations.
It is instructive to compare the metrics for spatial variability shown in Table 3-1 to those
calculated for PM25 and PM10.25 in the PM AQCD (U.S. Environmental Agency, 2004a). The
values for concentrations and concentration differences are unique to the individual species, but
the intersite correlation coefficients and the COD values can be directly compared. In general,
3-12
-------
Table 3-1. Summary Statistics for the Spatial Variability of O3 (in ppm) in Selected Urban Areas in the United States
Urban Area
Boston, MA
New York, NY
Philadelphia, PA
Washington, DC
Charlotte, NC
Atlanta, GA
Tampa, FL
Detroit, MI
Chicago, IL
Milwaukee, WI
V° St. Louis, MO
Baton Rouge, LA
Dallas, TX
Houston, TX
Denver, CO
El Paso, TX
Salt Lake City, UT
Phoenix, AZ
Seattle, WA
Portland, OR
Fresno, CA
Bakersfield, CA
Los Angeles, CA
Riverside, CA
Number of
Sites
18
29
12
20
8
12
9
7
24
9
17
7
10
13
8
4
8
15
5
5
6
8
14
18
Minimum
Mean Cone.
0.021
0.015
0.020
0.022
0.031
0.023
0.024
0.022
0.015
0.027
0.022
0.018
0.028
0.016
0.022
0.022
0.029
0.021
0.015
0.015
0.030
0.028
0.010
0.018
Maximum
Mean Cone.
0.033
0.041
0.041
0.041
0.043
0.047
0.035
0.037
0.039
0.038
0.038
0.031
0.043
0.036
0.044
0.032
0.048
0.058
0.038
0.036
0.047
0.047
0.042
0.054
Minimum
Corr. Coeff.
0.46
0.45
0.79
0.72
0.48
0.63
0.74
0.74
0.38
0.73
0.78
0.81
0.67
0.73
0.60
0.81
0.52
0.29
0.63
0.73
0.90
0.23
0.42
0.38
Maximum
Corr. Coeff.
0.93
0.96
0.95
0.97
0.95
0.94
0.94
0.96
0.96
0.96
0.96
0.95
0.95
0.96
0.92
0.94
0.92
0.95
0.94
0.91
0.97
0.96
0.95
0.95
Minimum
"90
0.012
0.0080
0.011
0.010
0.012
0.013
0.011
0.0090
0.0080
0.0090
0.0090
0.0090
0.011
0.0090
0.013
0.012
0.012
0.011
0.0080
0.011
0.0090
0.013
0.010
0.013
Maximum
p
^90
0.041
0.044
0.036
0.032
0.038
0.045
0.025
0.027
0.043
0.025
0.031
0.029
0.033
0.027
0.044
0.023
0.043
0.057
0.024
0.025
0.027
0.052
0.053
0.057
Minimum
COD11
0.17
0.17
0.23
0.17
0.17
0.24
0.20
0.19
0.16
0.18
0.15
0.23
0.16
0.20
0.16
0.24
0.13
0.15
0.16
0.20
0.17
0.20
0.22
0.15
Maximum
COD
0.45
0.55
0.46
0.45
0.32
0.55
0.35
0.36
0.50
0.33
0.41
0.41
0.36
0.38
0.46
0.31
0.51
0.61
0.46
0.50
0.40
0.58
0.59
0.64
, P90 = 90th percentile absolute difference in concentrations.
COD = coefficient of divergence for different site pairs.
-------
the variability in O3 concentrations is larger than for PM2 5 concentrations and comparable to that
obtained for PM10_2 5. Intersite correlation coefficients in some areas (e.g., Philadelphia, PA;
Atlanta, GA; Portland, OR) can be very similar for both PM2 5 and for O3. However, there is
much greater variability in the concentration fields of O3 as evidenced by the much higher COD
values. Indeed, COD values are higher for O3 than for PM2 5 in each of the urban areas
examined. In all of the urban areas examined for O3, some site pairs are always very highly
correlated with each other (i.e., r >0.9) as seen for PM2 5. These sites also show less variability
in concentration and are probably influenced most strongly by regional production mechanisms.
The above considerations indicate that caution should be observed in using data from the
network of ambient O3 monitors to approximate community-scale human exposures. A similar
conclusion was reached for PM using data from the PM2 5 FRM network, as indicated in
Section 3.4 of the PM AQCD (U.S. Environmental Protection Agency, 2004a).
3.3.1 Small-Scale Horizontal and Spatial Variability in
Ozone Concentrations
Ozone concentrations near roadways
Apart from the larger scale variability in surface O3 concentrations, there is also significant
variability on the micro-scale (< a few hundred meters), especially near roadways and other
sources of emissions that react with O3. These sources are not confined to urban areas. Sources
of emissions that react with O3 such as highways and power plants are also found in rural areas.
Johnson (1995) described the results of studies examining O3 upwind and downwind of
roadways in Cincinnati, OH. In these studies, O3 upwind of the roadway was about 50 ppb and
values as high as this were not found again until distances of about 100 m downwind. The O3
profile varied inversely with that of NO, as might be expected. For peak NO concentrations of
30 ppb immediately downwind of the road, the O3 mixing ratio was about 36 ppb, or about 70%
of the upwind value. The magnitude of the downwind depletion of O3 depends on the emissions
of NO, the rate of mixing of NO from the roadway and ambient temperatures. So depletions
of O3 downwind of roadways are expected, but with variable magnitude. In interpreting
historical data, it should be noted that scavenging of O3 by NO near roadways was more
pronounced before the implementation of stringent NOX emissions controls.
3-14
-------
Guidance for the placement of O3 monitors (U.S. Environmental Protection Agency, 1998)
states a separation distance that depends on traffic counts. For example, a minimum separation
distance of 100 m from a road with 70,000 vehicles per day (about 3,000 vehicles per hour) is
recommended for siting an O3 monitor to avoid interference that would mean a site is no longer
representative of the surrounding area. An average rate of about 3,000 vehicles per hour passing
by a monitoring site implies a road with rather heavy traffic. As noted in Section AX3.3.1 for
the Lakewood, CA monitoring, O3 levels are lower at sites located near traffic than those located
some distance away, and the scavenging of O3 by emissions of NO from roadways is a major
source of spatial variability in O3 concentrations.
Vertical Variations in Ozone Concentrations
In addition to horizontal variability in O3 concentrations, consideration must also be given
to variations in the vertical profile of O3 in the lowest layers of the atmosphere. The planetary
boundary layer (PEL) consists of an outer and an inner portion. The inner part extends from the
surface to about one-tenth the height of the PEL. Winds and transported pollutants, such as O3,
are especially susceptible to interactions with obstacles, such as buildings and trees in the inner
boundary layer (atmospheric surface layer) (e.g., Garratt, 1992). Inlets to ambient monitors
(typically at heights of 3 to 5 meters) are located in, and human and vegetation exposures occur
in this part of the boundary layer.
Photochemical production and destruction of O3 occur throughout the PEL. However, O3
is also destroyed on the surfaces of buildings, vegetation, etc. In addition, O3 is scavenged by
NO emitted by motor vehicles and soils. These losses imply that the vertical gradient of O3
should always be directed downward. The magnitude of the gradient is determined by the
intensity of turbulent mixing in the surface layer.
Most work characterizing the vertical profile of O3 near the surface has been performed in
nonurban areas with the aim of calculating fluxes of O3 and other pollutants through forest
canopies and to crops and short vegetation, etc. Corresponding data are sparse for urban areas.
However, monitoring sites are often set up in open areas such as parks and playgrounds where
surface characteristics may resemble those in rural areas more than those in the surrounding
urban area. The vertical profiles of O3 measured over low vegetation is shown in Figure 3-6.
These measurements were obtained as part of a field campaign to measure the fluxes of several
3-15
-------
4-1
Ul
'5
2-
1 -
0.5-
0
+ Average (n = 1797)
+ Stable (n= 937)
• Unstable (n = 860)
0.5
4
0.6 0.7 0.8
Relative O3 Concentration
0.9
Figure 3-6. Vertical profile of O3 obtained over low vegetation. Values shown are relative
to concentrations at 4 m above the surface. Ozone concentrations for stable
and unstable conditions were 41.3 and 24.1 ppb, and average O3 concentration
weighted by stability class was 33.1 ppb at 4 m.
Source. Horvath etal. (1995).
gas and aerosol phase pollutants using the gradient-flux technique (Horvath et al., 1995). The
labels stable and unstable in the figure refer to atmospheric stability conditions and average
represents the overall average. Ozone concentrations were normalized relative to their values at
a height of 4 m. As can be seen from the figure, there was a decrease of about 20% in going
from a height of 4 m down to 0.5 m above the surface during stable conditions, but O3 decreased
by only about 7% during unstable conditions. The average decrease was about 10% for all
measurements. As might be expected, O3 concentrations at all heights were very highly
correlated with one another. Of course, these values represent averages and there is scatter about
them. Under strongly stable conditions, they fall off toward the surface. However, these
conditions tend to occur mainly during night and the stability regime during the day in urban
areas tends more toward instability because of the urban heat island effect. Figure 3-7 shows the
vertical profile of O3 measured in a spruce forest (Horvath et al., 2003). The fall off in O3 for
3-16
-------
15-
D)
'5
2.5-
0.5-
+ Average (n = 1760)
+ Stable (n = 1272)
• Unstable (n = 488)
0.5
0.6 0.7 0.8
Relative O3 Concentration
0.9
Figure 3-7. Vertical profile of O3 obtained in a spruce forest. Values shown are relative to
concentrations at 19 m above the surface. Mean tree height is 14.5 m. Ozone
concentrations for stable and unstable conditions were 36.7 and 33.8 ppb, and
the average O3 concentration weighted by stability class was 34.6 ppb at 19 m.
Source: Horvath et al. (2003).
this case is due to uptake by trees, reaction with ambient NO and with NO emitted by the soil
in the forest, and reaction with hydrocarbons emitted by the trees in addition to deposition on
the surface.
3.4 DIURNAL AND SEASONAL VARIABILITY OF OZONE
Diurnal Variability
Diurnal variations in O3 at a given location are controlled by a number of factors, such as
the relative importance of transport versus local photochemical production and loss rates, the
timing for entrainment of air from the nocturnal residual boundary layer and the diurnal
variability in mixing layer height.
3-17
-------
Diurnal Patterns in the Nationwide Data Set
Composite urban, diurnal variations in hourly averaged O3 for April through October 2000
to 2004 are shown in Figure 3-8. As can be seen from Figure 3-8, daily 1-h maxima tend to
occur in mid-afternoon and daily 1-h minima tend to occur during the early morning. However,
there is also considerable spread in these times. Therefore, some caution must be exercised in
extrapolating results from one city to another and when attempting to judge the time of day when
the daily 1-h maximum occurs.
Urban Sites
0.200 n
•=• 0.150
Q.
Q.
C
o
C
0)
u
o
o
0)
C
o
N
o
0.100 -
0.050 -
0.000 -\
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
Figure 3-8. Composite, nationwide diurnal variability in hourly averaged O3 in
urban areas. Values shown are averages from April to October 2000 to
2004. Boxes define the interquartile range and the whiskers, the minima
and maxima.
Source: Fitz-Simons et al. (2005).
Corresponding data for 8-h average O3 variations are shown in Figure 3-9. As can be seen
from Figure 3-9, daily maximum eight hour O3 concentrations tend to occur from about 10 a.m.
to about 6 p.m. As can be seen from Figure 3-9, they can also occur at slightly different times
3-18
-------
Urban Sites
Q.
Q.
C
0)
o
c
o
O
0)
o
N
O
0.125 n
0.100 -
0.075 -
0.050 -
0.025 -
0.000 -I
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
Figure 3-9. Composite, nationwide diurnal variability in 8-h average O3 in urban areas.
Values shown are averages from April to October 2000 to 2004. Boxes define
the interquartile range and the whiskers, the minima and maxima. The hour
refers to the start of the 8-h averaging period.
Source: Fitz-Simons et al. (2005).
and the variation in the 8-h averages is smoother than for the 1-h averages. The minima in the
8-h averages tend to occur starting at about midnight.
Diurnal Patterns in EPA's 12 Cities
The diurnal variability of hourly averaged O3 in the twelve urban areas considered for
inclusion in EPA's human health exposure assessment risk assessment for the current review is
illustrated in Figures 3-10a-l for April to October. Daily maximum 1-h concentrations tend to
occur in mid-afternoon. However, as can be seen from the figures, the diurnal patterns vary
from city to city, with high values (>0.100 ppm) also occurring either late in the evening as in
Boston, past midnight as in Los Angeles and Sacramento, or midmorning as in Houston.
Typically, high values such as these are found during the daylight hours in mid to late afternoon.
The reasons for the behavior of O3 during the night at the above-mentioned locations are not
3-19
-------
a. Boston-Worcester-Manchester, MA-NH
b. New York-Newark-Bridgeport, NY-NJ-CT-PA
—- 0-125 •
E 0.125-
O
'•C 0,075
O
III
22 230001 02 03 04 05 06 07 OS 09 10 11 12 13 14 15 16 17 18 102021 2223 0001
hour
2223 00 01 0203 04 05 06 070809 10 11 12 13 14 15 '6 17 18 1920 21 ;
hour
c. Philadelphia-Camden-Vineland, PA-NJ-DE-ME
d. Washington-Baltimore-Northern Virginia, DC-MD-VA-WV
— 0125
— 0-125 •
22 23 00 01 02 03 04 05 060708 09 10 11 12 13 14 15 16 17 18 1920 21 2
hour
30001 2223 00 01 02030405 06 070809 '0 11 12 13 14 15 16 17 18 1920 21 ;
hour
e. Atlanta-Sandy Springs-Gainesville, GA-AL
—. 0,125
I
f. Cleveland-Akron-Elyria, OH
— 0125
O
O
2223 00 01 02 030405 06 070809 10 1" 12 1314 15 16 17 18 192021 22 23 0001
hour
22 23 0001 02 03 04 05 06070809 10 11 12 13 14 15 '6 17 15 1920 21 22 2300 01
hour
Figure 3-10a-f. Diurnal variability in hourly averaged O3 in selected urban areas. Values
shown are averages from April to October 2000 to 2004. Boxes define the
interquartile range and the whiskers, the minima and maxima.
Source: Fitz-Simons et al. (2005).
3-20
-------
g. Detroit-Warren-Flint, MI
h. Chicago-Naperville-Michigan City, IL-IN-WI
?? ?3 00 01 020304 05 06 070809 10 11 1? 13 14 15 16 17 18 1970 ?1 ?? 23 00 01
hour
22 23 0001 0? 030405 06 070809 10 11 12 13 14 15 16 17 18 192021 22 23 00 01
hour
i. St. Louis-St. Charles-Farmington, MO-IL
j. Houston-Baytown-Huntsville, TX
22 23 00 01 020304 05 06 070809 10 11 12 13 14 15 16 17 18 1920 21 22230001
hour
22 23 0001 02 03 04 05 060708 09 10 11 12 13 14 15 16 17 18 192021 22 2300 01
hour
k. Sacramento-Arden-Arcade-Truckee, CA-NV
a.
I. Los Angeles-Long Beach-Riverside, CA
E
S- 0.150
o
o
O 0.050 •
N
O
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 22 00 01
hour
22 23 00 01 02 03 0405 060708 09 10 11 12 13 14 15 16 17 18 192021 2223 CO 01
hour
Figure 3-10g-l. Diurnal variability in hourly averaged O3 in selected urban areas. Values
shown are averages from April to October 2000 to 2004. Boxes define the
interquartile range and the whiskers, the minima and maxima.
Source: Fitz-Simons et al. (2005).
3-21
-------
clear. Measurement issues may be involved, or there may be physical causes such as transport
phenomena, as discussed in Chapter 2. As discussed in Chapter 2, and in greater detail in
Section AX2.3.3, nocturnal low level jets are capable of producing secondary O3 maxima
at night.
The diurnal variability of O3 averaged over 8 hours in the same twelve urban areas is
shown in Figures 3-1 la-1. The diurnal patterns of 1-h O3 averages are broadly similar to 8-h
averages. Typically, although the 8-h daily maximum occurs between 10 a.m. and 6 p.m., actual
starting and ending times can differ from these characteristic times depending on location.
For example, as shown in Figures 3-1 la for Boston and 3-1 Ik for Sacramento, the highest 8-h
daily maximum values can start in mid-afternoon and extend into late evening. These results
suggest that transport processes are playing the dominant role in determining the timing of the
highest daily maxima in these areas.
On days with high 1-h daily maximum concentrations (e.g., >0.12 ppm), the maxima
tend to occur in a smaller time window centered in the middle of the afternoon, compared to
days on which the maximum is lower. For example, on high O3 days the 1-h maximum occurs
from about 11 a.m. to about 6 p.m. However, on days for which the 1-h daily maximum
is <0.080 ppm, the daily maximum can occur at any time during the day or night, with only a
50% probability that it occurs between 1 and 3 p.m., in each of the 12 cities. (The time of day
when the daily maximum 1-h O3 concentration occurs is illustrated for four of the cities in
Figures AX3-45a-d.). Photochemical reactions in combination with diurnal emissions patterns
are expected to produce mid-afternoon peaks in urban areas. These results suggest that transport
from outside the urban airshed plays a major role in determining the timing of the daily maxima
for low peak O3 levels. This pattern is typical for the Los Angeles-Long Beach-Riverside, CA
area even on high O3 days.
The same general timing patterns are found for 1-h daily maximum O3 concentrations as
for the daily maximum 8-h average O3 concentration. As mentioned above, the daily maximum
8-h O3 concentrations are generally found between the hours of 10 a.m. and 6 p.m. However,
there are a significant number of days when this is not the case, e.g., for high values in Houston,
TX and Los Angeles, CA, or in general for low values at any of the cities examined. (The time
of day when the daily maximum 8-h average O3 concentrations occurs is shown for four cities in
Figures AX3-46a-d.). Although the 8-h average O3 concentration is highly correlated with the
3-22
-------
a. Boston-Worcester-Manchester, MA-NH
0.125-,
b. New York-Newark-Bridgeport, NY-NJ-CT-PA
0125,
E 0.100-
O
'•C 0.075
22 23 00 01 02 03 04 05 06 07 OS 09 10 11 12 13 14 15 16'7 18 19 20 21 22230001
hour
22 230001 02 03 04 0506070809 10 11 12 13 14151617 18 1920 21 2223 00 01
hour
c. Philadelphia-Camden-Vineland, PA-NJ-DE-ME
3 1"5 i
8 0,025.
d. Washington-Baltimore-Northern Virginia, DC-MD-VA-WV
0.125 -i
E 0.100-
O
*= 0.075
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
2? 23 0001 02 03040506070809 10 11 1? 13 14 15 16 1718 19 ?0 ?1 ?2 ?3 00 01
hour
e. Atlanta-Sandy Springs-Gainesville, GA-AL
0 1?5 -,
E 0.100-
O
*= 0.075 •
f. Cleveland-Akron-Elyria, OH
0 125 i
O
o>
O
22230001 0203040506070809 10 11 12 13 14 15 16 "7 '8 1920 21 22 230001
hour
Figure 3-lla-f. Diurnal variability in 8-h O3 in selected urban areas. Values shown
are averages from April to October 2000 to 2004. Boxes define the
interquartile range and the whiskers, the minima and maxima.
The hour refers to the start of the 8-h averaging period.
Source: Fitz-Simons et al. (2005).
3-23
-------
g. Detroit-Warren-Flint, MI
h. Chicago-Naperville-Michigan City, IL-IN-WI
O
o
-^ 0.125 •
I
O
22230031 02030405080708091011 12 13'4 15 16 1718192021 22230001
hour
222300 01 0203040506 070809 10 11 12 13 14 15 16 17 16 1S 20 21 22 2300 01
hour
i. St. Louis-St. Charles-Farmington, MO-IL
j. Houston-Baytown-Huntsville, TX
— 0.125 •
O
o
s
o
22 2300 31 020304 0505 07 OS 09 10 11 1? 13 "4 15 16 17 18 192021 22 230001
hour
22230001 02 03 04 050607 0809 10 11 12 13 14 15 16 17 1 a 19 2021 22 23 QO 0
hour
k. Sacramento-Arden-Arcade-Truckee, CA-NV
I. Los Angeles-Long Beach-Riverside, CA
o
o
22 23 00 01 02 03 04 05 OS 07 08 09 10 "1 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
222300 01 0203 04 0506 07 08 09 1C 11 12 13 14 15 16 17 13 19 2021 222300 31
hour
Figure 3-llg-l. Diurnal variability in 8-h O3 in selected urban areas. Values shown
are averages from April to October 2000 to 2004. Boxes define the
interquartile range and the whiskers, the minima and maxima.
The hour refers to the start of the 8-h averaging period.
Source: Fitz-Simons et al. (2005).
3-24
-------
daily maximum 1-h average O3 concentration, there are situations where the daily maximum 8-h
average O3 concentration might be driven by very high values in the daily maximum 1-h average
O3 concentration as illustrated in Figure 3-lOj for Houston, TX. In cases such as these, the
predicted 8-h average may overestimate the short-term O3 concentration later in the day.
The patterns of diurnal variability for both 1-h and 8-h averages have remained quite stable
over the 15-year period from 1990 to 2004, with times of occurrence of the daily maxima
varying by no more than an hour from year to year in each of the 12 cities.
Weekday/Weekend Differences
Differences in the diurnal behavior of O3 have been observed in a number of cities (e.g.,
Heuss et al., 2003). Figures 3-12a-h show the contrast in the patterns of hourly averaged O3 in
the greater Philadelphia, Atlanta, Houston and Los Angeles areas from weekdays to weekends.
Daily maximum concentrations occur basically at the same time on either weekdays or
weekends. Differences are apparent in the hourly concentrations, especially in the extreme
values. Weekday/weekend differences in 8-h average O3 concentrations are shown in
Figures 3-13a-h. As can be seen from a comparison of the weekend versus weekday patterns,
there is a tendency for the lowest values in the distribution to be higher on weekends than on
weekdays. Lower traffic volumes, in particular diesel truck traffic, lead to less NO emissions
and titration of O3 on weekends. The spike in values for Houston in midmorning, shown in
Figure 3-12f, resulted from the release of highly reactive hydrocarbons from the petrochemical
industry (which could occur on any day of the week). Otherwise, the maximum O3
concentrations could be seen to occur on the weekdays as they do in Philadelphia and Atlanta,
in contrast to Los Angeles. Indeed, the diurnal pattern in Houston is similar to that observed in
Atlanta on weekdays, indicating some overall similarity in the sources of precursors of O3.
Spatial Variability in Diurnal Patterns in Urban Areas
Daily maxima in either the 1-h or 8-h averages do not necessarily occur at the same time of
day at each site in the 12 cities, and the diurnal pattern observed at individual sites can vary from
the composites shown in Figures 3-8 and 3-9. Differences between sites are not only related to
the distance between them; they also depend on nearby sources, such as highways, affecting one
site to a greater extent than another. For example, in the Los Angeles basin, daily 1-h maxima
3-25
-------
a. Philadelphia-Camden-Vineland, PA-NJ-DE-MD
(week day)
E
a.
a.
o
o
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
c. Atlanta-Sandy Springs-Gainesville, GA-AL
(week day)
B.
D.
O
O
b. Philadelphia-Camden-Vineland, PA-NJ-DE-MD
(week end)
~ 0,125
Q.
a
_o
S
o
O
o
8
22230001 02 03 04 05 06 07 OB 09 10 11 12131415161718192021 22230001
hour
d. Atlanta-Sandy Springs-Gainesville, GA-AL
(week end)
-~ 0.125 •
Q.
O.
O
I
S 0.075 •
-------
e. Houston-Baytown-Huntsville, TX
(week day)
f. Houston-Baytown-Huntsville, TX
(week end)
o.
Q.
O
O
Biaa
II
_o
o
o
O 0.050 •
s
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
g. Los Angeles-Long Beach-Riverside, CA
(week day)
22230001 02 03 04 05 06 07 OB 09 10 11 12131415161718192021 22230001
hour
h. Los Angeles-Long Beach-Riverside, CA
(week end)
O
o
O 0.050 •
N
O
£
S 0.100
o
o
O 0.050
N
O
2223 00 01 02 0304 05 0607 0809 10 11 12 13 14 15 16 17 18 192021 22 2300 01
hour
2223 0001 02 03 0405 06 0708 09 10 11 12 13 14 15 16 17 18 1920 21 22 230001
hour
Figure 3-12e-h. Diurnal variations in hourly averaged O3 on weekdays and weekends in
four cities. Values shown represent averages from May to September
of 2004. Boxes define the interquartile range and the whiskers, the
minima and maxima.
Source: Fitz-Simons et al. (2005).
remote from sources of precursors in urban/suburban areas, the behavior of O3 will follow these
basic patterns. Similar relations are found for the 8-h average O3 concentrations. Differences in
diurnal patterns between sites in urban cores and sites downwind of urban cores are illustrated
in Figures AX3-49a-b to AX3-51a-b for 1-h average O3 and in Figures AX3-52a-b to AX3-54a-b
for Detroit, MI, St. Louis, MO, and Riverside, CA areas.
3-27
-------
a. Philadelphia-Camden-Vineland, PA-NJ-DE-MD
(week day)
E 0,100
a.
o
S5 0075
2
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
c. Atlanta-Sandy Springs-Gainesville, GA-AL
(week day)
£ 0.100
B.
D.
O
5 0.075
b. Philadelphia-Camden-Vineland, PA-NJ-DE-MD
(week end)
E 0.100
a.
22230001 02 03 04 05 06 07 OB 09 10 11 12131415161718192021 22230001
hour
d. Atlanta-Sandy Springs-Gainesville, GA-AL
(week end)
£ 0.100
Q.
O.
o
g 0.025
2223 00 01 02 0304 05 0607 0809 10 11 12 13 14 15 16 17 18 192021 22 2300 01
hour
2223 0001 02 03 0405 06 0708 09 10 11 12 13 14 15 16 17 18 1920 21 22 230001
hour
Figure 3-13a-d. Diurnal variations in 8-h average O3 on weekdays and weekends in four
cities. Values shown represent averages from May to September of 2004.
Boxes define the interquartile range and the whiskers, the minima
and maxima. The hour refers to the start of the 8-h averaging period.
Source: Fitz-Simons et al. (2005).
Seasonal Variability
It should not be assumed that highest O3 levels are confined to the summer. Highest
average O3 concentrations generally occur at background monitoring sites at midlatitudes in the
Northern Hemisphere during late winter and spring versus summer, as for urban sites or for
nonurban sites heavily affected by regional pollution sources.
3-28
-------
e. Houston-Baytown-Huntsville, TX
(week day)
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
g. Los Angeles-Long Beach-Riverside, CA
(week day)
— 0,125
B.
D.
O
o
s
o
f. Houston-Baytown-Huntsville, TX
(week end)
22230001 02 03 04 05 06 07 OB 09 10 11 12131415161718192021 22230001
hour
h. Los Angeles-Long Beach-Riverside, CA
(week end)
-~ 0.125 •
Q.
O.
O
I
S 0.075 •
-------
a. Boston-Worcester-Manchester, MA-NH
0030-,
b. New York-Newark-Bridgeport, NY-NJ-CT-PA
11111 HID
. - - - .
TnTiiTiinni.
Ill null I ^
i °-M|> •
1.1 1 g
0
~ ~ O 0 020 -
LJ N
IIlniTTTTTII °
,„,„„„ n nnn .
r¥
T 1 T
llnnf 1 n
illmlllTT
T T
flrnnJLJLJL
2223 OC 01 02 03040506 0705 09 *0 11 12 13 14 15 16 17 18 1S 2021 222300 01
hour
222300 01 0203 040506 37 0809 10 11 12 13 1415 16 17 18 192021 2223 0001
hour
c. Philadelphia-Camden-Vineland, PA-NJ-DE-ME
OOSO i
d. Washington-Baltimore-Northern Virginia, DC-MD-VA-WV
o
I
« 0.040
II
£ C,C'40
IT
2223 0001 02 03 C4 05 06 07 08 OS 10 11 12 13 14 15 16 1718 192021 22 230001
hour
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 CO 01
hour
e. Atlanta-Sandy Springs-Gainesville, GA-AL
f. Cleveland-Akron-Elyria, OH
o
O
TTTT
TTTT
22 23 OC 01 02 03 04 05 06 0703 08 "011 12 13 14 15 16 17 18 IS 20 21 22230031
hour
?? 73 00 01 0? 03 040506 070809 10 11 1? 13 14 1516 17 18 19 ?0 ?1 7? ?3 CO 01
hour
Figure 3-14a-f. Diurnal variability in 8-h average O3 in selected urban areas. Values
shown are averages from November to March 2000 to 2004. Boxes define
the interquartile range and the whiskers, the minima and maxima. The
hour refers to the start of the 8-h averaging period.
Source: Fitz-Simons et al. (2005).
3-30
-------
g. Detroit-Warren-Flint, Ml
h. Chicago-Naperville-Michigan City, IL-IN-WI
o
o
o
o
2223000102030405060708091011 121314 151617 1819202122230001
hour
22 23 00 01 02 03 04 05 06 07 0809 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
I. St. Louis-St. Charles-Farmington, MO-IL
o
o
O 0-025 •
8
j. Houston-Baytown-Huntsville, TX
o
o
TT'i'V
?3 0001 0? 03 04 05 06 07 08 09 10 11 1? 13 14 15 16 17 18 1970 ?1 ?? ?3 00 01
hour
???3 0001 07 03 040506070809 1011 1? 13 14 15 16 1? 18 19 ?0 ?1 ?? ?3 00 01
hour
k, Sacramento-Arden-Arcade-Truckee, CA-NV
I. Los Angeles-Long Beach-Riverside, CA
5 0.050 •
o
o
I °-02H
TIT
| j
o
o
O 0-025 •
a
2223000102030405060708091011 12 131415181718192021 22230001
hour
22 23 00 01 02 030405060708091011 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
Figure 3-14g-l. Diurnal variability in 8-h average O3 in selected urban areas. Values
shown are averages from November to March 2000 to 2004. Boxes define
the interquartile range and the whiskers, the minima and maxima. The
hour refers to the start of the 8-h averaging period.
Source: Fitz-Simons et al. (2005).
3-31
-------
Diurnal Patterns in Nonurban Areas
Composite diurnal patterns of O3 are shown in Figure 3-15 for hourly averaged O3 and in
Figure 3-16 for 8-hour average O3 at rural (CASTNET) sites. As can be seen from a comparison
of Figures 3-15 and 3-16 with Figures 3-8 and 3-9, diurnal patterns of O3 are smoother and
shallower at the rural sites than at the urban sites. Maxima in hourly average O3 also tend to
occur in afternoon. However, highest concentrations observed during any particular hour at
night at the CASTNET sites (-0.130 ppm) are substantially higher than observed in urban areas
(<0.100 ppm) and daily 1-h maxima at CASTNET sites have exceeded 0.150 ppm. The diurnal
variations in 8-h average O3 concentrations are also much smaller at the CASTNET sites than at
the urban sites. Note also that the maxima in 8-h average O3 concentrations are higher at the
CASTNET sites than at the urban sites.
3.5 TRENDS IN OZONE CONCENTRATIONS
Year-to-year variability in the nationwide May to September, mean daily maximum 8-h O3
concentrations are shown in Figure 3-17. The corresponding year-to-year variability in the 95th
percentile concentrations is shown in Figure 3-18. Data flagged because of quality control issues
were removed with concurrence of the local monitoring agency. Only days with data for 18 of
24 hours were kept, and a minimum of 115 of 153 days were required in each year. Missing
years were filled in using simple linear interpolation, as done in EPA Trends reports. Year-to-
year variability in the 95th percentile values of the daily maximum 8-h O3 concentrations are
shown in Figure 3-18. Sites considered in this analysis are shown in the map in Figure AX3-3.
As was shown in Figures 3-1 and 3-2, most sites are located in the East. As can be seen from
Figure 3-17, the highest O3 concentrations have tended to decrease over the past 15 years, while
there has been little change in O3 concentrations near the center of the distribution. This is
consistent with observations in Europe (Volz-Thomas et al., 2003). Mean O3 concentrations
were slightly lower in 2003 and 2004 than in earlier years. The summer of 2003 was slightly
cooler than normal in the East (Levinson and Waple, 2004) and the summer of 2004 was much
cooler than normal in the East (Levinson, 2005) accounting in part for the dip in O3 during these
two years. Observations of O3 at a number of sites in the Northern Hemisphere likewise do not
show convincing evidence of strong upward trends during the 1990s (Oltmans et al., 1998).
3-32
-------
Rural (CASTNET) Sites
0.200 i
£
D- 0.150
a.
o
0.100 -
c
O
o
cu
O 0050
N
O
0.000 -
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
Figure 3-15. Composite diurnal variability in hourly O3 concentrations observed at
CASTNET sites. Values shown are averages from April to October 2000 to
2004. Boxes define the interquartile range and the whiskers, the minima
and maxima.
Source: Fitz-Simons et al. (2005).
Rural (CASTNET) Sites
0.200 i
ft 0.150
_O
I
C 0.100 -
o
c
o
o
CD
O 0050
O
0.000 -
22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01
hour
Figure 3-16. Composite diurnal variability in 8-h O3 concentrations observed at
CASTNET sites. Values shown are averages from April to October 2000 to
2004. Boxes define the interquartile range and the whiskers, the minima
and maxima. The hour refers to the start of the 8-h averaging period.
Source: Fitz-Simons et al. (2005).
3-33
-------
Nationwide Trends, May to September
Mean of Daily Maximum 8-Hour Values, 1990 - 2004
0.12-
0.11 -
-p 0.10-
e
3 0.09 -.
O 0.08 -
0.07 -
c
8 0.06 -
c
O 0.05 -.
O 0.03 -.
0.02-j
0.01 -
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Year
Figure 3-17. Year-to-year variability in nationwide mean daily maximum 8-h O3
concentrations. The whiskers on the box plot represent the 10th and 90th
percentile concentrations. The "X"s above and below the whiskers are the
values that fall below and above the 10th and 90th percentile concentrations.
The dots inside the box represent the mean, for the statistic, at all sites.
Source: Fitz-Simons et al. (2005)
There may even have been a slight increase in O3 concentrations near the bottom of the
distribution throughout the monitoring period. This would be consistent with data obtained in
Europe, showing that O3 minima increased during the 1990s because of reduced titration of O3
by reaction with NO in response to reductions in NOX emissions. As a result, the concentration
of Ox (NO2 + O3) shows little if any increase at all at the European sites (Volz-Thomas et al.,
2003). Trends in compliance metrics such as the fourth highest daily maximum 8-h O3
concentration can be found in the U.S. EPA Trends reports (http://www.epa.gov/airtrends) and
so are not repeated here.
3-34
-------
Nationwide Trends, May to September
95th Percentile of Daily Maximum 8-Hour Values, 1990 - 2004
E
Q.
o
c
o
O
0>
c
s
O
0.18-
0.17-
0.16-
0.15-
0.14-
0.13-
0.12-
0.11 -
0.10-
0.09-
0.08-
0.07-
0.06-
0.05-
0.04-
0.03-
0.02-
H
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Year
Figure 3-18. Year-to-year variability in nationwide 95th percentile value of the daily
maximum 8-h O3 concentrations. The whiskers on the box plot represent the
10th and 90th percentile values for the statistic. The "X"s above and below
the whiskers are the values that fall below and above the 10th and 90th
percentile values. The dots inside the box represent the mean, for the
statistic, at all sites.
Source: Fitz-Simons et al. (2005).
Figures 3-19a-h show year-to-year variability in mean daily 8-h O3 concentrations
observed at selected national park sites across the United States. Figures 3-20a-h show year-to-
year variability in the 95th percentile value of daily maximum 8-h O3 concentrations at the same
sites shown in Figures 3-19a-h. The same criteria used for calculating values in Figures 3-17
and 3-18 were used for calculating the May to September seasonal averages for the national
parks shown in Figures 3-19a-h and 3-20a-h. Sites at 22 national parks met these criteria, and
data for all 22 sites are given in Appendix AX3 in Figures AX3-66a-v and AX3-67a-v.
However, several monitoring sites were moved during the period from 1990 to 2004. Sites were
moved at Acadia NP in 1996, Joshua Tree NP in 1993, Mammoth Cave NP in 1996, Voyageurs
NP in 1996, and Yellowstone NP in 1996. These moves often resulted in offsets in O3 and so
3-35
-------
May to September Mean of Daily Maximum 8-Hour Values, 1990 - 2004
1996 19BB
Year
1996 1998
Year
2000 2002
0.09 :
0.08 -
0.07 -
o.oe -
0.05 -
0.04 -
0.03 •
0.02 -
0.01 ;
e, Theodore Roosevelt NP
0.09
1998 2000
0.08 •
£ 0.07 -
EX
£ 0.06 -
o
2 0.05 -
c
-------
May to September 95th Percentile of Daily Maximum 8-Hour Values, 1990 - 2004
0.09-
0.08-
0.07-
0.06-
0.05-
0.04-
0.03-
0.02
a, Brigantine NWR
1990 1992
1994
2000 2002
2004
2004
0.11
0.10;
0.09-
O.OB-
0.07-
0.06-
0.05-
0.04-
0.03-
0.02->r
c. Great Smoky Mountains NP
1996 1998
Year
1996 1998
Year
2002 2004
0.10 :
0.09-
0.08-
0.07-
0.06
0.05-
0.04-
0.03-
0.02 JT
e. Theodore Roosevelt NP
0.11
0.10:
0.09:
0.08 :
0.07 :
0.06-
0.05
0.04 :
0.03^
2000 2002
0.10-
0.09-
0.08-
0.07-
0.06:
0.05
0.04:
0.03-
0.024
g. Glacier NP
0.02
0.11
f. Yellowstone NP
1906 1998
Year
0.10:
0.09:
0.08:
0.07 :
0.06-
0.05:
0.04:
0.03 :
0.02
h. Chiricahua NM
1996 1998
Year
1996 1998
Year
Figure 3-20a-h. Year-to-year variability in 95th percentile of daily maximum 8-h O3
concentrations at selected national park (NP), national wildlife refuge
(NWR), and national monument (NM) sites.
Source: Fitz-Simons et al. (2005).
3-37
-------
trends for these locations have not been calculated (cf, Section AX3.6, Table AX3-9). As noted
in The Ozone Report—Measuring Progress through 2003 (U.S. Environmental Protection
Agency, 2004b), O3 trends in national parks in the South and the East are similar to nearby urban
areas and reflect the regional nature of O3 pollution. For example, O3 trends in Charleston, SC
and Charlotte, NC track those in nearby Cowpens NP and Cape Romaine NP in South Carolina;
O3 in Knoxville and Nashville, TN tracks O3 in Great Smoky NP; O3 in Philadelphia, PA and
Baltimore, MD tracks Brigantine NP in New Jersey; and New York, NY and Hartford, CT track
O3 in Cape Cod NS. The situation is not as clear in the West, where some national parks are
affected by local pollution sources (e.g., Lassen Volcanic National Park and Yosemite National
Park, CA) more than others. However, data obtained at these sites still provide valuable
information about the variability in regional background concentrations, especially since the
West has not been broken down into regions as has been done by Lehman et al. (2004) for the
East as shown in Figure 3-3. Comparison of Figures 3-19a-h and 3-20a-h shows that O3
concentrations near the center of the distribution do not necessarily track those at the upper end,
as pointed out earlier for the nationwide composite data set. Trends in the 98th and 95th
percentiles, and mean O3 concentrations for National Park sites presented in Table AX3-9
demonstrate this point. In addition, Table AX3-9 shows that trends reversed direction in going
from the 98th to 95th percentile values.
Caution should be exercised in using trends calculated at national parks to infer
contributions from distant sources either inside or outside of North America, because of the
influence of regional pollution. For example, using a 15-year record of O3 from Lassen Volcanic
NP and data from two aircraft campaigns, and observations spanning 18 years from five U.S.
west coast marine boundary layer sites, Jaffe et al. (2003) have estimated that the amount of O3
in air arriving from the Eastern Pacific in spring has increased by approximately 10 ppb from the
mid-1980s to the present. They suggested this positive trend might be due to increases of
emissions of O3 precursors in Asia. However, positive trends in O3 were found during all
seasons. Although the Lassen Volcanic NP site is not close to any major emission sources or
urban centers, maximum hourly average O3 concentrations of >0.080 ppm (during April-May)
and >0.100 ppm (during the summer) occur at Lassen Volcanic NP, reflecting the influence of
sources of O3 precursors in the area. Thus, although there is evidence that O3 levels may be
3-38
-------
increasing at some rural locations, there is also evidence that O3 levels at other locations have
either not increased or have decreased over the same period.
3.6 RELATIONSHIPS BETWEEN OZONE AND OTHER SPECIES
Correlations between Ozone and Other Species
In order to understand relationships among atmospheric species, an important distinction
must be made between primary (directly emitted) species and secondary (photochemically
produced) species. In general, it is likely that primary species will be highly correlated with
other primary species, and that secondary species will be highly correlated with other secondary
species. By contrast, primary species are less likely to be correlated with secondary species.
Secondary reaction products tend to correlate with each other, but there is considerable variation.
Some species (e.g., O3 and organic nitrates) are closely related photochemically and are highly
correlated. Others (e.g., O3 and H2O2) show a more complex correlation pattern. Further details
are given in Annex AX3 in Section AX3.7.
Relationships between primary and secondary components are illustrated by considering
data for O3 and PM2 5. Ozone and PM2 5 concentrations observed at a monitoring site in
Fort Meade, MD are plotted as binned means for different intervals in Figure 3-21, based on data
collected between July 1999 and July 2001. As can be seen from the figure, PM25 regarded as a
function of O3 increases to the left of the inflection point (at about 30 ppbv O3) and also
increases with O3 to the right of the inflection point. Data to the left of the minimum in PM2 5
were collected mainly during the cooler months of the year, while data to the right of the
minimum were collected during the warmer months. This situation arises because PM2 5
contains a large secondary component during the summer and has a larger primary component
during winter. During the winter, O3 comes mainly from the free troposphere, above the
planetary boundary layer and, thus, may be considered a tracer for relatively clean air, and it is
titrated by NO in the polluted boundary layer. Unfortunately, data for PM2 5 and O3 are collected
concurrently at relatively few U.S. sites throughout an entire year. So these results, while highly
instructive, are not readily extrapolated to areas where appreciable photochemical activity occurs
throughout the year. Ito et al. (2005) examined the relation between PM10 and O3 on a seasonal
basis in several urban areas (cf, Figure 7-24). Although PM10 contains proportionately more
3-39
-------
CO
E
c
o
§25 +
c
o
O
u> 20 +
re
0
E 10"
c
o
IE 5 +
c
o
O o
-+-
+
-+-
+
+
+
10 20 30 40 50 60
Conditional Mean 03 Concentration (ppbv)
70
Figure 3-21. Binned mean PM2 5 concentrations versus binned mean O3 concentrations
observed at Fort Meade, MD from July 1999 to July 2001.
Source: Chen (2002).
primary material than does PM25, relations similar to those shown in Figure 3-21 are found,
reflecting the dominant contribution of PM2 5 to PM10.
Other Oxidants
Measurements of gas phase peroxides in the atmosphere were reviewed by Lee et al.
(2000). Ground level measurements of H2O2 taken during the 1970s indicated values of 180 ppb
in Riverside, CA and 10 to 20 ppb during smog episodes in Claremont and Riverside, with
values approaching 100 ppb in forest fire plumes. However, later surface measurements always
found much lower values. For example, in measurements made in Los Angeles and nearby areas
in the 1980s, peak values were always less than about 2 ppb and in a methods intercomparison
study in Research Triangle Park, NC in June 1986, concentrations were <2.5 ppbv. Higher
values ranging up to 5 ppb were found in a few other studies in Kinterbish, Alabama and
Meadview, Arizona. Several of these studies found strong diurnal variations (typically about a
3-40
-------
factor of three) with maximum values in the mid-afternoon and minimum values in the early
morning. Mean concentrations of organic hydroperoxides at the surface at Niwot Ridge, CO in
the summer of 1988 and State Park, GA during the summer of 1991 were all less than a few ppb.
Aircraft measurements of hydroperoxide (H2O2, CH3OOH and HOCH2OOH)
concentrations were made as part of the Southern Oxidants Study intensive campaign in
Nashville, TN in July 1995 (Weinstein-Lloyd et al., 1998). The median concentration of total
hydroperoxides in the boundary layer between 1100 and 1400 CDT was about 5 ppbv, with more
than 50% contribution from organic hydroperoxides. Median O3 was about 70 ppbv at the same
time. The concentrations of the hydroperoxides depended strongly on wind direction with values
about 40% lower when winds originated from the N/NW as opposed to the S/SW suggesting that
local source areas were important.
Peroxyacetylnitrate (PAN) is produced during the photochemical oxidation of a wide range
of VOCs in the presence of NOX. It is removed by thermal decomposition and also by uptake to
vegetation (Sparks et al., 2003; Teklemariam and Sparks, 2004). PAN is the dominant member
of the broader family of peroxyacylnitrates (PANs) which includes as other significant
atmospheric components peroxypropionyl nitrate (PPN) of anthropogenic origin, and
peroxymethacrylic nitrate (MPAN) produced from oxidation of isoprene. Measurements and
models show that PAN in the United States includes major contributions from both
anthropogenic and biogenic VOC precursors (Horowitz et al., 1998; Roberts et al., 1998).
Measurements in Nashville during the 1999 summertime Southern Oxidants Study (SOS)
showed PPN and MPAN amounting to 14% and 25% of PANs respectively (Roberts et al.,
2002). Measurements during the TexAQS 2000 study in Houston indicated PAN concentrations
of up to 6.5 ppbv (Roberts et al., 2003). PAN measurements in southern California during the
SCOS97-NARSTO study indicated peak concentrations of 5-10 ppbv, which can be contrasted to
values of 60 to 70 ppbv measured back in 1960 (Grosjean, 2003). Vertical profiles measured
from aircraft over the U.S. and off the Pacific coasts show PAN concentrations above the
boundary layer of only a few hundred pptv, although there are significant enhancements
associated with long-range transport of pollution plumes from Asia (Kotchenruther et al., 2001a;
Roberts et al., 2004). Decomposition of this anthropogenic PAN as it subsides over North
America can lead to significant O3 production, enhancing the O3 background (Kotchenruther
et al., 2001b; Hudman et al., 2004).
3-41
-------
Oxidants are also present in airborne cloud droplets, rain drops and particulate matter.
Measurements of hydroperoxides, summarized by Reeves (2003), are available mainly for
hydrometeors, but are sparse for ambient particles. Venkatachari et al. (2005a) sampled the
concentrations of total reactive oxygen species (ROS) in particles using a cascade impactor in
Rubidoux, CA during July 2003. Although the species constituting ROS were not identified, the
results were reported in terms of equivalent H2O2 concentrations. Unlike O3 and gas phase H2O2
which show strong diurnal variability (i.e., about a factor of three variation between afternoon
maximum and early morning minimum), the diurnal variation of particle phase ROS was found
to be much weaker (i.e., less than about 20%) at least for the time between 8 a.m. and midnight.
Because the ROS were measured in the fine aerosol size fraction, which has a lifetime with
respect to deposition of much greater than a day, little loss is expected but their concentrations
might also be expected to increase because of nighttime chemistry, perhaps involving NO3
radicals. The ROS concentration, about 7 x io~9 M/m3 (expressed as equivalent H2O2), was at
most 1% that of O3 (6.2 to 38 x 10~7 M/m3 or 15 to 90 ppb), with highest values at night. In a
companion study conducted in Queens, NY during January and early February 2004,
Venkatachari et al. (2005b) found much lower concentrations of ROS of about 1.2 x 10~9 M/m3.
However, O3 levels were also substantially lower, but ROS concentrations were still less than
1% those of O3. It is of interest to note that gas phase OH concentrations measured at the same
time ranged from about 7.5 x lOVcm3 to about 1.8 x lOVcm3, implying the presence of
significant photochemical activity even in New York City during winter.
Co-Occurrence of Ozone with Other Pollutants
The characterization of co-occurrence patterns under ambient conditions is important for
relating human health and vegetation effects under ambient conditions to controlled research
results as described in Annex AX3.8. Several attempts have been made to characterize gaseous
air pollutant mixtures. The previous 1996 O3 AQCD discussed various patterns of pollutant
mixtures of SO2, NO2, and O3. Pollutant combinations can occur at or above a threshold
concentration at either the same or different times.
The 1996 O3 AQCD noted that studies of the joint occurrence of gaseous NO2/O3 and
SO2/O3 reached two conclusions: (1) hourly simultaneous and daily simultaneous-only
co-occurrences are fairly rare (when both pollutants were present at an hourly average
3-42
-------
concentration >0.05 ppm) and (2) when co-occurrences are present, complex-sequential and
sequential-only co-occurrence patterns predominate. Year-to-year variability was found to be
insignificant.
Using 2001 hourly data for O3 and NO2 and for O3 and SO2, the co-occurrence patterns for
the data are similar to those of previous studies. As shown in Figure 3-22, fewer than
10 co-occurrences of O3 and NO2 were found for most of the collocated monitoring sites.
Likewise, Figure 3-23 shows that fewer than 10 co-occurrences of O3 and SO2 were found for
most of the collocated monitoring sites analyzed.
40'
35-
J> 30"
55 25-
u—
2 20-
0)
I «H
z 1I
5-
co
CM
o
CM
CD
CO
O
CO
en
up
o
10
en
CO
o
CD
en
00
o
CO
CD
CD
O
CD
CD
O
O)
CM
CD
CO
CD
•si-
o
o
o
CN
O
CO
o
in
Number of Co-Occurances (Hours)
Figure 3-22. The co-occurrence pattern for O3 and nitrogen dioxide using 2001 data from
the AQS. There is co-occurrence when hourly average concentrations of O3
and another pollutant are both >0.05 ppm.
Since 1999, monitoring stations across the United States have routinely measured 24-h
average concentrations for PM2 5. Daily co-occurrence of PM2 5 and O3 over a 24-h period was
also characterized. Because PM2 5 data are mostly summarized as 24-h average concentrations in
the AQS database, a daily co-occurrence of O3 and PM25 was subjectively defined as an hourly
average O3 concentration >0.05 ppm and aPM25 24-h concentration >40 |ig/m3 (corresponding
to the EPA Air Quality Index, Level of Concern for PM2 5) occurring during the same 24-h
3-43
-------
60'
50-
w 4{H
"o
i- 30-|
O
E 20-
3
10-
co
CM
CD
CM
CD
CO
O
CO
O
in
CD
9
O
O>
op
O
oo
CO
CO
CD
CD
CO
O
o>
CM
CO
co
O)
IT)
O
O
O
CM
O
co
Number of Co-Occurances (Hours)
Figure 3-23. The co-occurrence pattern for O3 and sulfur dioxide using 2001 data
from AQS. There is co-occurrence when hourly average concentrations
of O3 and another pollutant are both ^0.05 ppm.
period. Using 2001 data from the AQS database, the daily co-occurrence of PM25 and O3 was
infrequent (Figure 3-24). Only limited data are available on the co-occurrence of O3 and other
pollutants (e.g., acid precipitation and acidic cloudwater). In most cases, routine monitoring data
are not available from which to draw general conclusions.
3.7 POLICY RELEVANT BACKGROUND OZONE CONCENTRATIONS
Background O3 concentrations used for purposes of informing decisions about NAAQS
are referred to as Policy Relevant Background (PRB) O3 concentrations. Policy Relevant
Background concentrations are those concentrations that would occur in the United States in the
absence of anthropogenic emissions in continental North America (defined here as the United
States, Canada, and Mexico). Policy Relevant Background concentrations include contributions
from natural sources everywhere in the world and from anthropogenic sources outside these
three countries. Background levels so defined facilitate separation of pollution levels that can be
controlled by U.S. regulations (or through international agreements with neighboring countries)
3-44
-------
160'
140-
J> 120-
w 100-
2 80-
0)
•g 60 H
^ 40-
20-
co
CM
CD
CM
CD
CO
O
CO
CO CD
"? 9
O O
m w
en
op
O
oo
CD
CO
CD
CD
CO
O
OJ
CM
CO
CO
O)
IT)
O
O
O
CM
O
co
Number of Co-Occurances (Hours)
Figure 3-24. The co-occurrence pattern for O3 and PM2 5 using 2001 data from AQS.
from levels that are generally uncontrollable by the United States. EPA assesses risks to human
health and environmental effects from O3 levels in excess of PRB concentrations. Issues
concerning the methodology for estimating PRB O3 concentrations are described in detail in
Annex AX3, Section AX3.9.
Contributions to PRB O3 include photochemical actions involving natural emissions of
VOCs, NOX, and CO as well as the long-range transport of O3 and its precursors from outside
North America and the stratospheric-tropospheric exchange (STE) of O3. Processes involved in
STE are described in detail in Annex AX2.3. Natural sources of O3 precursors include biogenic
emissions, wildfires, and lightning. Biogenic emissions from agricultural activities are not
considered in the formation of PRB O3.
Springtime maxima are observed at relatively remote (Annex AX3 and Figures 3-25a,b)
national park sites, located mainly in the western United States and at a number of other
relatively unpolluted monitoring sites throughout the Northern Hemisphere. The major issues
concerning the calculation of PRB O3 center on the capability of the current generation of global-
scale, three-dimensional, CTMs to simulate the causes of high O3 concentrations observed at
monitoring sites in relatively unpolluted areas of the United States from late winter through
spring (i.e., February through June). The issues raised do not affect interpretations of the causes
3-45
-------
Yellowstone National Park
Maximum Hourly Concentration
1998-2001
0 080 -
_. °-070 -
1"
g 0.060 -
IS
£ n n^n
o
o
ft nnn -
_
-
-
-
=
-
-
.
"
-
-
"
-
-
-
r
-
-
Month
11998 D2000 D2001
560391011
Figure 3-25a. Monthly maximum hourly average O3 concentrations at Yellowstone
National Park (WY) in 1998,1999, 2000, and 2001.
Source: U.S. Environmental Protection Agency (2003).
ooooooooooooooooooooooooooo
N«««CN««W«««WW«W«W«<>I<>I«N[«««««OI
Time
Figure 3-25b. Hourly average O3 concentrations at Yellowstone National Park (WY) for
the period January to December 2001.
Source: U.S. Environmental Protection Agency (2003).
3-46
-------
of summertime O3 episodes as strongly. Summertime O3 episodes are mainly associated with
slow- moving high-pressure systems characterized by limited mixing between the planetary
boundary layer and the free troposphere, as noted in Annex AX2, Section AX2.3.
A large number of case studies document the occurrence of STE mainly during winter and
spring in mid- and high-latitudes in Europe, Asia, and North America. These studies were based
on aircraft, satellite, and ground-based measurements. Considerable uncertainty exists in the
magnitude of the exchange; however, these studies have found that STE occurs throughout the
year, but with a distinct preference for the transport of O3 directly to the middle and lower
troposphere during late winter and spring. Transport to the upper troposphere occurs throughout
the year.
Springtime maxima in tropospheric O3 observed at high latitudes are also associated with
the winter buildup of O3 precursors and thermally labile reservoir species, such as PAN and
other reactive nitrogen species. These pollutants originate from all continents in the Northern
Hemisphere. Ozone precursor concentrations reach a maximum in late March; and as sunlight
returns to the Arctic, photochemical reactions generate tropospheric O3 (Section AX3.9.1).
The contribution of Asian sources to the U.S. levels is also largest during spring, reflecting the
efficient lifting of Asian pollution ahead of cold fronts originating in Siberia and transport by
strong westerly winds across the Pacific (e.g., Hudman et al., 2004). The longer lifetime of O3
during spring also contributes to springtime maxima (Wang et al., 1998).
Estimates of PRB concentrations cannot be obtained solely by examining measurements
of O3 obtained at RRMS in the United States (Annex AX3, Section AX3.2.3) because of the
long-range transport from anthropogenic source regions within North America. It should also be
noted that it is impossible to determine sources of O3 without ancillary data that could be used as
tracers of sources or to calculate photochemical production and loss rates. The current definition
of PRB implies that only CTMs can be used to estimate the range of PRB values. On the
synoptic and larger spatial scales at least, all evidence indicates that global CTMs are adequate
tools to investigate the factors controlling tropospheric O3; and three-dimensional CTMs, as
typified by Fiore et al. (2003) appear to offer the best methodology for estimating PRB
concentrations that cannot be measured directly (Annex AX3, Section AX3.9.2), at least for
averaging periods of longer than one hour.
3-47
-------
Previous estimates of background O3 concentrations, based on different concepts of
background, are given in Table 3-2. Results from global three-dimensional CTMs, where the
background is estimated by zeroing anthropogenic emissions in North America (Table 3-8) are
on the low end of the 25 to 45 ppbv range. Lefohn et al. (2001) have argued that frequent
occurrences of O3 concentrations above 50 to 60 ppbv at remote northern U.S. sites in spring are
mainly stratospheric in origin. Fiore et al. (2003) used a global CTM to determine the origin of
the high-O3 events reported by Lefohn et al. (2001), and to conduct a more general quantitative
analysis of background O3 as a function of season, altitude, and local O3 concentration.
Table 3-2. Previous Estimates of Background O3 in Surface Air Over the United States
Study
Method
Time Period Region
Background
Estimate (ppbv)
Trainer et al. y-intercept of O3 vs. NOy-NOx
(1993) regression linea
Hirsch et al. y-intercept of O3 vs. NOy-NOx
(1996) regression line
Altshuller y-intercept of O3 vs. NOy
and Lefohn regression line, and observations
(1996) at remote/rural sites
Liang et al. Sensitivity simulation in a 3-D
(1998) model with anthropogenic NOX
emissions in the continental U.S.
set to zero
Lin et al. Median O3 values for the lowest
(2000) 25th percentiles of CO and NOy
concentrations
Fiore et al. O3 produced outside of the North
(2002) American boundary layer in a
global 3-D model
Summer 1988 Eastern United
States
May-Sep Harvard Forest0
1990-1994
Apr-Oct Continental
1988-1993 United States
Full year
Continental
United States
30-40b
25 (Sept) - 40 (May)d
25-45 (inland)6
25-35 (coastal)
20-30 (East/
20-40 (West)
(spring maximum)
1990-1998 Harvard Forest 35 (fall) - 45 (spring)8
Summer 1995 Continental 15-30 (East)11
United States 25-35 (West)
aNOy is the chemical family including NOX and its oxidation products; NOy-NOx denotes the chemically
processed component of NOy.
b 1300-1700 local time (LT) in flatland and valley sites; all daytime measurements at elevated sites.
0 rural site in central Massachusetts.
d 1100-1700 EST hourly means.
e seasonal 7-h (0900-1559) daylight average.
f 1300-1600 LT monthly mean.
gdaily max 8-h averages.
h 1300-1700 average.
Source: Fiore et al. (2003).
3-48
-------
Figure 3-26 shows a comparison between observations obtained at CASTNet sites and
model results of Fiore et al. (2003). They classified the CASTNet monitoring sites into
low-lying sites (generally <1.5 km) and elevated sites (>1.5 km). All elevated sites are in the
West. Results were then aggregated to construct the cumulative probability distributions shown
in Figure 3-26 for the 58 low-altitude sites and the 12 high-altitude sites as well as for the three
seasons. The calculated mean background at the surface sites in spring is 27 ppbv, compared to
23 ppbv in summer and fall. At these sites, the background is highest for O3 concentrations near
the center of the distribution, and it declines as total surface O3 concentrations increase, for
reasons summarized below and discussed by Fiore et al. (2002). The observed O3 concentration
thus serves a surrogate for meteorological variability (i.e., stagnant versus ventilated conditions),
such that the background O3 is smaller on days when total O3 is highest. At the elevated sites,
the calculated mean background is 36 ppbv in spring versus 30 ppbv in the summer and fall.
Background concentrations in the fall resemble those in summer but show less variability and do
not exceed 40 ppbv anywhere in this analysis.
Major conclusions from the Fiore et al. (2003) study (discussed in detail in Annex AX3,
Sections AX3.9.3 and AX3.9.4) are:
• PRB O3 concentrations in U.S. surface air from 1300 to 1700 local time are generally
15 to 35 ppbv. They decline from spring to summer and are generally <25 ppbv under
the conditions conducive to high-O3 episodes.
• PRB O3 concentrations can be represented as a function of season, altitude, and total
surface O3 concentration, as illustrated in Figure 3-26.
• High PRB concentrations (40 to 50 ppbv) occur occasionally at high-elevation sites
(>1.5 km) in spring due to the free-tropospheric influence, including a 4- to 12-ppbv
contribution from hemispheric pollution (O3 produced from anthropogenic emissions
outside North America). These sites cannot be viewed as representative of low-elevation
surface sites (Cooper and Moody, 2000), where the background is lower when O3
>60 ppbv.
• The stratospheric contribution to surface O3 is of minor importance, typically well
<20 ppbv. While stratospheric intrusions might occasionally elevate surface O3
at high-altitude sites, these events are rare.
Appropriate background concentrations should thus be allowed to vary as a function of
season, altitude, and total O3 level. The diamonds in Figure 3-26 can be applied for this purpose.
In particular, the depletion of the background during high-O3 events should be taken into account
3-49
-------
120
100
« 80
, 60
I" 40
20
0
Elevated Sites (> 1.5 km)
i i i i i i i
1 2.5 16 50 84 97.5 99
120
100
80
60
40
20
0
Low-Lying Sites (< 1.5 km)
1 2.5 16 50 84 97.5 99
120
-
Q. < 80
Q. _•_
*T" 3 60
0 -»
O c 40
£ =
O ~> 20
0
I I I I I I I
1 2.5 16 50 84 97.5 99
120
100
80
60
40
20
0
1 2.5 16 50 84 97.5 99
120
100
O 80
O
' 60
I 40
20
0
"I I I I I I1!
1 2.5 16 50 84 97.5 99
120
100
80
60
40
20
0
1 2.5 16 50 84 97.5 99
Cumulative Probability (%)
Figure 3-26. Estimates of background contribution to surface afternoon (13 to 17 LT)
O3 concentrations in the United States as a function of local O3 concentration,
site altitude, and season. The figure shows cumulative probability
distributions of O3 concentrations for the observations (asterisks) and the
model (triangles). The corresponding distribution of background O3
concentrations is shown as grey diamonds.
Source: Fiore et al. (2003).
3-50
-------
(i.e., background O3 is depleted by reactions in the atmosphere and by deposition to the surface
but is not replenished at a significant rate in the stable, polluted boundary layer). This depletion
is shown in the right-hand panels of Figure 3-26 for the highest O3 values. Note that the model
is generally able to reproduce the overall frequency distributions in Figure 3-26. Typically,
models produce distributions flatter than are observed. Underpredictions, especially at the upper
end of the frequency distribution during the warmer months, are likely related to sub-grid-scale
processes that the model cannot resolve explicitly. The highest observed O3 concentrations in all
three seasons and at all altitudes are associated with regional pollution (i.e., North American
anthropogenic emissions), rather than stratospheric influence.
PRB ozone is not a directly observable quantity and must therefore be estimated from
models. Simple modeling approaches, such as the use of back-trajectories at remote U.S. sites to
identify background conditions, are subject to errors involving the reliability of the trajectories,
chemical production along the trajectories, and the hemispheric-scale contribution of North
American sources to ozone in air masses originating outside the continent. They also cannot
describe the geographical variability of the ozone background or the depletion of this
background during pollution episodes. Global 3-D chemical transport models such as
GEOS-Chem can provide physically-based estimates of the PRB and its variability through
sensitivity simulations with North American anthropogenic sources shut off. These models are
also subject to errors in the simulation of transport and chemistry, but the wealth of data that
they provide on ozone and its precursors for the present-day atmosphere enables extensive
testing with observations, and thus objective estimate of the errors on the PRB ozone values.
We present here such an error analysis for the Fiore et al. (2003a) PRB ozone estimates,
focusing on evaluation of their GEOS-Chem simulation for present-day conditions with ozone
observations of particular relevance to the PRB. Comparisons to ozonesonde observations of the
vertical distribution of tropospheric ozone at northern mid-latitudes are of particular interest.
These have been documented in a number of GEOS-Chem publications (Bey et al., 2001;
Liu et al., 2002; Martin et al., 2002; Fusco and Logan, 2003), including specifically over
North America (Li et al., 2002a, 2005). Results indicate no significant bias, and agreement to
generally within 5 ppbv for monthly mean concentrations at different altitudes.
Fiore et al. (2002, 2003a,b) presented detailed evaluations of GEOS-Chem with ozone
observations at U.S. surface sites. These evaluations focused on the afternoon hours (13-17 local
3-51
-------
time), when surface measurements are representative of a deep mixed layer that can be resolved
with the model. At night, surface ozone depletion often takes place by titration or deposition
under locally stratified conditions, but such conditions cannot be simulated with confidence by a
global model. The issue is not only one of vertical resolution (the lowest layers in GEOS-Chem
extend to 20, 50, 100, 200, and 400 m above the local surface) but also of horizontal resolution
(2°x2.5°). The PRB estimates presented by Fiore et al. (2003a) are also for 13-17 local time, and
lower values would apply at night.
The GEOS-Chem evaluation of U.S. surface ozone presented by Fiore et al. (2003a)
included monthly means, probability distributions, and time series at the sites previously used by
Lefohn et al. (2001) to estimate background ozone. The simulated monthly mean concentrations
in different seasons are typically within 5 ppbv of observations, with no significant bias, except
in the southeast in summer when the model is 8-12 ppbv too high due to excessive background
ozone advected in from the Gulf of Mexico. The time series comparisons for specific sites show
that the model simulates the day-to-day variability of ozone and reveals no further bias.
In particular, the model can reproduce the occurrences of relatively high ozone at remote sites
previously reported by Lefohn et al. (2001), and shows that these can generally be explained by
North American pollution.
Goldstein et al. (2004) presented comparisons of GEOS-Chem and MOZART global
model results with observations at Trinidad Head, California, during April-May 2002. The
observations, filtered to remove local influence, averaged 41 ± 5 ppbv, as compared to
GEOS-Chem (39 ± 5 ppbv) and MOZART (37 ± 9 ppbv). Neither model was successful at
reproducing the weak day-to-day structure in the observations, but they showed no bias in the
simulation of occasional >50 ppbv days. Hudman et al. (2004) found that GEOS-Chem
underestimated ozonesonde measurements in the free troposphere over Trinidad Head during the
same period by 10 ppbv on average, apparently due to insufficient regional stratospheric
influence. This bias does not extend to the surface, likely because stratospheric influence greatly
diminishes as free tropospheric air is diluted in the boundary layer.
Several other papers have evaluated the GEOS-Chem simulation for surface ozone and its
precursors over the United States. Fiore et al. (2003b) found a mean model bias of 6 ± 6 ppbv in
surface air over the eastern United States in summer, reflecting the overestimate in the southeast;
but they found that the model captured the large-scale spatial structure of the ozone observations
3-52
-------
as revealed by an EOF analysis. The model also reproduces without significant bias the
correlations of ozone with CO and NOy over North America and downwind (Fiore et al., 2002;
Li et al., 2002b, 2004). Comparisons to satellite observations of NO2 and formaldehyde
tropospheric columns over North America show excellent correspondence in the spatial patterns
and no significant biases.
In conclusion, we estimate that the PRB ozone values reported by Fiore et al. (2003a) for
afternoon surface air over the United States are likely 10 ppbv too high in the southeast in
summer, and accurate within 5 ppbv in other regions and seasons. The PRB ozone is likely
lower at night than in the afternoon though Fiore et al. (2003a) do not quantify the magnitude of
this diurnal effect.
Chemistry transport models should be evaluated with observations given earlier in
Chapter 3, in Annex AX3, and to simulate the processes causing the intra-day variability in O3
concentrations shown in Figure 3-27 in addition to those summarized in Chapter 2. The diurnal
patterns shown in Figure 3-27 do not fit the smooth pattern shown in Figure 3-15 and indicate
processes capable of producing rapid rises in O3 at times when substantial photochemical
activity is not present and may indicate stratospheric effects. Higher resolution models capable
of spatially and temporally resolving stratospheric intrusions and capable of resolving O3
variability on hourly timescales have not been applied to this problem. Ebel et al. (1991) have
demonstrated that regional-scale CTMs could be used to study individual stratospheric
intrusions. As an example of the utility of different types of models, Zanis et al. (2003) were
able to forecast, observe, and model a stratospheric intrusion (maximum penetration depth was
to slightly >2 km altitude) that occurred from June 20 to 21, 2001, over a large swath of central
Europe. Roelofs et al. (2003) compared results from six global tropospheric CTMs with lidar
observations obtained during that event and concluded that the models qualitatively captured the
features of this intrusion. It was also found that the coarser resolution models overestimated
transport to lower altitudes. The use of higher resolution models, perhaps nested inside the
coarser resolution models, may have helped solve this problem. They would also better address
issues related to temporal (i.e., 1-h versus 8-h averages) and spatial (i.e., populated versus
remote areas) scales needed by policymakers.
Although many of the features of the day-to-day variability of O3 at RRMS in the United
States are simulated reasonably well by Fiore et al. (2003), uncertainties in the calculation of the
3-53
-------
a. Denali National Park, AK
April 8-10, 2001
56-
-» 54"
^
0. 52-
_a
"^ 50-
O
48-
46-
AA-
f \
1 \ /""V"1 ""'"""""U..
,.,/ V ~"\
w
b. Voyageurs National Park, MN
May 19-20, 2001
Q.
o
O
50-
9D-
1 D-
n
>^N
S****~*S\ rl / \
*^~s ~Y\ / \
^/ V
OCOCOOTCMLnoOi-OCOCOOTOMLnoO —
T- T- T- CM T-T-T-CM
Time
Time
60-
c. Olympic National Park, WA
May 26-27, 2001
A
Q.
20-
10-
A/ * ,
Time
60-
50-
£t
Q.
O 20-
10-
d. Glacier National Park, MT
May 4-5, 2001
e. Yellowstone National Park, WY
April 25-26, 2001
f»»»» T^ »
/ %••_ A^/~ y
/ \ ^ ^
/ %
/
A
v
60-
£1
Qfl-
o
90-
10-
n.
A/ — v.
,****** **•*• « «^
« / ^*^ /
W^¥
OCO(DO)CMir>OOT-OCO(DO)CMir)00->-
T- T- T- CN T-T-T-CM
Time
Time
Figure 3-27. Time-series of hourly average O3 concentrations observed at five national
parks: Denali (AK), Voyageur (MN), Olympic (WA), Glacier (MT), and
Yellowstone (WY).
3-54
-------
temporal variability of O3 originating from different sources on shorter time scales must be
recognized. The uncertainties stem in part from an underestimate in the seasonal variability in
the STE of O3 (Fusco and Logan, 2003), the geographical variability of this exchange, and the
variability in the exchange between the free troposphere and the planetary boundary layer in the
model.
Ideally, the predictions resulting from an ensemble of models should be compared with
each other and with observations, so that the range of uncertainty inherent in the model
predictions can be evaluated. In this regard, it should be noted that only one model (GEOS-
Chem) is documented in the literature for estimating PRB concentrations.
3.8 OZONE EXPOSURE IN VARIOUS MICROENVIRONMENTS
Humans are exposed to O3 and related photochemical oxidants through the exchange
boundary, the skin and the openings into the body such as the mouth, the nostrils, and punctures
and lesions in the skin (U.S. Environmental Protection Agency, 1992; Federal Register, 1986).
Inhalation exposure to O3 and related photochemical oxidants is determined by pollutant
concentrations measured in the breathing zone that is not affected by exhaled air as the
individual moves through time and space. A discussion of the basic terminology associated with
exposure appears in AX3.
Quantification of Exposure
Ambient O3 concentrations vary with time of day (peaking during the latter portion of the
day) and season and among locations. Consequently, exposure to O3 will change as a function of
time of day and as an individual moves among locations. A hypothetical exposure is
demonstrated in Figure 3-28. The actual dose received also changes during the day and is
dependent on the O3 concentration in the breathing zone and the individual's breathing rate,
which is, in turn, dependent on the individual's level of exertion.
When measuring or modeling exposure to O3 and related photochemical oxidants
consideration should be given to the diurnal weekly (weekday-weekend) and seasonal
variability. Peak concentrations lasting for several hours typically occur toward the latter
portion of the day during the summer months. Regional O3 episodes often co-occur with high
3-55
-------
c:
.2
it--*
ro
•4—*
c
0
O
C
O
O
Peak
Exposure
Averaged
Exposure
Instantaneous
^jf-f^" Exposure
V*\ Area = Integrated ;;;
\VX Exposure '::. -
Time
Figure 3-28. Hypothetical exposure time profile: pollutant exposure as a function of time
showing how the average exposure, integrated exposure, and peak exposure
relate to the instantaneous exposure. (t2 - tt = T)
Source: U.S. Environmental Protection Agency (2004a).
concentrations of airborne fine particles, making it difficult to assess O3 dynamics and exposure
patterns. Also, while there are few indoor O3 sources, O3 will react with materials and other
pollutants in the indoor environment in an analogous fashion to that occurring in the ambient
atmosphere, potentially exposing subjects to other more toxic pollutants (Nazaroff and Weschler,
2004; Lee and Hogsett, 1999; Wainman et al., 2000; Weschler and Shields, 1997). (See
discussion on O3 chemistry and indoor sources and concentrations later in this chapter.).
Personal Exposure and Ambient Concentrations
The two approaches for measuring personal exposure are (a) the direct approach, using a
personal exposure monitor (PEM) consisting of a passive sampler worn around the breathing
zone, and (b) the indirect approach, which measures or estimates the O3 concentrations through
the use of models or biomarkers. Both approaches are associated with measurement error.
3-56
-------
Although it is difficult to develop passive monitors for personal exposure measurements
because of problems in identifying chemical or trapping reagents that can react with O3, several
modified passive samplers have been developed for use in personal O3 exposure measurements
(Bernard et al., 1999; Koutrakis et al., 1993; Avol et al., 1998b; Geyh et al., 1997, 1999).
Personal exposure measurements using passive samplers show O3 exposures below those O3
concentrations measured at outdoor stationary sites (Delfino et al., 1996; Avol et al., 1998a;
Sarnat et al., 2000; Geyh et al., 2000). However, there is a strong correlation between O3
measured at stationary sites and personal monitored concentrations (Liard et al., 1999; Brauer
and Brook, 1997; Linn et al., 1996; Lee et al., 2004; Avol et al., 1998a; O'Neill et al., 2003)
when the time spent outdoors, age, gender, and occupation of the subjects were considered.
The indirect approach determines and measures the concentrations in all of the locations or
"microenvironments." The concept of microenvironments is important in the understanding of
human exposure modeling. Often identified with a perfectly mixed compartment,
microenvironments are more recently viewed as a controlled volume, indoors or outdoors, that
can be characterized using a set of either mechanistic or phenomenological governing equations.
This allows for a nonhomogeneous environment, including sources and sinks within the
microenvironment. Microenvironments include indoor residences, other indoor locations,
outdoors near roadways, other outdoor locations, and areas within vehicles.
Microenvironmental Concentration and Ozone Exposure Models
Outdoor concentrations of O3 are estimated either through emissions-based mechanistic
modeling, or through ambient-data-based modeling. Emissions-based models determine the
spatiotemporal fields of the O3 concentrations using precursor emissions and meteorological
conditions as inputs. (They are described in Annex AX2.). The ambient-data-based models
determine spatial or spatiotemporal distributions of O3 through the use of interpolation schemes.
The kriging approach provides standard procedures for generating an interpolated O3 spatial
distribution for a given period of time (Georgopoulos et al., 1997a,b). The Spatio-Temporal
Random Field (STRF) approach has been used to interpolate monitoring data in both space and
time (Christakos and Vyas, 1998a,b). The STRF approach can analyze information on temporal
trends which cannot be directly incorporated by kriging.
3-57
-------
Several approaches are available for modeling microenvironmental concentrations:
empirical, mass balance, and detailed computational fluid dynamics (CFD) models. Empirical
relationships provide the basis for future, "prognostic" population exposure models. Mass
balance modeling is the most common approach used to model pollutant concentrations in
enclosed microenvironments. Mass balance modeling ranges from very simple formulations,
assuming ideal (homogeneous) mixing and only linear physicochemical transformations with
sources and sinks, to models that account for complex multiphase chemical and physical
interactions and nonidealities in mixing. Mass balance models take into account the effects of
ventilation, filtration, heterogeneous removal, and direct emission as well as photolytic, thermal,
and chemical reactions. The simplest form of the model is represented by the following
differential equation:
/Af „
—£- = VCOUT + -y
where dCIN is the indoor pollutant concentration (mass/volume), dt is time in hours, v is the air
exchange rate, COUT is the outdoor pollutant concentration (mass/volume), Fis the volume of the
microenvironment, and S is the indoor source emission rate. When the model was used to
estimate indoor O3 concentrations, indoor concentrations were found to be 33% of outdoor O3
concentrations (Freijer and Bloemen, 2000). A more in-depth discussion of the mass balance
model has been reported in Nazaroff and Cass (1986). The pNEM/O3 model, discussed later in
this chapter, includes a sophisticated mass balance model for indoor and vehicle
microenvironments (Johnson, 2003). CFD models take into account the complex, multiphase
processes that affect indoor concentrations of interacting gas phase pollutants, such as the
interactions of O3 with indoor sinks and sources (surfaces, gas releases) and with entrained gas
(Sarwar et al., 2001, 2002; S0rensen and Weschler, 2002).
Exposure modeling is often used in evaluating exposure to large populations over time.
The use of models is complicated by the fact that O3 is a secondary pollutant with complex
nonlinear and multiscale dynamics in space and time. Ozone is formed in the atmosphere
through a series of chemical reactions involving precursor VOCs and NOX. Therefore, O3
exposures may be affected by: (1) emission levels and spatiotemporal patterns of VOCs
and NOX; (2) ambient atmospheric as well as indoor microenvironmental transport, removal
3-58
-------
and mixing processes; and (3) chemical transformations that take place over a multitude of
spatial scales. The transformations are dependent on the presence of co-occurring pollutants and
the nature of surfaces interacting with the pollutants.
Exposure models may be classified as (1) potential exposure models, typically the
maximum outdoor concentrations versus "actual" exposure, including locally modified
microenvironmental outdoor and indoor exposures; (2) population versus "specific individual"-
based exposure models; (3) deterministic versus probabilistic models; and (4) observation versus
mechanistic air quality model-driven estimates of spatially and temporally varying O3
concentrations.
There are several steps involved in defining exposure models. The steps are based on
frameworks described in the literature over the last 20 years and the structure of various existing
inhalation exposure models (NEM/pNEM, MENTOR/SHEDS, REHEX, TREVI.Expo also known
as APEX, AIRPEX, AIRQUIS). The steps include (1) estimation/ determination of the
background or ambient levels of O3; (2) estimation/determination of levels and temporal profiles
of O3 in various microenvironments; (3) characterization of relevant attributes of individuals or
populations under study (age, gender, weight, occupation, other physiological characteristics);
(4) development of activity event or exposure event sequences; (5) determination of appropriate
inhalation rates during the exposure events; (6) determination of dose; (7) determination of
event-specific exposure and intake dose distributions for selected time periods; and
(8) extrapolation of population sample (or cohort) exposures and doses to the entire populations
of interest. Figure 3-29 provides a conceptual overview of a current exposure model. A more
detailed overview of an exposure model can be found in Annex AX3.
To estimate the actual O3 dose delivered to the lung, information on the concentration,
minute ventilation rate, activity level, and the morphology of the respiratory tract are needed.
Limited data have been compiled for ventilation rates for different age groups, both healthy and
compromised individuals, at varies levels of activity (Klepeis et al., 1996, 2001; Avol et al.,
1998a; Adams, 1993). Based on the available information, the highest level of outdoor activity
occurs during the spring and summer months, during the mid- to late afternoon and early
evening—the times when O3 concentrations are highest. Children are likely more susceptible to
the effects of O3 than other groups. School-age children spend more time outdoors engaged in
high-level activities than do other groups and breath more air in than adults relative to body
3-59
-------
Air Exchange
Rates and
Building
Volumes
Emission Rates and
Use Patterns for
Indoor Sources
(e.g., gas appliances,
passive smoking)
Ambient
Fixed-Site
Concentrations
Air Quality
Specification
Seasonal
Considerations
(Temperature)
Mass Balance
Model for Indoor
Microenvironments
Outdoor
Microenvironment
Concentrations
Human Activity
and Exertion
Patterns
(Exercise Level
Patterns)
Population and
Commuting
Data
Exposure Algorithms
Distribution of People and
Occurrences of Exposures
Linked with Breathing Rate
(Minute Ventilation Rate)
Figure 3-29. Conceptual overview of an exposure model. Model inputs (e.g., activity
patterns, ambient monitoring data, air exchange rates) are in round-corner
boxes and model calculations are shown in rectangles.
Source: Johnson etal. (1999).
surface area, breathing frequency, and heart rate. Asthmatic children spend the same amount of
time outdoors as other more healthy children but the time spent engaged in high levels of activity
are less.
Estimates of activity level have been compiled based on questionnaire data. The National
Human Activity Pattern Survey (NHAPS), a probability-based telephone survey, was conducted
in the early 1990s. The survey concluded that outdoor work-related activities were highest
during the springtime and were more frequent during the morning and early afternoon.
Exercise/sports-related activities were highest from noon to 3 p.m. during the summer months.
During the spring months, exercise/sports-related activities were highest from mid- to late
afternoon (Klepeis et al., 1996, 2001). A pilot study by Gonzales et al. (2003) evaluated the use
of retrospective questionnaires for reconstructing past time-activity and location pattern
information. Ozone concentration estimates using ambient stationary monitors and estimates
3-60
-------
derived from diaries and questionnaires differed slightly. However, both estimates were greater
than O3 personal exposure measurements.
Existing comprehensive inhalation exposure models (NEM and pNEM) (Johnson, 2003;
Johnson et al., 1996a,b,c, 1997), (MENTOR7SHEDS) Burke et al., 2001; McCurdy et al., 2000),
and the APEX model treat human activity patterns as sequences of exposure events in which
each event is defined by a geographic location and microenvironment and then assigned activity
diary records from the CHAD (Consolidated Human Activities Database;
www.epa.gov/chadnetl) (Glen et al., 1997; McCurdy, 2000; McCurdy et al., 2000). There are
now about 22,600 person-days of sequential daily activity pattern data in CHAD representing all
ages and both genders. The data for each subject consist of one or more days of sequential
activities, in which each activity is defined by start time, duration, activity type (140 categories),
and microenvironment classification (110 categories). Activities vary from 1 min to 1 h in
duration. Activities longer than 1 h are subdivided into clock-hour durations to facilitate
exposure modeling. A distribution of values for the ratio of oxygen uptake rate to body mass
(referred to as metabolic equivalents or METs) is provided for each activity type listed. A table
listing the activity patterns included in CHAD appears in AX3.
pNEM divides the population of interest into representative cohorts based on the
combinations of demographic characteristics (age, gender, and employment), home/work
district, and residential cooking fuel. APEX and MENTOPJSHEDS generate a population
demographic file containing a user-defined number of person-records for each census tract of
the population based on proportions of characteristic variables (age, gender, employment,
and housing) obtained for the population of interest, and then assigns the matching activity
information from CHAD to each individual record of the population based on the characteristic
variables.
The APEX model is capable of simulating individual movement through time and space to
provide estimates of exposure to a given pollutant in various microenvironments (e.g., indoor,
outdoor, and in-vehicle microenvironments). One of the key strengths of the APEX model is its
ability to estimate hourly exposures and doses for all simulated individuals in a sampled
population. However, APEX is limited in that uncertainties in the predicted distributions (e.g.,
age, activity data, commuting patterns, personal activities) have not been addressed. The APEX
model has not been evaluated, however, the pCNEM, a Canadian conceptional version of the
3-61
-------
NEM model, has been evaluated for estimation of PM10. Since pCNEM is similar to the
pNEM/NEM, the APEX model should work as well as pCNEM. (See discussion on pCNEM
later is this section.)
MENTOR/SHEDS is capable of simulating individuals exposures in various
microenvironments (outdoors, residence, office, school, store, restaurant, bar, and vehicles)
using spatial concentration data for each census tract. The indoor and in-vehicle pollutant
concentrations are calculated using specific equations for the microenvironment and ambient
pollutant concentration relationship. Randomly selected characteristics for a fixed number of
individual are selected to match demographics within the census tract for age, gender,
employment status, and housing type. Smoking prevalence statistics by gender and age is
randomly selected for each individual in the simulation. Diaries for activity patterns are matched
for the simulated individual by demographic characteristics (Burke et al., 2001).
Zidek et al. (2000, 2003, 2005) described a methodology for predicting human exposure to
environmental pollutants. The methodology builds on earlier models such as SHEDS and
pNEM/NEM and provides a WWW platform for developing a wide variety of models.
pCNEM, a platform model developed from this methodology, is a Canadian PC version of NEM.
pCNEM was used to estimate a conditional predictive exposure distribution for PM10 in London.
An important feature of pCNEM is it's ability to estimate the effects of reductions in ambient
levels of pollutants.
An important source of uncertainty in existing exposure modeling involves the creation of
multiday, seasonal, or year-long exposure activity sequences based on 1- to 3-day activity data
for any given individual from CHAD. Activity pattern data sets in CHAD vary for different
studies and may not be representative of the population. The commuting data are for home-to-
work and may not be representative of other commuting patterns. Correlations among human
activities that can impact microenvironmental concentrations are not captured. See Annex AX3.
Currently, appropriate longitudinal data are not available and the existing models use various
rules to derive longer-term activity sequences utilizing 24-h activity data from CHAD.
Of the above models, only NEM/pNEM have been used in the prior NAAQS review. The
pNEM probabilistic model builds on the earlier NEM deterministic exposure model. The model
takes into consideration the temporal and spatial distribution of people and O3 in the area of
consideration, variations in O3 concentrations in the microenvironment, and the effects of
3-62
-------
exercise-increased ventilation on O3 uptake. The pNEM models have been applied to nine urban
areas and a summer camp. The models used activity data from the Cincinnati Activity Diary
Study (CADS) along with time-activity data from several other studies. Data from stationary
monitoring sites were used to estimate outdoor O3 exposure. Indoor O3 decay was assumed to be
proportional to the indoor O3 concentration. An algorithm assigned the EVR associated with
each exposure event. The EVR for the outdoor children model was generated using a module
based on heart rate data by Spier et al. (1992) and Linn et al. (1992).
Characterization of Exposure
The use of ambient air monitoring stations is the most common surrogate for assigning
exposure in epidemiological studies. Since the primary source of O3 exposure is the ambient air,
monitoring concentration data would provide the exposure outdoors while exercising, a potential
important exposure to evaluate in epidemiological studies. Monitored concentrations are useful
for a relative assignment of exposure with time if the concentration were uniform across the
region; the time-activity pattern were the same across the population; and the housing
characteristics, such as ventilation rates and the O3 sinks contributing to its indoor decay rates,
were constant for the study area. Since these factors vary by population and location there will
be errors in the magnitude of the total exposure and in the relative total exposure assignment
based solely on ambient monitoring data.
Personal O3 exposure measurements have been made for potentially susceptible
populations (children, outdoor workers, the elderly, and individuals with chronic obstructive
pulmonary disease). Children and outdoor workers have somewhat higher exposures than other
individuals because they spend more time outdoors engaged in moderate and heavy exertion.
Children are also more active outside and, therefore, have a higher minute ventilation rate than
most adults (Klepeis et al., 1996, 2001). Available exposure studies suggest trends in exposure
magnitude for some populations, however, additional exposure studies are needed to generalize
differences in exposure between the general population and potentially susceptible populations.
Table 3-3 summaries the findings of available exposure studies.
Ozone concentrations in various microenvironments under a variety of environmental
conditions have been reported in the literature. In the absence of an indoor O3 source,
concentrations of O3 indoors are lower than that found in the ambient air. Ozone concentrations
3-63
-------
Table 3-3. Personal Exposure Concentrations
Personal Exposure Mean
Location, Population, Sample Duration n (range) (ppb)
San Diego, CA, Asthmatics ages 9-18 years, 12
12 hour
Vancouver, Canada, Adult Workers, Daily 585
High indoor time
Moderate indoor time
Only outdoor
12 ± 12 (0-84)
10 weekend
12 weekday
(ND-9)
(ND-12)
(2-44)
a
Reference
Delfinoetal. (1996)
Brauer and Brook
(1997)
Southern California, Subjects 10-38 years 24
Spring
Fall
Montpellier, France, Adults, Hourly 16
Winter
Summer
Southern California, Children 6-12 years, 169
>6 days
Upland - winter
- summer
Mountain - winter
- summer
Baltimore, MD, Technician, Hourlyb 1
Winter
Summer
Baltimore, MD, Adults 75 ± 7 years, Daily 20
Winter
Summer
13.6 ± 2.5 (- to 80)
10.5 ± 2.5 (- to 50)
34.3 ± 17.6 (6.5-88)
15.4 ±7.7 (6.5-40)
44.1 ±18.2(11-88)
6.2 ±4.7 (0.5-41)
19 ± 18(0.5-63)
5.7 ±4.2 (0.5-31)
25 ± 24 (0.5-72)
3.5±7.5(ND-49)
15 ± 18 (ND-76)
3.5±3.0(ND-9.9)
0. ± 1.8 (ND-2.8)
Liu etal. (1997)
Bernard etal. (1999)
Geyh et al. (2000)
Chang et al. (2000)
Sarnat et al. (2000)
aND = not detected.
bMeasurements made following scripted activities for 15 days.
in microenvironments were found to be primarily controlled by ambient O3 concentrations and
the AER: they increase with increasing AER. To a lesser extent, O3 concentrations in
microenvironments are influenced by the ambient temperature, time of day, indoor
characteristics (e.g., presence of carpeting), and the presence of other pollutants in the
microenvironment. Table 3-4 describes the findings of the available studies.
Factors Affecting Ozone Concentrations
Ozone and other photochemical oxidants are formed in the ambient air from the reaction of
sunlight with vehicle emissions, gasoline fumes, solvent vapors, and power plant and industrial
3-64
-------
Table 3-4. Indoor/Outdoor Ozone Concentrations in Various Microenvironments
Location and Ventilation
Conditions
New England States (9)
. Fall
Indoor/Outdoor
Concentrations
20 ppb/40 ppb
Comments
Schools represented a variety of environmental conditions - varying ambient O3
concentrations, sources, geographic locations, population density, traffic patterns,
building types. Average O3 concentrations were low in the morning and peaked
during the early afternoon. O3 concentrations averaged for all schools monitored.
Reference
NESCAUM
(2002)
Mexico City, School
Windows/Doors Open (27)
Windows/Doors Closed Cleaner
Off (41)
Windows/Doors Closed Cleaner
On (47)
Mexico City
Homes
Schools
Boston, MA, Homes (9)
Winter - continuously
Summer - continuously
Los Angeles, Homes (239)
0 to 247 ppb/
64 to 361 ppb
5 ppb/27ppb (7 d)
7 ppb/37 ppb (14 d)
22 ppb/73 ppb
0 to 20.4 ppb/4.4 to
24.5 ppb
0 to 34.2 ppb/8.2 to
51.8 ppb
13 ppb/37 ppb
Study conducted over 4 d period during winter months. Two-minute averaged Gold et al.
measurements were taken both inside and outside of the school every 30 min from (1996)
10 a.m. to 4 p.m. Estimated air exchange rates were 1.1, 2.1, and 2.5 h'1 for low,
medium, and high flow rates. Ozone concentrations decreased with increasing
relative humidity.
Ozone monitoring occurred between September and July. Study included 3 schools Romieu
and 145 homes. Most of the homes were large and did not have air conditioning. et al. (1998)
Ninety-two percent of the homes had carpeting, 13% used air filters, and 84% used
humidifiers. Thirty-five percent opened windows frequently, 43% sometimes, and
22% never between 10 a.m. and 4 p.m. Ozone monitored at schools sites
from 8 a.m. to 1 p.m. daily for 14 consecutive days. Homes monitored for
continuous 24-h periods for 7 and 14 consecutive days.
Study examined the potential for O3 to react with VOCs to form acid aerosols. Reiss et al.
Carbonyls were formed. No clear trend of O3 with AERs. The average AER was (1995)
0.9 IT1 during the winter and 2.6 IT1 during the summer. Four residences in winter
and nine in summer with over 24 h samples collected.
Four hundred and eighty-one samples collected inside and immediately outside of Avol et al.
home from February to December. Concentrations based on 24-h average O3 (1998b)
concentrations indoors and outdoors. Low outdoor concentrations resulted in low
indoor concentrations. However, high outdoor concentrations resulted in a range of
indoor concentrations.
Burbank, CA
Telephone Switching Station
0.2 to 1.0/1.0-21.1
ppb
Major source of O3 was transport from outdoors. From early spring to late fall O3 Weschler
concentrations peaked during the early afternoon and approach zero at sunset. et al. (1994)
AER ranged from 1.0 to 1.9 h~'.
-------
Table 3-4 (cont'd). Indoor/Outdoor Ozone Concentrations in Various Microenvironments
Location and Ventilation
Conditions
Indoor/Outdoor
Concentrations
Comments
Reference
Oi
Oi
Montpellier, France
Homes (110)
Southern CA, Homes
Upland
Mountains
Krakow, Poland, Museums
Cloth Hall
Matejko
Wawel Castle
National
Buildings, Greece
Thessalonki
Athens
Patrol cars, NC
University of CA
Photocopy room
Home/office
O3 generators
15.5/32.0 ppb
11.8/48.2 ppb
2.8/35.7 ppb
3.2/25.7-27.4 ppb
8.5/20.0 ppb
2.5/14.7 ppb
1.5/11.0 ppb
9.39/15.48 ppb
8.14/21.66 ppb
11.7/28.3 ppb
<20 to 40 ppb/—
BLD* to 290 ppb/-
Ozone measurements made over 5-d periods in and outside of 21 homes during Bernard et al.
summer and winter months. The winter I/O ratio was 0.31 compared to 0.46 during (1999)
summer months.
Ozone measurements were taken at 119 homes (57 in Upland and 62 in towns Geyh et al.
located in the mountains) during April and May. Concentrations were based on (2000);
average monthly outdoor and average weekly indoor concentrations. Indoor based Lee et al.
on the home location, number of bedrooms, and presence/absence of an (2002)
air conditioner.
Ozone continuously monitored at five museums and cultural centers. Monitoring Salmon et al.
conducted during the summer months for 21 to 46 h or 28 to 33 days at each site. (2000)
Indoor concentration found to be dependent on ventilation rate, i.e., when the
ventilation rate was high, the indoor O3 concentrations approached that of ambient
O3. Rooms sequestered from outdoor air, or where air was predominantly recycled
through charcoal filters, O3 levels indoors greatly reduced.
There was no heating/air conditioning system in the building at Thessaloniki. Drakou et al.
Windows were kept closed during the entire monitoring period. Complete air (1998)
exchange took place every 3 h. The air conditioning system in continuous use at
the Athens site recirculated the air. Complete air exchange estimated to be 1 h.
Monitoring done for 30 days at each site, but only 7 most representative days used.
Patrol cars were monitored Mon. through Thurs. between the hours of 3 p.m. to Riediker
midnight on 25 occasions during the months of Aug., Sept., and Oct. Outdoor O3 et al. (2003)
concentrations were taken from ambient monitoring station. Air inside the patrol
car was recirculated cool air.
Room volume was 40 m3. Ozone concentrations increased proportionately with Black et al.
increasing use of photocopier. (2000)
Room volume was 27 m3. Doors and windows closed. Heating/air conditioning Steiber et al.
and mechanical ventilation systems off. Ozone generators operated for 90 min. (1995)
High O3 concentrations noted when O3 generator used at high setting. AER was
0.3 h"1. Ozone concentrations varied depending on unit tested.
*BLD = Below limit of detection
-------
emissions (See Chapter 2 for a discussion of O3 atmospheric chemistry). Ozone enters the
indoor environment primarily through infiltration from outdoors through building components,
such as windows, doors, and ventilation systems. There are also a few indoor sources of O3
(photocopiers, facsimile machines, laser printers, and electrostatic air cleaners and precipitators)
(Weschler, 2000). Generally O3 emissions from office equipment and air cleaners are low
except under improper maintenance conditions. Reported O3 emissions from office equipment
range from 1300 to 7900 |ig/h (Leovic et al., 1996, 1998). Most air cleaners (particulate
ionizers) emitted no or only a small amount (56 to 2757 |ig/h) of O3 during operation (Niu et al.,
2001). Emissions from O3 generators can range from tens to thousands of micrograms per hour
(Weschler, 2000; U.S. Environmental Protection Agency, 1996).
Other photochemical oxidants (peroxyacyl nitrates; PAN and PPN) have no known direct
emission sources indoors. Although not a significant source of indoor PAN. PAN can form in
the indoor environment from the reaction of the OH- or NO3 with acetaldehyde to form the acetyl
radical, CH3CO (Grosjean et al., 1996). The acetyl radical then reacts with oxygen to for an
acetylperoxy radical which reacts with NO2 to form PAN. Peroxyacyl nitrates primarily occur in
the indoor environment from infiltration through the building envelop and through openings in
the building envelopment.
The concentration of O3 in indoor environments is dependent on the outdoor O3
concentration, the AER or outdoor infiltration, indoor circulation rate, and O3 removal processes
through contact with indoor surfaces and reactions with other indoor pollutants. Since O3
concentrations are generally higher during the warmer months, indoor concentrations will likely
be highest during that time period. (See earlier discussion on ambient concentrations of O3.)
Air exchange rates vary depending on temperature differences, wind effects, geographical
region, type of heating/mechanical ventilation system, and building type (Weschler and Shields,
2000; Colome et al., 1994). The balance of the flow of air in and out of a microenvironment is
greatest in a residential building when a window or door is open (Johnson et al., 2004; Howard-
Reed et al., 2002). The opening of windows or doors is dependent on the building occupancy,
season, housing density, the presence of air conditioning, and wind speed (Johnson and Long,
2004). When windows and doors are closed, the dominant mechanism controlling AERs is
infiltration through unintentional openings in the building envelope. Williams et al. (2003a,
2003b) reported AERs of 0.001 to 4.87 h'1 in 37 homes in Research Triangle Park, NC. Chan
3-67
-------
et al. (2005) compared air leakage measurements for 70,000 houses. Older and smaller houses
had higher normalized leakage areas than newer and larger houses. Meng et al. (2004) also
attributed higher AERs to the age of the housing stock. AERs for homes in Houston, TX and
Elizabeth, NJ were averaged for all four seasons, the highest AER, 1.22 h"1, was noted for homes
in Elizabeth, NJ where the homes were older. Evaluations of AERs for residential structures was
reported by Murray and Burmaster (1995) and includes AERs for 2,844 residential structures in
four different climatic regions by season (winter, spring, summer, and fall). The AER for all
seasons across all regions was 0.76 h"1 (arithmetic mean) (Region 1: IN, MN, MT, NH, NY1,
VT, WI; Region 2: CO, CT, IL, NJ, NY2, OH, PA, WA; Region 3: CAS, MD, OR, WA;
Region 4: AZ, CA4, FL, TX). The AERs were generally higher during the warm seasons, when
ambient O3 concentrations are highest. Data for the warmest region during the summer months
may not be representative of all homes because measurements were made in southern California
where windows are open and air conditioning is not used.
Average mean (median) AERs of 2.45 (2.24), 1.35 (1.09), and 2.22 (1.79) h"1 were
reported by Lagus Applied Technology, Inc. (1995) for schools, offices, and retail
establishments in California. Mean AERs for schools, offices, and retail establishments in
Oregon and Washington were 0.32, 0.31, and 1.12 h"1 (Turk et al., 1989)—considerably less than
that reported by Lagus Applied Technology. Park et al. (1998) reported mean AERs ranging
from 1.0 to 47.5 h"1 for stationary vehicles under varying ventilating conditions. Where
available, AERs for other studies are included in Table 3-10.
The most important removal process for O3 in the indoor environment is deposition on, and
reaction with, indoor surfaces. The rate of deposition is material-specific. The removal rate will
depend on the indoor dimensions, surface coverings, and furnishings. Smaller rooms generally
have larger surface-to-volume ratio (A/V) and remove O3 faster than larger rooms. Fleecy
materials, such as carpets, have larger surface-to-volume ratios and remove O3 faster than
smooth surfaces (Weschler, 2000). However, the rate of O3 reaction with carpet diminishes with
cumulative O3 exposure (Morrison and Nazaroff, 2000, 2002). Weschler (2000) compiled the O3
removal rates for a variety of microenvironments. Generally, the removal rates ranged between
3.0 and 4.3 kd (A/VyiT1. The highest removal rate, 7.6 kd (A/VyiT1, was noted for a clean room
(Weschler et al, 1989).
3-68
-------
Ozone chemical reactions in the indoor environment are analogous to those reactions
occurring in the ambient air (See discussion on atmospheric chemistry in Chapter 2). Ozone
reacts with unsaturated VOCs in the indoor environment, primarily terpenes or terpene-related
compounds from cleaning products, air fresheners, and wood products. The reactions are
dependent on the O3 indoor concentration, the indoor temperature and, in most cases, the air
exchange rate/ventilation rate. It is beyond the scope of this document to discuss actual
concentrations of O3 reaction products in the indoor environment; however, many of the reaction
products may more negatively impact human health and artifacts in the indoor environment than
their precursors (Wolkoff et al., 1999; Wilkins et al., 2001; Weschler et al., 1992; Weschler and
Shields, 1997; Rohr et al., 2002; N0jgaard et al., 2005). Primary reaction products are Criegee
biradicals, nitrate radicals, and peroxyacetyl radicals. Secondary reaction products are hydroxy,
alkyl, alkylperoxy, hydroperoxy, and alkoxy radicals. Reactions with alkenes can produce
aldehydes, ketones, and organic acids (Weschler and Shields, 2000; Weschler et al., 1992).
Hydroxyl radicals formed from the reaction of O3 with VOCs, nitric oxide and
hydroperoxy, and other intermediate products can react with nitrogen compounds, sulfur
dioxide, and carbon monoxide to produce significantly more toxic compounds (Sarwar et al.,
2002; Orzechowska and Paulson, 2002; Pick et al., 2003, 2004; Van den Bergh et al., 2000; Fan
et al., 2003; Wilkins et al., 2001; Clausen et al., 2001; Rohr et al., 2002, 2003; Poupard et al.,
2005; Blondeau et al., 2005). The reaction between O3 and terpenes also has been shown to
increase the concentration of indoor particles (Weschler and Shields, 1999, 2003; Weschler,
2004; Clausen et al., 2001; Fan et al, 2003; Wainman et al., 2000), possibly from further
reactions of the hydroxy radical with terpenes (Sarwar et al., 2002).
Decomposition and formation of PAN in the indoor environment are influenced by NO2
and NO. Decomposition of PAN is expected to be a relatively fast process when indoor O3
levels are low and when motor vehicle emissions are large or there is an indoor source of NOX
(Weschler and Shields, 1997).
Factors Affecting the Relationship between Ambient Concentrations and Personal
Exposures to O3
Ambient O3 concentrations vary with the time of day, season of the year, and among
locations. Personal exposure to O3 is influenced by the microenvironmental concentration and
the amount of time spent in each microenvironment. The majority of the population spends, on
3-69
-------
average, nearly 90% of their time in an indoor microenvironment. Since there are few indoor
sources of O3, O3 ambient concentration may be the most important factor that affects average
population exposure in the indoor environment.
Indoor O3 concentrations also are affected by several other factors and mechanisms.
Studies have shown that in addition to the ambient O3 concentrations, indoor O3 concentrations
are influenced by the air exchange rate or outdoor infiltration, increasing with increasing air
exchange. Once indoors, the O3 concentration is affected by the indoor circulation rate and O3
removal through contact with indoor surfaces and reactions with other indoor pollutants.
Studies on the effect of elevation on O3 concentrations found that concentrations increased
with increasing elevation (Vakeva et al., 1999; Johnson, 1997). Since O3 monitors are frequently
located on rooftops in urban settings, the concentrations measured there may overestimate the
exposure to individuals outdoors in streets and parks, locations where people exercise and their
maximum O3 exposure is more likely to occur.
In epidemiologic studies investigating acute and chronic health outcomes using ambient
monitoring data from stationary monitoring sites, O3 exposure assessment was affected by the
distance between home and the monitoring site, gender, time-activity patterns (e.g., percentage
of time spent outdoors, type of outdoor activity, time of day during outdoor activity), and indoor
air exchange rates (e.g., ventilation conditions, home characteristics) (Geyh et al., 2000; Lee
et al., 2002, 2004; Liu et al., 1995, 1997; Chang et al., 2000; Chan et al., 2005; O'Neill et al.,
2003; Brauer and Brook, 1997). Geyh et al. (2000) observed higher indoor and personal O3
concentrations in a southern California community with 2% air-conditioned homes compared to
a community with 93% air-conditioned homes during the summer (high O3) months, but showed
no difference in O3 levels during the winter (low O3) months. People that work outdoors tend to
be exposed to higher levels of O3 (Brauer and Brook, 1997; O'Neill et al., 2003). Lee et al.
(2004) observed that personal O3 exposure was positively correlated with outdoor time (r = 0.19,
p < 0.01) and negatively correlated with indoor time (r = -0.17, p < 0.01). Additional factors
that affected indoor O3 levels were air conditioning, window fans, and window opening. The O3
exposure assessment study by Liu et al. (1995) found that after adjusting for time spent in
various indoor and outdoor microenvironments (e.g., car with windows open, car with windows
closed, school, work, home, outdoors near home, outdoors other than near home), mean 12-hour
ambient O3 concentrations explained 32% of the variance in personal exposure in the summer.
3-70
-------
In a southern California study by Avol et al. (1998a), boys were found to spend more time
outdoors and be more physically active than girls. Another southern California study found that
boys were outdoors 30 minutes longer than girls and had higher personal O3 exposure during
both high and low O3 months (Geyh et al., 2000).
The announcement of smog alerts or air quality indices may influence personal exposures
to O3 by causing individuals to alter behaviors (avoidance behavior). Neidell (2004), in his
evaluation of the effect of pollution on childhood asthma, examined the relationship between the
issuance of smog alerts or air quality indices for several counties in California and hospital
admissions for asthma in children under age 18 years (not including newborns). Smog alerts are
issued in California on days when O3 concentrations exceed 200 ppb. There was a significant
reduction in the number of asthma-related hospital admissions in children ages 1 to 12 years on
smog alert days, indicating that avoidance behavior might be present on days of high O3
concentrations. Changes in population behavior as a function of concentration complicate the
estimation of health effects from population-based studies; thus, it may be desirable to include
sensitivity analyses that eliminate high O3 days, particularly in areas where avoidance behavior
is expected.
Potential Sources of Error Resulting from the Use of Ambient Ozone Concentrations in
Epidemiological Analyses
There is no clear consensus among exposure analysts as to how well stationary monitor
measurements of ambient O3 concentrations represent a surrogate for personal O3 exposure.
The microenvironmental (indirect) approach and the personal sampling (direct) approach
(Navidi et al., 1999; Ott, 1982, 1985) have been used to assess personal exposure in air pollution
epidemiologic studies; however, both approaches are associated with measurement error.
To determine personal exposure using the microenvironmental approach, the concentrations of
the various microenvironments are multiplied by the time spent in each microenvironment.
Both the concentration and time component contribute to the measurement error. There is no
time component to the measurement error in the personal sampling approach, however, the
estimation of exposure using personal monitoring devices contributes to measurement error,
especially in the case of O3. Passive badges are commonly used for monitoring O3 integrated
personal exposure. Their sensitivity to wind velocity, badge placement, and interference with
other copollutants may result in measurement error.
3-71
-------
Results from the error analysis models developed by Navidi et al. (1999) using two
different statistical designs (bidirectional case-crossover and multi-level analytic designs)
indicated that neither the microenvironmental nor personal sampling approach gave reliable
health effect estimates when measurement errors were uncorrected. The nondifferential
measurement error biased the effect estimates toward zero under the model assumptions.
However, if the measurement error was correlated with the health response, a bias away from the
null could result. The use of central ambient monitors to estimate personal exposure has a
greater potential to introduce bias since most people spend the majority of their time indoors,
where O3 levels tend to be much lower than outdoor ambient levels. If the error is of a fixed
amount (i.e., absolute differences do not change with increasing concentrations), there is no bias.
However, if the error is not a fixed difference, this error will likely attenuate the O3 risk estimate
as personal O3 exposures are generally lower than ambient O3 concentrations.
Several studies have examined relationships between measured ambient O3 concentrations
from fixed monitoring sites and personal O3 exposure (Avol et al., 1998b; Brauer and Brook,
1995, 1997; Chang et al., 2000; Delfino et al., 1996; Lee et al., 2004; Liard et al., 1999; Linn
et al., 1996; Liu et al., 1995, 1997; O'Neill et al., 2003; Sarnat et al., 2001). Two studies by
Sarnat et al. (2001, 2005) examined relationships between individual variations in personal
exposure and ambient O3 concentrations. In the first study conducted in Baltimore, MD, the
association between 24-h average ambient O3 concentrations from a centrally-located monitoring
site and 24-h average personal O3 exposures was evaluated in a cohort of older adults (n = 20),
individuals with COPD (n = 15), and children (n = 21) (Sarnat et al., 2001). Personal exposures
were measured repeatedly for each subject for a total of 800 person-days; thus, analyses of
personal exposure data were conducted using mixed models with subjects modeled as random
variables to account for between-subject variation. The mixed regression effect estimates
were P = 0.01 (95% CI: -0.01, 0.03) for the summer (196 paired samples) and P = 0.00 for the
winter (449 paired samples). However, in the second study conducted in Boston, MA with a
cohort of 20 healthy senior citizens and 23 school children, significant associations were
observed between 24-h average ambient O3 concentrations and 24-h average personal O3
exposures (Sarnat et al., 2005). The mixed regression effect estimates were P = 0.27 (95% CI:
0.18, 0.37) and P = 0.04 (95% CI: 0.00, 0.07) during the summer (332 paired samples) and
winter (288 paired samples), respectively. In the Boston study, the regression coefficients
3-72
-------
indicated that ambient O3 concentrations were predictive of personal O3 exposures; however,
ambient O3 levels overestimated personal exposures 3- to 4-fold in the summer and 25-fold in
the winter.
Important predictors for personal O3 exposure were examined (Xue et al., 2005) in the
Harvard Southern California Chronic Ozone Exposure Study. In addition to ambient O3
concentrations from central monitors, indoor and outdoor O3 concentrations at the participant's
homes, and personal O3 concentrations were measured using passive O3 samplers. During a
one-year period, 160 children from two southern California communities (Upland and the
San Bernardino mountains) were monitored for six consecutive days for at least six of the
12 months. Outdoor O3 concentrations at the participant's homes were very similar to
central ambient O3 concentrations, and personal O3 concentrations were close to indoor O3
concentrations. At different time points throughout the year, the average ratios of personal to
central O3 concentrations were relatively stable, being around 0.3 (SD 0.13). Central O3
concentrations were approximately 3 times higher compared to personal concentrations during
the O3 season (May to September) and approximately 5 times higher during the non-O3 season
(October to April). Xue et al. (2005) found that ambient O3 concentrations from central
monitors, after adjusting for time-activity patterns and housing characteristics from
questionnaire data, reasonably predicted personal O3 concentrations; a 1 ppb increase in
ambient O3 concentration was associated with a 0.54 ppb (95% CI not provided) increase in
personal O3 exposure (Pearson r = 0.76). The regression coefficient for the relationship between
ambient O3 concentrations and personal O3 exposures without adjustment for time-activity and
housing characteristics, which is of most relevance to epidemiologic time-series studies, was
not presented.
Chang et al. (2000) compared 1-h personal and ambient O3 measurements in older adults in
various microenvironments in Baltimore, MD, using activity data from the National Human
Activity Pattern Survey study (Klepeis, 1999). Activities were scripted to simulate activities
performed by older adults (65+ years of age). In total, 180 1-h personal samples were collected
in each season (summer and winter). There was no correlation between personal and ambient O3
concentrations in the indoor residence (r = 0.09 in summer and r = 0.05 in winter), although a
moderate correlation was found in other indoor environments such as restaurants, hospitals, and
shopping malls (r = 0.34 and r = 0.46 for summer and winter, respectively). In comparison, the
3-73
-------
correlation in outdoor environments (near and away from roads) was moderate to high (0.68 <
r < 0.91) and statistically significant. Regression coefficients for the relationship between
personal and ambient O3 concentrations were not reported in this study.
Brauer and Brook (1995, 1997) observed that the daily averaged personal O3 measurements
and ambient concentrations were well-correlated after stratifying groups by time spent outdoors.
Clinic workers (n = 25; 24-hour samples), teenage camp counselors (n = 25; 24-hour samples),
and farm workers (n = 15; 6-14 h work shift samples) spent 0 to 25%, 7.5 to 45%, and 100% of
their monitored time outdoors, respectively. The personal to ambient O3 concentration ratios
were significantly different for the clinic workers (0.28) and farm workers (0.96). Ambient O3
concentrations and time spent outdoors explained more of the variability in the personal O3
measurements for outdoor farm workers compared to the clinical workers. However, the
Spearman correlation coefficients were comparable, 0.60 and 0.64 for the clinic workers and
farm workers, respectively, indicating that even though the clinic workers spent considerable
amounts of time indoors or in transit there was still reasonable correlation between the day-to-
day variations in mean personal O3 exposures and mean O3 concentrations.
A study by O'Neill et al. (2003) examined 107 pairs of ambient and personal O3
measurements from 39 outdoor workers in Mexico City using a longitudinal analysis method.
Two to seven personal measurements were collected on each of the 26 monitoring days, which
were averaged and then compared with the ambient concentrations. They estimated that a 1 ppb
increase in ambient O3 concentration was associated with a 0.56 ppb (95% CI: 0.43, 0.69)
increase in personal O3 concentration. In a Paris, France study by Liard et al. (1999), adults
(n = 55) and children (n = 39) wore passive O3 monitors for 4 consecutive days during three
periods. For each period, all adults wore the O3 monitors over the same 4 days. Likewise, all
children wore monitors over the same 4 days for each of the three periods, but on different days
from the adults. The ambient O3 concentrations from the stationary monitoring sites did not
explain a high percentage of the variance of personal O3 exposure (nonsignificant [value not
stated] in adults and 21% in children). However, when personal measurements from all subjects
were aggregated for each of the six periods, the 4-day mean personal O3 exposure was found to
be highly correlated with the corresponding mean ambient concentration (r = 0.83, p < 0.05).
Similarly, a study of Los Angeles school children by Linn et al. (1996) found that daily 24-h
3-74
-------
average ambient O3 concentrations from a central site were well-correlated (r = 0.61) with daily
averaged personal O3 exposures.
The relationship between personal O3 exposures and ambient O3 concentrations from
central monitoring sites differed by city in the two studies by Sarnat et al. (2001, 2005). In the
Baltimore study (Sarnat et al., 2001) no association was observed, while in the Boston study
(Sarnat et al., 2005) ambient O3 concentrations were significantly associated with personal O3
exposures in mixed effects regression models. These results suggest that O3 concentrations
measured at central ambient monitors may explain, at least partially, the variance of individual
personal exposures; however, this relationship is influenced by factors such as air exchange rates
in housing and time spent outdoors, which may vary by city. This was further supported by
results from the southern California study by Xue et al. (2005). They found that incorporating
information on time-activity and housing characteristics allowed reasonable prediction of
personal O3 exposures from ambient O3 concentrations.
Other studies observed that the daily averaged personal O3 exposures from the population
were well correlated with monitored ambient O3 concentrations, although substantial variability
existed among the personal measurements. In other words, centrally-located ambient O3
monitors are likely to be representative of day-to-day changes in O3 exposure experienced by the
population. Brauer and Brook (1997) indicated that these results have implications for large
epidemiologic studies which often depend upon fixed-site outdoor ambient monitors to estimate
exposures. It should be noted, however, that although there are correlations between aggregate
personal and monitored ambient O3 concentrations, the absolute personal concentrations are
generally considerably lower than the monitored ambient O3 concentrations.
In summary, results indicate that the relationship between ambient O3 concentrations and
personal exposure will vary depending on individual- or city-specific factors such as time-
activity patterns, indoor air exchange rates, and housing conditions, creating potential
measurement errors. The expectations based on statistical modeling considerations are that these
exposure measurement errors or uncertainties will decrease estimates of effect in O3 health
effects analyses, making it difficult to detect a true underlying association between the exposure
and the health outcome studied. Also of special concern in interpreting results from mortality
and hospitalization time-series studies is to what extent the ambient O3 concentrations are
representative of personal O3 exposures in a particularly susceptible group of individuals, the
3-75
-------
debilitated elderly, as the correlation between the two measurements have not been examined in
this population. However, until more data on O3 exposure become available, the use of
monitored ambient O3 concentrations as a surrogate for exposures is not expected to change the
principal conclusions from O3 epidemiologic studies.
Exposure to Related Photochemical Oxidants
A variety of related photochemical oxidants produced outdoors, such as PAN and
peroxypropionyl nitrate (PPN), can infiltrate into indoor environments. These compounds are
thermally unstable and decompose to peroxacyl radicals and NO2. Exposure to related
photochemical oxidants has not been measured, nor are these compounds routinely monitored at
stationary monitoring sites. Available monitored concentrations of related photochemical
oxidants may be found in Annex AX3.
3.9 SUMMARY OF KEY POINTS
The median of the daily maximum 8-h O3 concentration averaged over May to September
2000 to 2004 was about 0.049 ppm. The daily maximum 1-h O3 concentrations could have been
much higher in large urban areas or in areas downwind of large urban areas. For example, in
Houston, TX, the daily maximum 1-h O3 concentrations have approached 0.20 ppm during this
period.
Daily maximum 8-h average O3 concentrations are lower than the maximum 1-h O3
concentrations, but they are highly correlated. Within individual MSAs, O3 concentrations tend
to be well correlated across monitoring sites. However, there can be substantial variations
in O3 concentrations. Ozone in city centers tends to be lower than in regions either upwind or
downwind because of titration by NO emitted by motor vehicles.
Ozone concentrations tend to peak in early- to mid-afternoon in areas where there is strong
photochemical activity and later in the day in areas where transport is more important in
determining the O3 abundance. Summertime maxima in O3 concentrations occur in areas in the
United States where there is substantial photochemical activity involving O3 precursors emitted
from human activities. Monthly maxima can occur anytime from June through August.
However, springtime maxima are observed in national parks, mainly in the western United States
3-76
-------
and at a number of other relatively unpolluted monitoring sites throughout the Northern
Hemisphere. For example, the highest O3 concentrations at Yellowstone NP tend to occur
during April and May. Generally, monthly minima O3 concentrations tend to occur from
November through February at polluted sites and during the fall at relatively remote sites.
Nationwide, daily maximum 8-h O3 concentrations have decreased at the upper end of the
distribution from 1990 to 2004. However, the daily maximum 8-h O3 concentrations toward the
center of the distribution have not reflected these changes. Trends have not been consistent at
national park sites; with downward trends observed at some sites and upward or no trends
observed at others. At some sites, trends reversed direction in going from the 98th to the 95th
percentile values.
Sufficient data are not available for other atmospheric oxidants (e.g., H2O2, PAN) and
oxidation products (e.g., HNO3, H2SO4) to relate concentrations of O3 to these species for use in
time series studies. Data for these species are only obtained as part of specialized field studies.
In general, secondary species, such as HNO3, H2SO4, H2O2, and PAN, are expected to be at least
moderately correlated with O3. On the other hand, primary species are expected to be more
highly correlated with each other than with secondary species, provided that the primary species
originate from common sources. Concentrations of other oxidants are much lower than for O3
and range from < 1% for oxidants in particles to several percent for gas phase species. The
relationship of O3 to PM25 is complex, because PM is not a distinct chemical species but is a mix
of primary and secondary species. PM25 concentrations were positively correlated with O3
during summer, but negatively correlated with O3 during winter at Ft. Meade, MD. PM10
concentrations show similar relations with O3.
Co-occurrences of O3 (defined when both pollutants are present at an hourly average
concentration of >0.05 ppm) with NO2 and SO2 are rare. For example, there were fewer than
10 co-occurrences with either NO2 or SO2 in 2001. The number of co-occurrences for O3 and
PM2 5 (defined as an hourly average O3 concentration >0.05 ppm and a 24-h average PM25
concentration >40 |ig/m3 occurring during the same 24-h period) also tended to be infrequent
(<10 times) at most sites, but there were up to 20 such co-occurrences at a few sites.
Policy relevant background O3 concentrations are used for assessing risks to human health
associated with O3 produced from anthropogenic sources in continental North America. Because
of the nature of the definition of PRB concentrations, they cannot be directly derived from
3-77
-------
monitored concentrations, instead they must be derived from modeled estimates. Current model
estimates indicate that ambient air PRB concentrations in the United States are generally
0.015 ppm to 0.035 ppm. They decline from spring to summer and are generally <0.025 ppm
under conditions conducive to high O3 episodes. However, PRB concentrations can be higher,
especially at elevated sites during spring, due to enhanced contributions from hemispheric
pollution and stratospheric exchange.
Ozone exposure changes as a function of time of day, season, and microenvironment.
Ambient O3 concentrations are generally higher during warmer seasons and during the weekday,
peaking during the later portion of the day. Ozone concentrations in indoor microenvironments
are generally lower than those concentrations encountered in the ambient air. There are few
indoor sources of O3. Ozone occurs in indoor microenvironments primarily through infiltration
through the building envelop and through windows, doors, and ventilation systems. The indoor
O3 concentration is dependent on the outdoor concentration, the AER, indoor circulation rate,
and removal processes. Consequently, measured and modeled exposures should take into
consideration O3 diurnal weekly and seasonal variability and varying microenvironmental
concentrations.
Once indoors, O3 reacts with indoor surfaces, including surface coverings and furnishings.
Ozone also will react with VOCs in indoor environments, primarily terpenes or terpene-related
compounds. Ozone reactions with pollutants indoors are analogous to those reactions occurring
in the ambient air, potentially exposing subjects to compounds significantly more toxic than O3.
The available approaches for measuring personal O3 exposure include the direct approach,
using a PEM, and the indirect approach, which measures or models exposure in the
microenvironments the individual encounters. Both approaches are associated with
measurement errors.
There are difficulties in identifying chemical trapping agents for PEMs that can react
with O3, and PEMs are sensitive to wind velocity, badge placement, and interference with other
copollutants. Studies using PEMs show personal O3 exposures below those concentrations
measured at stationary monitoring sites, when measurements are not adjusted for time spent
outdoors, housing characteristics, age, gender, and occupation.
The use of measured O3 concentrations from stationary ambient monitoring sites as
surrogates for personal exposure may be affected by the O3 ambient concentration, percentage of
3-78
-------
time spent outdoors, and type of outdoor activity. Epidemiologic studies investigating health
outcomes using data from stationary monitoring sites found O3 exposure to be affected by the
distance between the subjects' location and the stationary monitor, individual activity patterns,
and the O3 concentration in the microenvironment.
The use of exposure models to evaluate O3 exposure to large populations over time is
complicated by the fact that O3 is a secondary pollutant with complex nonlinear and multiscale
dynamics in space and time. The existing comprehensive inhalation exposure models (NEM,
pNEM, MENTOR/SHEDS, APEX, pCNEM) treat human activity patterns as sequences of
exposure events. Estimates of activity levels are assigned from CHAD, the Consolidated Human
Activities Database.
Ambient O3 concentrations are estimated using emissions-based mechanistic models or
ambient-data-based models. Models for estimating microenvironmental concentrations include
the empirical, mass balance, and detailed CFD models. Mass balance modeling is the most
common modeling approach to estimating concentrations in enclosed microenvironments.
The pNEM/O3 population exposure model, the model used more extensively in O3 exposure
modeling, includes a sophisticated mass balance model for indoor and vehicle
microenvironments. There are three versions of the pNEM/O3 model: the general population,
outdoor workers, and outdoor children.
Results from O3 exposure studies indicate that the relationship between ambient O3
concentrations and personal exposure/dose will vary depending on O3 concentrations and time
spent in the various microenvironments, particularly the time spent outdoors where O3
concentrations tend to be higher, and the personal activity level. Consequently, the O3
exposure/dose may differ from the concentrations measured at stationary monitoring sites.
However, until more data on O3 exposure become available, the use of monitored ambient O3
concentrations as a surrogate for exposures is not expected to change the principal conclusions
from O3 epidemiologic studies using community average health and pollution data.
3-79
-------
REFERENCES
Adams, W. C. (1993) Measurement of breathing rate and volume in routinely performed daily activities [final
report]. Sacramento, CA: California Environmental Protection Agency, Air Resources Board; contract no.
A033-205.
Altshuller, A. P.; Lefohn, A. S. (1996) Background ozone in the planetary boundary layer over the United States.
J. Air Waste Manage. Assoc. 46: 134-141.
Avol, E. L.; Navidi, W. C.; Rappaport, E. B.; Peters, J. M. (1998a) Acute effects of ambient ozone on asthmatic,
wheezy, and healthy children. Cambridge, MA: Health Effects Institute; research report no. 82.
Avol, E. L.; Navidi, W. C.; Colome, S. D. (1998b) Modeling ozone levels in and around southern California homes.
Environ. Sci. Technol. 32: 463-468.
Bernard, N. L.; Gerber, M. J.; Astre, C. M.; Saintot, M. J. (1999) Ozone measurement with passive samplers:
validation and use for ozone pollution assessment in Montpellier, France. Environ. Sci. Technol. 33: 217-222.
Bey, L; Jacob, D. J.; Yantosca, R. M.; Logan, J. A.; Field, B.; Fiore, A. M.; Li, Q.; Liu, H.; Mickley, L. J.; Schultz,
M. G. (2001) Global modeling of tropospheric chemistry with assimilated meteorology: model description
and evaluation. J. Geophys. Res. (Atmos.) 106: 23,073-23,095.
Black, D. R.; Harley, R. A.; Hering, S. V.; Stolzenburg, M. R. (2000) A new, portable, real-time monitor. Environ.
Sci. Technol. 34:3031-3040.
Blondeau, P.; lordache, V.; Poupard, O.; Genin, D.; Allard, F. (2005) Relationship between outdoor and indoor air
quality in eight French schools. Indoor Air 15: 2-12.
Brauer, M.; Brook, J. R. (1995) Personal and fixed-site ozone measurements with a passive sampler. J. Air Waste
Manage. Assoc. 45: 529-537.
Brauer, M.; Brook, J. R. (1997) Ozone personal exposures and health effects for selected groups residing in the
Fraser Valley. In: Steyn, D. G.; Bottenheim, J. W., eds. The Lower Fraser Valley Oxidants/Pacific '93 Field
Study. Atmos. Environ. 31: 2113-2121.
Burke, J. M.; Zufall, M. J.; Ozkaynak, H. (2001) A population exposure model for paniculate matter: case study
results forPM25 in Philadelphia, PA. J. Exposure Anal. Environ. Epidemiol. 11: 470-489.
Chan, C.-C.; Wu, T.-H. (2005) Effects of ambient ozone exposure on mail carriers' peak expiratory flow rates.
Environ. Health Perspect. 113: 735-738.
Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H. H. (2000) Hourly personal exposures to fine particles and
gaseous pollutants—results from Baltimore, Maryland. J. Air Waste Manage. Assoc. 50: 1223-1235.
Chen, L.-W. A. (2002) Urban fine paniculate matter: chemical composition and possible origins (dissertation).
College Park, MD: University of Maryland, Department of Chemical Physics. Available from: University
Microfilms, Ann Arbor, MI; AADAA-I3078297.
Christakos, G.; Vyas, V. M. (1998a) A composite space/time approach to studying ozone distribution over eastern
United States. Atmos. Environ. 32: 2845-2857.
Christakos, G.; Vyas, V. M. (1998b) A novel method for studying population health impacts of spatiotemporal
ozone distribution. Soc. Sci. Med. 47: 1051-1066.
Clausen, P. A.; Wilkins, C. K.; Wolkoff, P.; Nielsen, G. D. (2001) Chemical and biological evaluation of a reaction
mixture of R-(+)-limonene/ozone: formation of strong airway irritants. Environ. Int. 26: 511-522.
Code of Federal Regulations. (2000) Appendix D to part 58—Network design for state and local air monitoring
stations (SLAMS), national air monitoring stations (NAMS), and photochemical assessment monitoring
stations (PAMS). C. F. R. 40: pt. 58, app. D.
Colome, S. D.; Wilson, A. L.; Tian, Y. (1994) California residential indoor air quality study. Volume 2. Carbon
monoxide and air exchange rate: an univariate and multivariate analysis. Chicago, IL: Gas Research Institute;
report no. GRI-93/0224.3.
Cooper, O. R.; Moody, J. L. (2000) Meteorological controls on ozone at an elevated eastern United States regional
background monitoring site. J. Geophys. Res. [Atmos.] 105: 6855-6869.
Delfino, R. J.; Coate, B. D.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Koutrakis, P. (1996) Daily asthma severity in
relation to personal ozone exposure and outdoor fungal spores. Am. J. Respir. Crit. Care Med. 154: 633-641.
Drakou, G.; Zerefos, C.; Ziomas, I.; Voyatzaki, M. (1998) Measurements and numerical simulations of indoor O3
and NO-j in two different cases. Atmos. Environ. 32: 595-610.
Ebel, A.; Hass, H.; Jakobs, J. H.; Laube, M.; Memmesheimer, M.; Oberreuter, A. (1991) Simulation of ozone
intrusion caused by a tropopause fold and cut-off low. Atmos. Environ. Part A25: 2131-2144.
Fan, Z.; Lioy, P.; Weschler, C.; Fiedler, N.; Kipen, H.; Zhang, J. (2003) Ozone-initiated reactions with mixtures of
volatile organic compounds under simulated indoor conditions. Environ. Sci. Technol. 37: 1811-1821.
3-80
-------
Federal Register. (1986) Guidelines for estimating exposures. F. R. (September 24) 51: 34,042-34,054.
Pick, I; Pommer, L.; Nilsson, C.; Andersson, B. (2003) Effect of OH radicals, relative humidity, and time on the
composition of the products formed in the ozonolysis of a-pinene. Atmos. Environ. 37: 4087-4096.
Pick, I; Nilsson, C.; Andersson, B. (2004) Formation of oxidation products in a ventilation system. Atmos. Environ.
38: 5895-5899.
Fiore, A. M; Jacob, D. I; Bey, L; Yantosca, R. M; Field, B. D.; Fusco, A. C.; Wilkinson, J. G. (2002) Background
ozone over the United States in summer: origin, trend, and contribution to pollution episodes. J. Geophys.
Res. (Atmos.) 107(D15): 10.1029/2001JD000982.
Fiore, A.; Jacob, D. J.; Liu, H.; Yantosca, R. M.; Fairlie, T. D.; Li, Q. (2003a) Variability in surface ozone
background over the United States: implications for air quality policy. J. Geophys. Res. (Atmos.) 108(D24):
10.1029/2003 JD003855.
Fiore, A. M.; Jacob, D. J.; Mathur, R.; Martin, R. V. (2003b) Application of empirical orthogonal functions to
evaluate ozone simulations with regional and global models. J. Geophys. Res. (Atmos.) 108(D14):
10.1029/2002JD003151.
Fitz-Simons, T.; McCluney, L.; Rizzo, M. (2005) Analysis of 2004 ozone data for the ozone NAAQS review
[memorandum to Dr. Joseph Pinto]. Research Triangle Park, NC: U.S. Environmental Protection Agency;
August 22.
Freijer, J. I.; Bloemen, H. J. T. (2000) Modeling relationships between indoor and outdoor air quality. J. Air Waste
Manage. Assoc. 50: 292-300.
Fusco, A. C.; Logan, J. A. (2003) Analysis of 1970-1995 trends in tropospheric ozone at Northern Hemisphere
midlatitudes with the GEOS-CHEM model. J. Geophys. Res. (Atmos.) 108: 10.1029/2002JD002742.
Garratt, J. R. (1992) The atmospheric boundary layer. Cambridge, United Kingdom: Cambridge University Press.
(Houghton, J. T.; Rycroft, M. J.; Dessler, A. J., eds. Cambridge atmospheric and space science series).
Georgopoulos, P. G.; Arunachalam, S.; Wang, S. (1997a) Alternative metrics for assessing the relative effectiveness
of NOX and VOC emission reductions in controlling ground-level ozone. J. Air Waste Manage. Assoc.
47: 838-850.
Georgopoulos, P. G.; Walia, A.; Roy, A.; Lioy, P. J. (1997b) Integrated exposure and dose modeling and analysis
system. 1. Formulation and testing of microenvironmental and pharmacokinetic components. Environ. Sci.
Technol. 31: 17-27.
Geyh, A. S.; Wolfson, J. M.; Koutrakis, P.; Mulik, J. D.; Avol, E. L. (1997) Development and evaluation of a small
active ozone sampler. Environ. Sci. Technol. 31: 2326-2330.
Geyh, A. S.; Roberts, P. T.; Lurmann, F. W.; Schoell, B. M.; Avol, E. L. (1999) Initial field evaluation of the
Harvard active ozone sampler for personal ozone monitoring. J. Exposure Anal. Environ. Epidemiol.
9: 143-149.
Geyh, A. S.; Xue, J.; Ozkaynak, H.; Spengler, J. D. (2000) The Harvard Southern California chronic ozone exposure
study: assessing ozone exposure of grade-school-age children in two southern California communities.
Environ. Health Perspect. 108: 265-270.
Glen, G.; Lakkadi, Y.; Tippett, J. A.; del Valle-Torres, M. (1997) Development of NERL/CHAD: the National
Exposure Research Laboratory consolidated human activity database. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development; contract no. 68-D5-0049.
Gold, D. R.; Allen, G.; Damokosh, A.; Serrano, P.; Hayes, C.; Castillejos, M. (1996) Comparison of outdoor and
classroom ozone exposures for school children in Mexico City. J. Air Waste Manage. Assoc. 46: 335-342.
Goldstein, A. H.; Millet, D. B.; McKay, M.; Jaegle, L.; Horowitz, L.; Cooper, O.; Hudman, R.; Jacob, D. J.;
Oltmans, S.; Clarke, A. (2004) Impact of Asian emissions on observations at Trinidad Head, California,
during ITCT 2K2. J. Geophys. Res. 109(D23S17): 10.1029/2003JD004406.
Gonzales, M.; Ngo, L.; Hammond, S. K.; Tager, I. (2003) Validation of a questionnaire and microenvironmental
model for estimating past exposures to ozone. Int. J. Environ. Health Res. 13: 249-260.
Grosjean, D. (2003) Ambient PAN and PPN in southern California from 1960 to the SCOS97-NARSTO. Atmos.
Environ. 37(suppl. 2): S221-S238.
Grosjean, E.; Grosjean, D.; Fraser, M. P.; Cass, G. R. (1996) Air quality model evaluation data for organics: 3.
Peroxyacetyl nitrate and peroxypropionyl nitrate in Los Angeles air Environ. Sci. Technol. 30: 2704-2714.
Heuss, J. M.; Kahlbaum, D. F.; Wolff, G. T. (2003) Weekday/weekend ozone differences: what can we learn from
them? J. Air Waste Manage. Assoc. 53: 772-788.
Hirsch, A. I.; Munger, J. W.; Jacob, D. J.; Horowitz, L. W.; Goldstein, A. H. (1996) Seasonal variation of the ozone
production efficiency per unit NOx at Harvard Forest, Massachusetts. J. Geophys. Res. [Atmos.]
101: 12,659-12,666.
3-81
-------
Horowitz, L. W.; Liang, J.; Gardner, G. M; Jacob, D. J . (1998) Export of reactive nitrogen from North America
during summertime: sensitivity to hydrocarbon chemistry. J. Geophys. Res. [Atmos.] 103(D11):
13451-13476.
Horvath, L.; Nagy, Z.; Weidinger, T.; Artz, R.; Luke, W. T.; Valigura, R.; Pinto, J. P.; Womack, J. (1995)
Measurement effluxes of trace gases (O3, NOX, SO2, CO2, HNO3), paniculate sulfate and nitrate, water
vapour over short vegetation by gradient and eddy correlation techniques in Hungary. EGS XX. General
Assembly; April; Hamburg, Germany. Ann. Geophys. 13(suppl. 2): C490.
Horvath, L.; Pinto, J.; Weidinger, T. (2003) Estimate of the dry deposition of atmospheric nitrogen and sulfur
species to spruce forest. Idojaras (Q. J. Hung. Meteorol. Serv.) 107: 249-255.
Howard-Reed, C.; Wallace, L. A.; Ott, W. R. (2002) The effect of opening windows on air change rates in two
homes. J. Air Waste Manage Assoc. 52: 147-159.
Hudman, R. C.; Jacob, D. J.; Cooper, O. C.; Evans, M. J.; Heald, C. L.; Park, R. J.; Fehsenfeld, F.; Flocke, F.;
Holloway, J.; Hubler, G.; Kita, K.; Koike, M.; Kondo, Y.; Neuman, A.; Nowak, J.; Oltmans, S.; Parrish, D.;
Roberts, J. M.; Ryerson, T. (2004) Ozone production in transpacific Asian pollution plumes and implications
for ozone air quality in California. J. Geophys. Res. (Atmos.) 109(D23): 10.1029/2004JD004974.
Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and
meta-analysis. Epidemiology 16: 446-457.
Jaffe, D.; Price, H.; Parrish, D.; Goldstein, A.; Harris, J. (2003) Increasing background ozone during spring on the
west coast of North America. Geophys. Res. Lett. 30: 10.1029/2003GL017024.
Johnson, T. R. (1995) Recent advances in the estimation of population exposure to mobile source pollutants.
J. Exposure Anal. Environ. Epidemiol. 5: 551-571.
Johnson, T. (1997) A pilot study in Los Angeles to measure personal ozone exposures during scripted activities.
Washington, DC: American Petroleum Institute, Health and Environmental Sciences Department; API
publication no. DR 218.
Johnson, T. (2003) A guide to selected algorithms, distributions, and databases used in exposure models developed
by the Office of Air Quality Planning and Standards. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development; EPA grant no. CR827033. Available:
http://www.epa.gov/ttn/fera/data/human/report052202.pdf [9 April, 2004].
Johnson, T.; Long, T. (2004) Determining the frequency of open windows in residences: a pilot study in Durham,
North Carolina during varying temperature conditions. J. Exposure Anal. Environ. Epidemiol.:
10.1038/sj.jea.7500409.
Johnson, T.; Capel, J.; McCoy, M. (1996a) Estimation of ozone exposures experienced by urban residents using a
probabilistic version of NEM and 1990 population data. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards; contract no. 68-DO-0062.
Johnson, T.; Capel, J.; McCoy, M.; Mozier, J. W. (1996b) Estimation of ozone exposures experienced by outdoor
workers in nine urban areas using a probabilistic version of NEM. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards; contract no. 63-D-30094,
work assignment nos. 0-1 and 1-4.
Johnson, T.; Capel, J.; Mozier, J.; McCoy, M. (1996c) Estimation of ozone exposures experienced by outdoor
children in nine urban areas using a probabilistic version of NEM. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards; contract no. 63-D-30094.
Johnson, T.; Mozier, J.; Capel, J. (1997) Supplement to "Estimation of ozone exposures experienced by outdoor
children in nine urban areas using a probabilistic version of NEM (April 1996)". Research Triangle Park, NC:
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards.
Johnson, T.; Mihlan, G.; LaPointe, J.; Fletcher, K.; Capel, J. (1999) Estimation of carbon monoxide exposures and
associated carboxyhemoglobin levels in Denver residents using pNEM/CO (version 2.0) [draft]. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
March 15.
Johnson, T.; Myers, J.; Kelly, T.; Wisbith, A.; Ollison, W. (2004) A pilot study using scripted ventilation conditions
to identify key factors affecting indoor pollutant concentrations and air exchange rate in a residence. J.
Exposure Anal. Environ. Epidemiol. 14: 1-22.
Klepeis, N. E. (1999) An introduction to the indirect exposure assessment approach: modeling human exposure
using microenvironmental measurements and the recent National Human Activity Pattern Survey. Environ.
Health Perspect. Suppl. 107(2): 365-374.
3-82
-------
Klepeis, N. E.; Tsang, A. M; Behar, J. V. (1996) Analysis of the national human activity pattern survey (NHAPS)
respondents from a standpoint of exposure assessment. Washington, DC: U.S. Environmental Protection
Agency, Office of Research and Development; report no. EPA/600/R-96/074.
Klepeis, N. E.; Nelson, W. C.; Ott. W. R.; Robinson, J. P. Tsang, A. M.; Switzer, P.; Behar, J. V.; Hern, S. C.;
Engelmann, W. H. (2001) The National Human Activity Pattern Survey (NHAPS): a resource for assessing
exposure to environmental pollutants. J. Exposure Anal. Environ. Epidemiol. 11: 231-252.
Kotchenruther, R. A.; Jaffe, D. A.; Beine, H. J.; et al. (2001a) Observations of ozone and related species in the
northeast Pacific during the PHOBEA campaigns. 2. Airborne observations. J. Geophys. Res. (Atmos.)
106(07): 7463-7483.
Kotchenruther, R. A.; Jaffe, D. A.; Jaegle, L. (200Ib) Ozone photochemistry and the role of peroxyacetyl nitrate in
the springtime northeastern Pacific troposphere: results from the photochemical ozone budget of the eastern
north Pacific atmosphere (PHOBEA) campaign. J. Geophys. Res. (Atmos.) 106(D22): 28731-28742.
Koutrakis, P.; Sioutas, C.; Ferguson, S. T.; Wolfson, J. M.; Mulik, J. D.; Burton, R. M. (1993) Development and
evaluation of a glass honeycomb denuder/filter pack system to collect atmospheric gases and particles.
Environ. Sci. Technol. 27: 2497-2501.
Lagus Applied Technology, Inc. (1995) Air change rates in non-residential buildings in California. Sacramento, CA:
California Energy Commission; contract no. 400-91-034; July.
Lee, E. H.; Hogsett, W. E. (1999) Role of concentrations and time of day in developing ozone exposure indices for a
secondary standard. J. Air Waste Manage. Assoc. 49: 669-681.
Lee, M.; Heikes, B. G.; O'Sullivan, D. W. (2000) Hydrogen peroxide and organic hydroperoxide in the troposphere.
Atmos. Environ. 34: 3475-3494.
Lee, K.; Xue, J.; Geyh, A. S.; Ozkaynak, H.; Leaderer, B. P.; Weschler, C. J.; Spengler, J. D. (2002) Nitrous acid,
nitrogen dioxide, and ozone concentrations in residential environments. Environ. Health Perspect.
110: 145-150.
Lee, K.; Parkhurst, W. J.; Xue, J.; Ozkaynak, H.; Neuberg, D.; Spengler, J. D. (2004) Outdoor/indoor/personal ozone
exposures of children in Nashville, Tennessee. J. Air Waste Manage. Assoc. 54: 352-359.
Lefohn, A. S.; Runeckles, V. C. (1987) Establishing standards to protect vegetation - ozone exposure/dose
considerations. Atmos. Environ. 21: 561-568.
Lefohn, A. S.; Oltmans, S. J.; Dann, T.; Singh, H. B. (2001) Present-day variability of background ozone in the
lower troposphere. J. Geophys. Res. [Atmos.] 106: 9945-9958.
Lehman, J.; Swinton, K.; Bortnick, S.; Hamilton, C.; Baldridge, E.; Ender, B.; Cox, B. (2004) Spatio-temporal
characterization of tropospheric ozone across the eastern United States. Atmos. Environ. 38: 4357-4369.
Leovic, K. W.; Sheldon, L. S.; Whitaker, D. A.; Hetes, R. G.; Calcagni, J. A.; Baskir, J. N. (1996) Measurement of
indoor air emissions from dry-process photocopy machines. J. Air Waste Manage. Assoc. 46: 821-829.
Leovic, K.; Whitaker, D.; Northeim, C.; Sheldon, L. (1998) Evaluation of a test method for measuring indoor air
emissions from dry-process photocopiers. J. Air Waste Manage. Assoc. 48: 915-923.
Levinson, D. H., ed. (2005) State of the climate in 2004. Bull. Am. Meteorol. Soc. 86: S1-S86.
Levinson, D. H.; Waple, A. M., eds. (2004) State of the climate in 2003. Bull. Am. Meteorol. Soc. 85: S1-S72.
Li, Q. B.; Jacob, D. J.; Bey, I.; Palmer, P. I.; Duncan, B. N.; Field, B. D.; Martin, R. V.; Fiore, A. M.; Yantosca,
R. M.; Parrish, D. D.; Simmonds, P. G.; Oltmans, S. J. (2002a) Transatlantic transport of pollution and its
effects on surface ozone in Europe and North America. J. Geophys. Res. 107(D13): 10.1029/2001JD001422.
Li, Q.; Jacob, D. J.; Fairlie, T. D.; Liu, H.; Martin, R. V.; Yantosca, R. M. (2002b) Stratospheric versus
pollution influences on ozone at Bermuda: reconciling past analyses. J. Geophys. Res.
107(D22): 10.1029/2002JD002138.
Li, Q. B.; Jacob, D. J.; Yantosca, R. M.; Munger, J. W.; Parrish, D. D. (2004) Export of NOy from the North
American boundary layer: reconciling aircraft observations and global model budgets. J. Geophys. Res.
109(002313): 10.1029/2003 JD004086.
Li, Q.; Jacob, D. J.; Park, R.; Wang, Y.; Heald, C. L.; Hudman, R.; Yantosca, R. M.; Martin, R. V.; Evans, M.
(2005) North American pollution outflow and the trapping of convectively lifted pollution by upper-level
anticyclone. J. Geophys. Res. 110 (D10301): 10.1029/2004JD005039.
Liang, J.; Horowitz, L. W.; Jacob, D. J.; Wang, Y.; Fiore, A. M.; Logan, J. A.; Gardner, G. M.; Munger, J. W.
(1998) Seasonal budgets of reactive nitrogen species and ozone over the United States, and export fluxes to
the global atmosphere. J. Geophys. Res. (Atmos.) 103: 13,435-13,450.
Liard, R.; Zureik, M.; Le Moullec, Y.; Soussan, D.; Glorian, M.; Grimfeld, A.; Neukirch, F. (1999) Use of personal
passive samplers for measurement of NO2, NO, and O3 levels in panel studies. Environ. Res. 81: 339-348.
3-83
-------
Lin, C.-Y.; Jacob, D. I; Munger, J. W.; Fiore, A. M. (2000) Increasing background ozone in surface air over the
United States. Geophys. Res. Lett. 27: 3465-3468.
Linn, W. S.; Shamoo, D. A.; Hackney, J. D. (1992) Documentation of activity patterns in 'high-risk' groups exposed
to ozone in the Los Angeles area. In: Tropospheric ozone and the environment II: effects, modeling and
control: papers from an Air & Waste Management Association international specialty conference; November,
1991; Atlanta, GA. Pittsburgh, PA: Air & Waste Management Association; pp. 701-712. (A&WMA
publication TR-20).
Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996)
Short-term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal.
Environ. Epidemiol. 6: 449-472.
Liu, L.-J. S.; Koutrakis, P.; Leech, J.; Broder, I. (1995) Assessment of ozone exposures in the greater metropolitan
Toronto area. J. Air Waste Manage. Assoc. 45: 223-234.
Liu, L.-J. S.; Delfino, R.; Koutrakis, P. (1997) Ozone exposure assessment in a southern California community.
Environ. HealthPerspect. 105: 58-65.
Liu, H.; Jacob, D. J.; Chan, L. Y.; Oltmans, S. J.; Bey, I.; Yantosca, R. M.; Harris, J. M.; Duncan, B. N.; Martin, R.
V. (2002) Sources of tropospheric ozone along the Asian Pacific Rim: an analysis of ozonesonde
observations. J. Geophys. Res. [Atmos.] 107(D21): 10.1029/2001JD002005.
Martin, R. V.; Jacob, D. J.; Logan, J. A.; Bey, I.; Yantosca, R. M.; Staudt, A. C.; Li, Q. B.; Fiore, A. M.; Duncan, B.
N.; Liu, H.; Ginoux, P.; Thouret, V. (2002) Interpretation of TOMS observations of tropical tropospheric
ozone with a global model and in-situ observations. J. Geophys. Res. 107(D18) 4351:
10.1029/2001JDOO1480.
McCurdy, T. (2000) Conceptual basis for multi-route intake dose modeling using an energy expenditure approach.
J. Exposure Anal. Environ. Epidemiol. 10: 86-97.
McCurdy, T.; Glen, G.; Smith, L.; Lakkadi, Y. (2000) The National Exposure Research Laboratory's Consolidated
Human Activity Database. J. Exposure Anal. Environ. Epidemiol. 10: 566-578.
Meng, Q. Y.; Turpin, B. J.; Korn, L.; Weisel, C. P.; Morandi, M.; Colome, S.; Zhang, J.; Stock, T.; Spektor, D.;
Winer, A.; Zhang, L.; Lee, J. H.; Giovanetti, R.; Cui, W.; Kwon, J.; Alimokhtari, S.; Shendell, D.; Jones, J.;
Farrar, C.; Maberti, S. (2005) Influence of ambient (outdoor) sources on residential indoor and personal PM25
concentrations: analyses of RIOPA data. J. Exposure Anal. Environ. Epidemiol. 15: 17-28.
Morrison, G. C.; Nazaroff, W. W. (2000) The rate of ozone uptake on carpets: experimental studies. Environ. Sci.
Technol. 34: 4963-4968.
Morrison, G. C.; Nazaroff, W. W. (2002) Ozone interactions with carpet: secondary emissions of aldehydes.
Environ. Sci. Technol. 36: 2185-2192.
Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
Nazaroff, W. W.; Cass, G. R. (1986) Mathematical modeling of chemically reactive pollutants in indoor air.
Environ. Sci. Technol. 20: 924-934.
Nazaroff, W. W.; Weschler, C. J. (2004) Cleaning products and air fresheners: exposure to primary and secondary
air pollutants. Atmos. Environ. 38: 2841-2865.
Neidell, M. J. (2004) Air pollution, health, and socio-economic status: the effect of outdoor air quality on childhood
asthma. J. Health Econ. 23: 1209-1236.
Niu, J.; Tung, T. C. W.; Burnett, J. (2001) Ozone emission rate testing and ranking method using environmental
chamber. Atmos. Environ. 35: 2143-2151.
Nejgaard, J. K.; Christensen, K. B.; Wolkoff, P. (2005) The effect on human eye blink frequency of exposure to
limonene oxidation products and methacrolein. Toxicol. Lett. 156: 241-251.
Northeast States for Coordinated Air Use Management (NESCAUM). (2002) Indoor/outdoor school air monitoring
project. Boston, MA. Available: http://www.nescaum.org/pdf/schoolmonitoring.pdf [29 October, 2003].
Oltmans, S. J.; Lefohn, A. S.; Scheel, H. E.; Harris, J. M.; Levy, H., II; Galbally, I. E.; Brunke, E.-G.; Meyer, C. P.;
Lathrop, J. A.; Johnson, B. J.; Shadwick, D. S.; Cuevas, E.; Schmidlin, F. J.; Tarasick, D. W.; Claude, H.;
Kerr, J. B.; Uchino, O.; Mohnen, V. (1998) Trends of ozone in the troposphere. Geophys. Res. Lett.
25: 139-142.
O'Neill, M. S.; Ramirez-Aguilar, M.; Meneses-Gonzalez, F.; Hernandez-Avila, M.; Geyh, A. S.; Sienra-Monge, J. J.;
Romieu, I. (2003) Ozone exposure among Mexico City outdoor workers. J. Air Waste Manage. Assoc.
53: 339-346.
3-84
-------
Orzechowska, G. E.; Paulson, S. E. (2002) Production of OH radicals from the reactions of C4-C6 internal alkenes
and styrenes with ozone in the gas phase. Atmos. Environ. 36: 571-581.
Ott, W. R. (1982) Concepts of human exposure to air pollution. Environ. Int. 7: 179-196.
Ott, W. R. (1985) Total human exposure: an emerging science focuses on humans as receptors of environmental
pollution. Environ. Sci. Technol. 19: 880-886.
Park, J.-H.; Spengler, J. D.; Yoon, D.-W.; Dumyahn, T.; Lee, K.; Ozkaynak, H. (1998) Measurement of air
exchange rate of stationary vehicles and estimation of in-vehicle exposure. J. Exposure Anal. Environ.
Epidemiol. 8: 65-78.
Pinto, J. P.; Lefohn, A. S.; Shadwick, D. S. (2004) Spatial variability of PM2 s in urban areas in the United States.
J. Air Waste Manage. Assoc. 54: 440-449.
Poupard, O.; Blondeau, P.; lordache, V.; Allard, F. (2005) Statistical analysis of parameters influencing the
relationship between outdoor and indoor air quality in schools. Atmos. Environ. 39: 2071-2080.
Reeves, C. E.; Penkett, S. A. (2003) Measurements of peroxides and what they tell us. Chem. Rev. 103: 5199-5218.
Reiss, R.; Ryan, P. B.; Koutrakis, P.; Tibbetts, S. J. (1995) Ozone reactive chemistry on interior latex paint. Environ.
Sci. Technol. 29: 1906-1912.
Riediker, M.; Williams, R.; Devlin, R.; Griggs, T.; Bromberg, P. (2003) Exposure to paniculate matter, volatile
organic compounds, and other air pollutants inside patrol cars. Environ. Sci. Technol. 37: 2084-2093.
Roberts, J. M.; Williams, J.; Baumann, K.; Buhr, M. P.; Goldan, P. D.; Holloway, J.; Htibler, G.; Kuster, W. C.;
McKeen, S. A.; Ryerson, T. B.; Trainer, M.; Williams, E. J.; Fehsenfeld, F. C.; Bertman, S. B.; Nouaime, G.;
Seaver, C.; Grodzinsky, G.; Rodgers, M.; Young, V. L. (1998) Measurements of PAN, PPN, and MPAN
made during the 1994 and 1995 Nashville Intensives of the Southern Oxidant Study: implications for regional
ozone production from biogenic hydrocarbons. J. Geophys. Res. [Atmos.] 103: 22,473-22,490.
Roberts, J. M.; Flocke, F.; Stroud, C. A.; Hereid, D.; Williams, E.; Fehsenfeld, F.; Brune, W.; Martinez, M.; Harder,
H. (2002) Ground-based measurements of peroxycarboxylic nitric anhydrides (PANs) during the 1999
Southern Oxidants Study Nashville intensive. J. Geophys. Res. [Atmos.] 107(D21): 10.1029/2001JD000947.
Roberts, J. M.; Jobson, B. T.; Kuster, W.; Goldan, P.; Murphy, P.; Williams, E.; Frost, G.; Riemer, D.; Apel, E.;
Stroud, C.; Wiedinmyer, C.; Fehsenfeld, F. (2003) An examination of the chemistry of peroxycarboxylic
nitric anhydrides and related volatile organic compounds during Texas Air Quality Study 2000 using
ground-based measurements. J. Geophys. Res. [Atmos.] 108(D16): 10.1029/2003JD003383.
Roberts, J. M.; Flocke, F.; Chen, G.; de Gouw, J.; Holloway, J. S.; Htibler, G.; Neuman, J. A.; Nicks, D. K., Jr.;
Nowak, J. B.; Parrish, D. D.; Ryerson, T. B.; Sueper, D. T.; Warneke, C.; Fehsenfeld, F. C. (2004)
Measurement of peroxycarboxylic nitric anhydrides (PANs) during the ITCT 2K2 aircraft intensive
experiment. J. Geophys. Res. 109(D23S21): 10.1029/2004JD004960.
Roelofs, G. J.; Scheeren, H. A.; Heland, J.; Ziereis, H.; Lelieveld, J. (2003) A model study of ozone in the eastern
Mediterranean free troposphere during MINOS (August 2001). Atmos. Chem. Phys. 3: 1199-1210.
Rohr, A. C.; Wilkins, C. K.; Clausen, P. A.; Hammer, M.; Nielsen, G. D.; Wolkoff, P.; Spengler, J. D. (2002) Upper
airway and pulmonary effects of oxidation products of (+)-a-pinene, J-limonene, and isoprene in B ALB/c
mice. Inhalation Toxicol. 14: 663-684.
Rohr, A. C.; Shore, S. A.; Spengler, J. D. (2003) Repeated exposure to isoprene oxidation products causes enhanced
respiratory tract effects in multiple murine strains. Inhalation Toxicol. 15: 1191-1207.
Romieu, L; Lugo, M. C.; Colome, S.; Garcia A. M.; Avila, M. H.; Geyh, A.; Velasco, S. R.; Rendon, E. P. (1998)
Evaluation of indoor ozone concentration and predictors of indoor-outdoor ratio in Mexico City. J. Air Waste
Manage. Assoc. 48: 327-335.
Salmon, L. G.; Cass, G. R.; Bruckman, K.; Haber, J. (2000) Ozone exposure inside museums in the historic central
district of Krakow, Poland. Atmos. Environ. 34: 3823-3832.
Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal paniculate and gaseous
exposures of senior citizens living in Baltimore, MD. J. Air Waste Manage. Assoc. 50: 1184-1198.
Sarnat, J. A.; Schwartz, J.; Catalano, P. J.; Suh, H. H. (2001) Gaseous pollutants in paniculate matter epidemiology:
confounders or surrogates? Environ. Health Perspect. 109: 1053-1061.
Sarnat, J. A.; Brown, K. W.; Schwartz, J.; Coull, B. A.; Koutrakis, P. (2005) Ambient gas concentrations and
personal paniculate matter exposures: implications for studying the health effects of particles. Epidemiology
16: 385-395.
Sarwar, M.; Corsi, R.; Kimura, Y.; Allen, D.; Weschler, C. (2001) Hydroxyl radicals in indoor environments. In:
Proceedings of the Air & Waste Management Association's 94th Annual Conference & Exhibition; June;
Orlando, FL. Pittsburgh, PA: Air & Waste Management Association.
3-85
-------
Sarwar, G.; Corsi, R.; Kumura, Y.; Allen, D.; Weschler, C. J. (2002) Hydroxyl radicals in indoor environments.
Atmos. Environ. 36: 3973-3988.
Serensen, D. N.; Weschler, C. J. (2002) Modeling-gas phase reactions in indoor environments using computational
fluid dynamics. Atmos. Environ. 36: 9-18.
Sparks, J. P.; Roberts, J. M; Monson, R. K. (2003) The uptake of gaseous organic nitrogen by leaves: a significant
global nitrogen transfer process. Geophys. Res. Lett. 30(23): 10.1029/2003GL018578.
Spier, C. E.; Little, D. E.; Trim, S. C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992) Activity patterns in
elementary and high school students exposed to oxidant pollution. J. Exposure Anal. Environ. Epidemiol.
2: 277-293.
Steiber, R. S. (1995) Ozone generators in indoor air settings. Research Triangle Park, NC: U.S. Environmental
Protection Agency, National Risk Management Research Laboratory; report no. EPA-600/R-95-154.
Available from: NTIS, Springfield, VA; PB96-100201.
Teklemariam, T. A.; Sparks, J. P. (2004) Gaseous fluxes of peroxyacetyl nitrate (PAN) into plant leaves. Ecol. Soc.
Am. Ann. Meeting Abst. 89: 501-502.
Trainer, M.; Parrish, D. D.; Buhr, M. P.; Norton, R. B.; Fehsenfeld, F. C.; Anlauf, K. G.; Bottenheim, J. W.; Tang,
Y. Z.; Wiebe, H. A.; Roberts, J. M.; Tanner, R. L.; Newman, L.; Bowersox, V. C.; Meagher, J. F.; Olszyna,
K. J.; Rodgers, M. O.; Wang, T.; Berresheim, H.; Demerjian, K. L.; Roychowdhury, U. K. (1993) Correlation
of ozone with NOy in photochemically aged air. J. Geophys. Res. [Atmos.] 98: 2917-2925.
Turk, B. H.; Grimsrud, D. T.; Brown, J. T.; Geisling-Sobotka, K. L.; Harrison, J.; Prill, R. J. (1989) Commercial
building ventilation rates and particle concentrations. ASHRAE Trans. 95(part 1): 422-433.
U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA;
PB87-142949.
U.S. Environmental Protection Agency. (1992) Guidelines for exposure assessment. Washington, DC: Risk
Assessment Forum, USEPA 600Z-92/001.
U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
http://cfpub2.epa.gov/ncea/.
U.S. Environmental Protection Agency. (1998) Guideline on ozone monitoring site selection. Research Triangle
Park, NC: U.S. Environmental Protection Agency, National Center for Environmental Assessment;
EPA-454/R-98-002.
U.S. Environmental Protection Agency. (2003) Technology Transfer Network: Air Quality System (AQS).
Washington, DC: Office of Air and Radiation. Available: http://www.epa.gov/ttn/airs/airsaqs/ [24 August,
2005].
U.S. Environmental Protection Agency. (2004a) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
http://cfpub.epa.gov/ncea/ [9 November, 2004].
U.S. Environmental Protection Agency. (2004b) The ozone report: measuring progress through 2003. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/K04-001. Available:
http://www.epa.gov/air/airtrends/pdfs/2003ozonereport.pdf [12 May, 2005].
Vakeva, M.; Hameri, K.; Kulmala, M.; Lahdes, R.; Ruuskanen, J.; Laitinen, T. (1999) Street level versus rooftop
concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon. Atmos.
Environ. 33: 1385-1397.
Van den Bergh, V.; Vanhees, I.; De Boer, R.; Compernolle, F.; Vinckier, C. (2000) Identification of the oxidation
products of the reaction between a-pinene and hydroxyl radicals by gas and high-performance liquid
chromatography with mass spectrometric detection. J. Chromatogr. A 896: 135-148.
Venkatachari, P.; Hopke, P. K.; Grover, B. D.; Eatough, D. J. (2005a) Measurement of particle-bound reactive
oxygen species inRubidoux aerosols. J. Atmos. Chem. 50: 49-58.
Venkatachari, P.; Hopke, P .K.; Brune, W. H.; Ren, X.; Lesher, R.; Mao, J.; Mitchell, M. (2005b) Characterization
of reactive oxygen species trends in Flushing, New York. Atmos. Environ.: in press.
Volz-Thomas, A.; Geiss, H.; Hofzumahaus, A.; Becker, K.-H. (2003) Introduction to special section: photochemistry
experiment in BERLIOZ. J. Geophys. Res. [Atmos.] 108(D4): 10.1029/JD002029.
Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of submicron
particle exposure. Environ. Health Perspect. 108: 1139-1145.
3-86
-------
Wang, Y.; Logan, J. A.; Jacob, D. J. (1998) Global simulation of tropospheric O3-NOx-hydrocarbon chemistry 2.
Model evaluation and global ozone budget. J. Geophys. Res. (Atmos.) 103: 10,727-10,755.
Weinstein-Lloyd; et al. (1998) Measurements of peroxides and related species during the 1995 summer intensive of
the Southern oxidants study in Nashville, Tennessee. J. Geophys. Res. (Atmos.) 103: 22361-22373.
Weschler, C. J. (2000) Ozone in indoor environments: concentration and chemistry. Indoor Air 10: 269-288.
Weschler, C. J. (2004) Chemical reactions among indoor pollutants: what we've learned in the new millennium.
Indoor Air 14(suppl. 7): 184-194.
Weschler, C. J.; Shields, H. C. (1997) Potential reactions among indoor pollutants. Atmos. Environ. 31: 3487-3495.
Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
Environ. 33:2301-2312.
Weschler, C. J.; Shields, H. C. (2000) The influence of ventilation on reactions among indoor pollutants: modeling
and experimental observations. Indoor Air. 10: 92-100.
Weschler, C. J.; Shields, H. C. (2003) Experiments probing the influence of air exchange rates on secondary organic
aerosols derived from indoor chemistry. Atmos. Environ. 37: 5621-5631.
Weschler, C. J.; Shields, H. C.; Naik, D. V. (1989) Indoor ozone exposures. JAPCA 39: 1562-1568.
Weschler, C. J.; Hodgson, A. T.; Wooley, J. D. (1992) Indoor chemistry: ozone, volatile organic compounds, and
carpets. Environ. Sci. Technol. 26: 2371-2377.
Weschler, C. J.; Shields, H. C.; Naik, D. V. (1994) Indoor chemistry involving O3, NO, and NO2 as evidenced by
14 months of measurements at a site in southern California. Environ. Sci. Technol. 28: 2120-2132.
Wilkins, C. K.; Clausen, P. A.; Wolkoff, P.; Larsen, S. T.; Hammer, M.; Larsen, K.; Hansen, V.; Nielsen, G. D.
(2001) Formation of strong irritants in mixtures of isoprene/ozone and isoprene/ozone/nitrogen dioxide.
Environ. Health Perspect. 109: 937-941.
Williams, R.; Suggs, J.; Rea, A.; Leovic, K.; Vette, A.; Croghan, C.; Sheldon, L.; Rodes, C.; Thornburg,, J.;
Ejire, A.; Herbst, M.; Sanders, W., Jr. (2003a) The Research Triangle Park paniculate matter panel study: PM
mass concentration relationships Atmos. Environ. 37: 5349-5363.
Williams, R.; Suggs, J.; Rea, A.; Sheldon, L.; Rodes, C.; Thornburg, J. (2003b) The Research Triangle Park
paniculate matter panel study: modeling ambient source contribution to personal and residential PM mass
concentrations. Atmos. Environ. 37: 5365-5378.
Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Hougaard, K. S.; Nielsen, G. D. (1999) Formation of strong airway
irritants in a model mixture of (+)-a-pinene/ozone. Atmos. Environ. 33: 693-698.
Xue, J.; Liu, S. V.; Ozkaynak, H.; Spengler, J. D. (2005) Parameter evaluation and model validation of ozone
exposure assessment using Harvard Southern California Chronic Ozone Exposure Study data. J. Air Waste
Manage. Assoc. 55: 1508-1515.
Zanis, P.; Trickl, T.; Stohl, A.; Wernli, H.; Cooper, O.; Zerefos, C.; Gaeggeler, H.; Schnabel, C.; Tobler, L.; Kubik,
P. W.; Priller, A.; Scheel, H. E.; Kanter, H. J.; Cristofanelli, P.; Forster, C.; James, P.; Gerasopoulos, E.;
Delcloo, A.; Papayannis, A.; Claude, H. (2003) Forecast, observation and modelling of a deep stratospheric
intrusion event over Europe. Atmos. Chem. Phys. 3: 763-777.
Zidek, J. V.; Meloche, J.; Le, N. D.; Sun, L. (2000) Combining statistical and computer models for health nsk
assessment (exposure analysis). In: Nunez-Anton, V.; Feffeirra, E., eds. Proceedings of New Trends in
Statistical Modelling, the 15th international workshop on statistical modelling; Bilbao, Spain. Stat. Model.:
pp. 95-106.
Zidek, J. V.; Meloche, J.; Shaddick, G.; Chatfield, C.; White, R. A. (2003) A computational model for estimating
personal exposure to air pollutants with application to London's PM10 in 1997. Research Triangle Park, NC:
Statistical and Applied Mathematical Sciences Institute; TR 2003-03. Available:
http://www.samsi.info/TR/tr2003-03.pdf [26 January, 2006].
Zidek, J. V.; Shaddick, G.; White, R.; Meloche, J.; Chatfield, C. (2005) Using a probabilistic model (pCNEM) to
estimate personal exposure air pollution. Environmetrics 16: 481-493.
3-87
-------
4. DOSIMETRY, SPECIES HOMOLOGY, SENSITIVITY,
AND ANIMAL-TO-HUMAN EXTRAPOLATION
4.1 INTRODUCTION
The dosimetry of ozone (O3) in humans has been examined in a series of studies published
in the past decade. These studies further characterize the dose of O3 delivered to various sites in
the respiratory tract (RT). Ozone, classified as a reactive gas, interacts with surfactant,
antioxidants, and other compounds in the epithelial lining fluid (ELF). Researchers have
attempted to obtain a greater understanding of how these complex interactions affect O3 uptake
and O3-induced injury. New work has also been completed evaluating species differences in
responses to O3 exposures, which allow more accurate quantitative extrapolation from animals
to humans.
This chapter is not intended to be a complete overview of O3 dosimetry and animal-to-
human comparisons, but rather, it is an update of the dosimetry/extrapolation chapter from the
last O3 criteria document (U.S. Environmental Protection Agency, 1996), or 1996 O3 AQCD, and
other reviews of the earlier published literature. The framework for presenting this chapter is
first a discussion in Section 4.2 of general concepts of the dosimetry of O3 in the RT. Bolus-
response studies are then presented in Section 4.2.1 followed by general uptake studies in
Section 4.2.2. Dosimetry modeling is presented in Section 4.2.3 followed by the summary and
conclusions for the dosimetry material in Section 4.2.4. The chapter continues in Section 4.3
with a discussion of species comparisons and ends with a discussion of animal-to-human
extrapolation. More detailed discussions of the studies are presented in the supporting material
to this chapter (Annex AX4). The toxicological effects of O3 in laboratory animals and in vitro
test systems are discussed in Chapter 5 and direct effects of O3 in humans are discussed in
Chapter 6. The historical O3 literature is very briefly summarized in this chapter, providing a
very concise overview of previous work. The reader is referred to the 1996 O3 AQCD for more
detailed discussion of the literature prior to the early 1990s.
4-1
-------
4.2 DOSIMETRY OF OZONE IN THE RESPIRATORY TRACT
Ozone dosimetry refers to the measurement or estimation of the amount of O3 or its
reaction products reaching and persisting at specific sites in the RT following an exposure. The
compound most directly responsible for toxic effects may be the inhaled gas O3 or one of its
chemical reaction products. Complete identification of the actual toxic agents and their
integration into dosimetry is a complex issue that has not been resolved. Dosimetric studies
attempt to quantify the amount of O3 retained in the lung (i.e., not exhaled) or the dose of O3 or
its active metabolites (e.g., aldehydes or peroxides) delivered to target cells or tissues (i.e., dose
per cell or tissue surface area). For comparison, epidemiologic studies may simply consider
exposure concentration while clinical studies may consider the total amount of O3 inhaled
(product of exposure concentration, duration, and minute ventilation). Hence, dosimetric studies
seek to accurately quantify dose to target lung regions or tissues, whereas epidemiologic and
clinical studies typically consider exposures.
Figure 4-1 illustrates the structure of the lower airways with progression from the large
airways to the alveolus. Understanding dosimetry as it relates to O3-induced injury is complex
due to the fact that O3 interacts primarily with the ELF, which contains surfactant and
antioxidants. Reactive products created by O3 can diffuse within the lung or be transported out
of the lung to generate systemic effects. Antioxidant enzymes are the primary cellular defense
against reactive species created by O3. The level and type of antioxidants varies between
species, regions of the RT itself, and can be altered by O3 exposure.
A considerable number of dosimetric studies were summarized in the 1996 O3 AQCD.
These studies provided estimates of absorbed O3 in the RT as a whole or in regions such as the
upper airways (URT) or lower airways (LRT), defined as being proximal or distal to the tracheal
entrance, respectively. Estimates were obtained for both humans and animals via direct
measurement and mathematical modeling. The mathematical models also estimated O3 doses to
specific target sites such as the proximal alveolar region (PAR; first generation distal to the
terminal bronchioles), which is also referred to as the centriacinar region (CAR; junction of
conducting airways and gas exchange region) in some studies.
4-2
-------
a.
EP
BM
Bronchus
Bronchiolus
Aveolus
b.
Air
Liquid
Tissue
Tissue
Air
Air
Figure 4-1. Structure of lower airways with progression from the large airways to the
alveolus. Panel (a) illustrates basic airway anatomy. Structures are epithelial
cells, EP; basement membrane, BM; smooth muscle cells, SM; and
fibrocartilaginous coat, FC. Panel (b) illustrates the relative amounts of liquid,
tissue, and blood with distal progression. In the bronchi there is a thick liquid
lining over a relatively thick layer of tissues. Even highly soluble materials
moving from the air into the liquid layer have minimal systemic access via the
blood. With distal progress, the protective liquid lining diminishes allowing
increased access of compounds crossing the air-liquid interface to the tissues
and the blood.
Source: Panel (a) reproduced with permission (Weibel, E. R. [1980] Design and structure of the human lung.
In: Fishman, A. P., ed. Pulmonary Diseases and Disorders. New York, NY: McGraw-Hill; p. 23 1).
-------
In general, the consensus of experimental and modeling studies summarized in the
1996 O3 AQCD supported the following conclusions: (1) for the URT, animal and human
studies suggested that O3 uptake is greater in the nose than the mouth but the effect of flow on
uptake was equivocal; (2) for the LRT, predicted tissue doses (O3 flux to liquid-tissue interface)
were very low in the trachea, increased to a maximum in the terminal bronchioles or first airway
generation in the pulmonary region, and rapidly decreased with distal progression; (3) increasing
tidal volume (VT) increases O3 uptake, whereas, increasing flow or breathing frequency (fB)
decreases O3 uptake; (3) increasing flow shifts O3 uptake to the smaller peripheral airways, i.e.,
toward the CAR; and (4) similarly, the effect of exercise is to significantly increase the
pulmonary region total dose (mass of O3) and the CAR dose (mass per unit surface area).
Some cross-species in vivo comparisons were described in the 1996 O3 AQCD.
For instance, comparing bronchoalveolar lavage (BAL) cells from rats and humans, it was
estimated that a 0.4 ppm O3 exposure in exercising humans gave 4 to 5 times the O3 dose
(retained) relative to rats exposed at rest to the same concentration. In vitro dosimetry studies in
the 1996 O3 AQCD using isolated lung preparations showed that uptake efficiency is chemical-
reaction dependent, indicating the importance of reaction product formation. These reaction
products, created mainly by the ozonolysis of polyunsaturated fatty acids, included hydrogen
peroxide, aldehydes, and hydroxyhydroperoxides, which are mediators of O3 toxicity. Other
products are created by the reaction of O3 with other ELF constituents, all of which must be
considered in understanding the dosimetry of O3.
The next two sections (4.2.1 and 4.2.2) review the available new experimental studies
on O3 dosimetry. Table AX4-1 in Annex AX4 summarizes the new human studies.
4.2.1 Bolus-Response Studies in Humans
The bolus-response method has been used by the Ultman group as an approach to explore
the distribution of O3 absorption in the airways of humans. This noninvasive method consists
of an injection of a known volume and concentration of O3 at a predetermined point during
inspiration. Ozone uptake is the amount of O3 absorbed during a breath relative to the amount
contained in the inhaled bolus. Figure 4-2 illustrates the uptake of a series of O3 boli as a
4-4
-------
i.u -
c
o
'•§
5
"- 0.5-
0)
1
Q.
D
0.0 -
v,
JA vPso% v
/
• w
1 «*
• *
• $t
*•
*
:
* •
.
f •* *
t .%**
20 40 60 80 100 120 140 160 180 200
Penetration Volume (ml_)
Figure 4-2. Ozone uptake fraction as a function of volumetric penetration (VP) in a
representative subject. Each point represents the O3 uptake of a bolus inspired
by the subject. The volumes, VUA and VD, are the volume of the upper airways
and anatomical dead space, respectively, and VP50% is the VP at which 50% of
the inspired bolus was absorbed. In 47 healthy subjects (24 M, 23 F), Ultman
et al. (2004) found that VP500/0 was well correlated with VD (r = 0.57, p < 0.001)
and better correlated with the volume of the conducting airways, i.e., VD minus
VUA, (r = 0.65, p = 0.001).
Source: Adapted from Ultman et al. (2004).
function of volumetric penetration (VP), i.e., the volume between the center of mass of an
inhaled bolus and the end of inspiration. The inspired O3 boli (for which the uptake fractions are
illustrated in Figure 4-2) were 20 mL of 2 ppm O3. Kabel et al. (1994) have previously shown
that varying the O3 concentration of inspired boli between 0.4 and 4 ppm does not affect the
distribution of uptake as a function of VP.
The O3 bolus-technique was used by Bush et al. (1996a) to ascertain differences in lung
anatomy and gender that can alter the exposure-dose cascade. Forced vital capacity (FVC), total
lung capacity (TLC) and anatomic dead space (VD) were determined for ten male and ten female
subjects. Differences between subjects in absorption of a 20 mL O3 bolus injected into the
inhaled breath could be explained by differences in VD. In particular, they concluded that the
4-5
-------
intrinsic mass transfer parameter (Ka) was proportional to the ratio of the respiratory flow to VD.
In a subsequent study, Ultman et al. (2004) showed that the volume at which 50% of an
inspired O3 bolus is absorbed was better associated with the volume of the lower conducting
airways than VD (see Figure 4-2). Bush et al. (1996a) pointed out that the applicability of their
results may be limited because of their assumptions that Ka was independent of location in the
RT and that there was no mucous resistance. They further suggested that the dependence
of Ka on flowrate and VD be restricted to flowrates < 1000 mL/s until studies at higher rates
have been performed.
Nodelman and Ultman (1999) demonstrated that the uptake distributions of O3 boli were
sensitive to the mode of breathing and to the airflow rate. As flowrates increased from 150 to
1000 mL/s, O3 penetrated deeper into the lung and penetration was further increased by oral
relative to nasal breathing. The authors suggest that the switch from nasal to oral breathing
coupled with increases in respiratory flow as occurs during exercise causes a shift in the O3 dose
distribution, allowing O3 to penetrate deeper into the lung, increasing the potential for damage to
bronchiolar and alveolar tissues.
More recently, Ultman et al. (2004) measured O3 uptake using the bolus technique in
60 young heathy nonsmoking adults (32 M, 28 F). Bolus were inspired at a rate of 1 L/s,
equivalent to a moderate exercise rate with a minute ventilation of 30 L/min. Figure 4-2
illustrates the O3 uptake fraction as a function of VP in a representative subject. Anatomic dead
space was measured in 47 of the subjects (24 M, 23 F). In these subjects, the volume at which
50% of an inhaled bolus was absorbed (VP50o/0) was correlated with VD (r = 0.57, p < 0.001) and
the volume of the lower conducting airways, i.e., VD minus the volume of the upper airways,
(r = 0.65, p = 0.001). The better correlation found by subtracting off the upper airways from VD
can be explained by the fact that very little O3 is absorbed in the upper airways during oral
breathing at 1 L/s. Both VP50o/0 and lower airways volume were greater in males than females.
These findings suggest that in females the smaller airways, and associated larger surface-to-
volume ratio, enhance local O3 uptake and cause reduced penetration of O3 into the distal lung.
It is not clear from these findings, however, if the actual anatomical location of VP500/0 differed
between males and females.
4-6
-------
A few studies have measured the effect of a continuous pollutant exposure on O3 bolus
uptake. Asplund et al. (1996) randomly exposed young healthy adults (8 M, 3 F) for 2 h
[presumably at rest] to 0.0 (air), 0.12, or 0.36 ppm O3 on 3 separate occasions separated by at
least 1-wk. Ozone bolus uptake was measured preexposure and subsequently at 30 min intervals
during the exposure. Ozone uptake over the VP range of 70 to 120 mL increased after the air
exposure (0.045, absolute change in absorbed fraction), decreased slightly after the 0.12 ppm O3
exposure (-0.005), and decreased more substantially following the 0.36 ppm O3 exposure
(-0.03). For clarification, these absolute changes in uptake due to filtered air or O3 exposures
are increases or decreases from an average uptake of-0.70 over the VP range from 70 to
120 mL (Hu et al., 1994). Relative to uptake during the air exposure, Asplund et al. (1996)
found O3 bolus uptake was significantly decreased by 30 min of the 0.12 and 0.36 ppm O3
exposures and remained significantly decreased for the duration of these exposures.
Using a similar protocol, Rigas et al. (1997) randomly exposed young healthy adults (6 M,
6 F) for 2 h at rest to filtered air, 0.36 ppm NO2, 0.75 ppm NO2, 0.36 ppm SO2, or 0.36 ppm O3.
Ozone bolus uptake (VP range of 70 to 120 mL ) was measured preexposure and every 30 min
during the exposures. The results of an F test indicated that exposure duration (30-, 60-, 90-,
120-min) was not a significant factor, but treatment (NO2, SO2, etc.) was (p < 0.01). Ozone
bolus uptake was increased by 30 min during the NO2 and SO2 exposures and decreased during
the O3 exposure. The authors suggested that there may be increased production of an O3-reactive
substrate in the ELF due to airway inflammation. During NO2 and SO2 exposures the substrate
was not depleted by these gases and so could react with the O3 bolus. During O3 exposure the
substrate was depleted, causing the fractional absorption of the O3 bolus to decrease.
4.2.2 General Uptake Studies
Ultman and colleagues have recently completed some general uptake studies to determine
the ratio of O3 uptake to the quantity of O3 inhaled. Uptake efficiency was determined at
exposures of 0.2 or 0.4 ppm O3 while exercising at a minute volume of approximately 20 L/min
for 60 min or 40 L/min for 30 min in both men and women (Rigas et al., 2000). Uptake
efficiency ranged from 0.56 to 0.98 and had a statistically significant but weak dependence on
4-7
-------
concentration, minute volume, and exposure time. Intersubject differences had the largest
influence on uptake efficiency, resulting in a variation of approximately 10%. As the quantity
of O3 retained by the RT is equal to uptake efficiency times the quantity of O3 inhaled, relatively
large changes in concentration, minute volume, or exposure time may result in relatively large
changes in the amount of O3 retained by the RT or absorbed locally. The authors concluded that
for exposure times <2 h, inhaled dose (product of O3 concentration, exposure duration, and
minute ventilation) is a reasonable predictor of actual uptake as long as there are fixed
concentrations of O3 and fixed levels of exercise. More importantly, similarly exposed
individuals vary in the amount of actual dose received.
Santiago et al. (2001) studied the effects of airflow rate (3 to 15 L/min) and O3
concentration (0.1, 0.2, or 0.4 ppm) on O3 uptake in nasal cavities of males and females.
As would be expected, uptake efficiency in the nose was inversely related to the flowrate and the
concentration of O3 in the inlet air. They computed a gas-phase diffusion resistance of <24% of
overall diffusion resistance, which suggested to them that simultaneously occurring diffusion
and chemical reactions in the mucous layer were the limiting factors in O3 uptake. Difference
in O3 uptake ranged from 0.63 to 0.97 at flowrates of 3 L/min and 0.25 to 0.50 at 15 L/min.
The small effects of flowrate and concentration on uptake efficiency were statistically
significant, but intersubject differences accounted for approximately half of the total variation
in uptake efficiency. Both these general uptake studies, done at environmentally relevant O3
concentrations, indicate that interindividual differences in fractional uptake are extremely
important in O3 dose-response relationships.
In the research mentioned above, Ultman et al. (2004) also completed continuous exposure
studies. The same 60 subjects were exposed continuously for 1 h to either clean air or 0.25 ppm
O3 while exercising at a target minute ventilation of 30 L/min. This is the first study to assess
ventilatory and dosimetric parameters for an entire hour of exposure. In addition to measuring
pre-to-post exposure changes in FEVj, they used the peripheral bronchial cross-sectional area
available for diffusion (inferred from the alveolar slope of CO2 expirograms) as an alternative
response variable. At a fixed minute ventilation of 30 L/min, the uptake fraction of O3 decreased
with increasing fB (see Figure 4-3) and increased with increasing VT. The uptake fraction was
4-8
-------
i.u -
c 0.9 -
0)
"o
it
LJJ
OJ
« 0.8 -
a.
3
0.7 -
++ + °
+ ++ + + o
+ , +
+ + <$ _p $
T
"*" + + + + O
O
O + -|-O
+
a1- o
O 0
o
o
O 28 Women
+ 32 Men 0
i i i i
20 30 40 50
Breathing Frequency (bpm)
Figure 4-3. Ozone uptake efficiency as a function of breathing frequency at a minute
ventilation of 30 L/min. The uptake efficiency was well correlated with
breathing frequency (r = -0.723, p < 0.001) and tidal volume (not illustrated;
r = 0.490, p < 0.001).
Source: From Ultman et al. (2004).
significantly greater in males (91.4%) than females (87.1%), which is consistent with the
larger fB and smaller VT of the females than males. There was a small but statistically significant
reduction in the breath-by-breath uptake of O3 from 90.6% on average for the first 15 min to
87.3% on average for the last 15 min of exposure, although the biological significance of this
small change is questionable. Ozone uptake rate correlated with percent changes in individual
bronchial cross-sectional area but did not correlate with individual FEVj responses. Neither of
these parameters correlated with the VP50o/0 determined in the bolus studies mentioned above.
The authors concluded that the intersubject differences in forced respiratory responses were not
due to differences in O3 uptake. However, these data did partially support the hypothesis that
changes in cross-sectional area available for gas diffusion are related to overall O3 retention.
4-9
-------
Plopper et al. (1998) examined the relationship between O3 dose and epithelial injury in
Rhesus monkeys. Using 18O content in lung tissues, the respiratory bronchioles were confirmed
as the site receiving the greatest O3 dose (mass 18O per dry lung weight). Furthermore, the
greatest cellular injury occurred in the vicinity of the respiratory bronchioles and was dependent
on the delivered O3 dose to these tissues. After a 2 h exposure, the antioxidant glutathione
(GSH) was increased in the proximal intrapulmonary bronchus after 0.4 ppm O3 and decreased
in the respiratory bronchiole after 1.0 ppm O3. Perhaps an adaptive response, chronic O3
exposure leads to increased GSH levels in distal bronchioles of both rats and monkeys relative to
GSH levels in filtered air-exposed animals (Duan et al., 1996).
4.2.3 Dosimetry Modeling
When all of the animal and human in vivo O3 uptake efficiency data are compared, there is
a good degree of consistency across data sets, which raises the level of confidence with which
these data sets can be used to support dosimetric model formulations. Models predict that the
net O3 dose (O3 flux to air-liquid interface) gradually decreases distally from the trachea toward
the end of the TB and then rapidly decreases in the pulmonary region. However, the tissue
dose (O3 flux to liquid-tissue interface) is low in the trachea, increases to a maximum in the
terminal bronchioles and the first generation of the pulmonary region, and then decreases rapidly
distally into the pulmonary region. The increased VT and flow, associated with exercise in
humans or CO2-stimulated ventilation increases in rats, shifts O3 dose further into the periphery
of the lung, causing a disproportionate increase in distal lung dose.
Table AX4-2 in the annex presents a summary of new theoretical studies of the uptake
of O3 by the RTs (or regions) of humans and laboratory animals that have become available
since the 1996 review. These studies are briefly described below. Virtually all of these models
have assumed that the reaction rate of O3 in the liquid lining layer and in tissues is quasi first-
order with respect to O3 concentration. However, there is considerable discrepancy between rate
constants used in models. For instance, Hu et al. (1994) and Bush et al. (2001) estimate a
reaction rate constant that is more than 1000 times as large as that used by Cohen-Hubal et al.
(1996). Both the uptake of O3 at the gas-liquid interface and the fraction of O3 (or its reaction
4-10
-------
products) that reach epithelial cells are sensitive to the value of these reaction rate constants.
A large disparity in rate constants between studies illustrates limitations in our current state of
knowledge and affects the interpretation of O3 model predictions.
Overton and Graham (1995) created a rat model combining multiple path anatomic models
and one-dimensional convection-dispersion equations, which simulates transport and uptake
of O3 in airways and airspaces of the modeled TB region. Predictions from this model
realistically detail O3 transport and uptake of different but morphologically equivalent sites.
Modeled lung tissue doses show variation of O3 dose among anatomically equivalent ventilatory
units as a function of path length from the trachea with shorter paths generally showing greater
doses (Overton and Graham, 1995). This conflicts with Schelegle et al. (2001), who exposed
rats to 1 ppm O3 for 8 h and found that the terminal bronchioles supplied by short and long paths
had similar epithelial injury. Interestingly, in rats with a C-fiber conduction block to prevent O3-
induced rapid shallow breathing, it was the long path terminal bronchioles that received the
greatest epithelial injury. Overall, O3-induced rapid shallow breathing appears to protect the
large conducting airways while producing a more even distribution of injury to the terminal
bronchioles (load et al., 2000; Schelegle et al., 2001). Postlethwait et al. (2000) have also
identified the conducting airways as a primary site of acute O3-induced cell injury. Such data
must be considered when developing models that attempt to predict site-specific locations of O3-
induced injury. The early models computed relationships between delivered regional dose and
response with the assumption that O3 was the active agent responsible for injury. It is now
known that reactive intermediates such as hydrohydroxyperoxides and aldehydes are important
agents mediating the response to O3 (further discussed in Section 5.3.1). Thus, models must
consider O3 reaction/diffusion in the ELF and ELF-derived reactions products.
Using computational fluid dynamics (CFD), Cohen-Hubal et al. (1996) modeled the effect
of the mucus layer thickness in the nasal passage of a rat. Predictions of overall uptake were
within the range of measured uptake. Predicted regional O3 flux was correlated with measured
cell proliferation for the CFD simulation that incorporated two regions, each with a different
mucus thickness.
4-11
-------
With a RT dosimetry model, Overton et al. (1996) investigated the sensitivity of uptake
efficiency, proximal alveolar region (PAR) dose (O3 mass per unit surface area per unit time),
and PAR dose ratio to TB region volume (VTB) and TB region expansion in humans and rats.
The PAR was defined as the first generation distal to terminal bronchioles and the PAR dose
ratio was defined as the ratio of a rat's predicted PAR dose to a human's predicted PAR dose.
This ratio relates human and rat exposure concentrations so that both species receive the same
PAR dose. In rats, the PAR is a region of major damage from O3. For each species, three values
of Vra were used: a mean value from the literature and the mean ± twice the SD. For both the
rat and human simulations, there were several general findings: (1) uptake efficiency and
PAR dose both increased with decreasing VTB, e.g., using the highest TB region mass transfer
coefficient (kTB), the PAR dose for V^ - 2SD was five times greater than the PAR dose
for V-re + 2SD, (2) uptake efficiency and PAR dose both decreased with TB expansion relative
to no expansion, (3) PAR dose increased with tidal volume, (4) PAR dose increased with
decreasing k^, and (5) uptake efficiency increased with k^.
Bush et al. (2001) modified their single-path model (Bush et al., 1996b) so that simulations
would coincide with experimental uptake efficiency data for O3 and C12 during oral and nasal
breathing. Relative to their original model, the Bush et al. (2001) model added lung expansion
and modified the mass transfer coefficients for both the gas-phase (kg) and the liquid-phase (k,).
Consistent with Overton et al. (1996), considering expansion of the TB airways reduced uptake
efficiency versus no expansion. As very little inhaled O3 reaches the peripheral lung, it was not
surprising that alveolar expansion had minimal affect on uptake efficiency. Ignoring the O3
reaction rate constant (k,.), the simulations for O3 and C12 were nearly the same since the gas-
phase diffusion coefficients of O3 and C12 are similar. But for a given VP the TB airways of the
lung, experimental bolus uptake are always less for O3 than for C12. The authors surmised that
the difference between the uptake for these gases could be explained adequately based solely on
the diffusive resistance of O3 in airways surface fluid (modeled by k,.). Qualitatively, model
simulations also agreed well with the experimental data of Gerrity et al. (1995).
4-12
-------
Age- and gender-specific differences in both regional and systemic uptake in humans was
modeled using a physiologically-based pharmacokinetic (PBPK) approach (Sarangapani et al.,
2003). The model estimated that regional (URT, TB, pulmonary) extraction efficiency of O3 is
relatively insensitive to age and gender.
A recent attempt was made (Mudway and Kelly, 2004) to model O3 dose-inflammatory
response using a meta-analysis of 23 exposures in published human chamber studies. The O3
concentrations ranged from 0.08 to 0.6 ppm and the exposure durations ranged from 60 to
396 min. The analysis showed linear relationships between O3 dose and neutrophilia in
bronchoalveolar lavage fluid (BALF). Linear relationships were also observed between O3 dose
and protein leakage into BALF, which suggested to the authors that a large-scale study could
determine a possible O3 threshold level for these inflammatory responses. These recent findings
seem consistent with the linear relationship between O3 dose to pulmonary tissues normalized
for body weight and lavage fluid protein in rats, guinea pigs, and rabbits (Miller et al., 1988).
4.2.4 Summary and Conclusions - Dosimetry
Ozone is a highly reactive gas and powerful oxidant with a short half-life. Uptake occurs
in mucous membranes of the RT where O3 reacts with components of the ELF. Uptake
efficiency is chemical-reaction dependent and the reaction products (hydrogen peroxide,
aldehydes, and hydroxyhydroperoxides) created by ozonolysis of polyunsaturated fatty acids
mediate O3 toxicity. The 1996 O3 AQCD reported that uptake of O3 in rats is about 0.50 and in
humans at rest is about 0.8 to 0.95. In humans, about 0.07 of the O3 is removed in the
larynx/trachea, about 0.50 in the head, and about 0.43 in the lungs, where the primary site of
damage was believed to be the CAR. Increasing flow shifted O3 uptake distally toward smaller
airways of the lung. Studies in humans showed that increasing minute ventilation with exercise
(by increasing both breathing frequency and tidal volume) causes only a small decrease in
uptake efficiency by the total RT. The nasal passages appeared to absorb more O3 than the oral
passages. Comparing BAL cells, a 0.4 ppm exposure in exercising humans showed 4 to 5 times
the retained dose of O3 relative to rats exposed at rest to the same concentration.
4-13
-------
Most new research on O3 uptake has been performed in humans. Bolus-response studies
demonstrated that a previous continuous exposure to O3 decreases the absorption of a bolus
of O3, probably due to depletion of compounds able to absorb O3. Continuous exposure to NO2
and SO2 increased absorption of a bolus of O3. These data are of some relevance to
environmental exposures where humans may receive differing concentrations of O3 depending
on time of day. Verifying prior work, the bolus-response method was used to demonstrate
that O3 bolus uptake is sensitive to the mode of breathing and to the airflow rate. As flow is
increased from 150 to 1000 mL/s, O3 boli penetrated deeper into the lung and penetration was
further increased by oral versus nasal breathing. This suggests that the switch from nasal to oral
breathing coupled with increases in respiratory flow as occurs during exercise causes a shift in
regional O3 dose deeper into the lung, increasing the potential of damage to bronchiolar and
alveolar tissues. The finding that O3 uptake is inversely related to airflow also agrees with
earlier animal studies.
New general uptake study data demonstrate that exercising men and women receiving
0.2 or 0.4 ppm O3 at 20 L/min for 60 min or 40 L/min for 30 min absorb 0.56 to 0.98. The
absorbed fraction was affected only by large changes in concentration, minute volume, and
exposure time. This suggests that for exposure times <2 h, inhaled dose (i.e., product of O3
concentration, minute ventilation, and exposure duration) is a reasonable predictor of actual O3
dose as long as the O3 exposure concentration and level of exercise are relatively constant.
However, individuals exposed to similar concentrations vary considerably in the amount of
actual dose received. This intersubject variability has also been demonstrated in a study of O3
uptake in nasal cavities of men and women. The absorbed fraction in the nose was inversely
related to the flowrate and the concentration of O3 suggesting that both gas phase diffusion and
chemical reactions in the mucous layer were limiting O3 uptake.
The consistency of uptake data generated in animal and human studies allow a high level
of confidence in their use in dosimetry modeling. Early models predicted that net O3 dose to
ELF and tissue gradually decreases distally from the trachea toward the end of the TB and then
rapidly decreases in the pulmonary region. Exercise-induced or CO2-stimulated increases in VT
and flow, shift O3 dose further into the periphery of the lung, causing a disproportionate increase
4-14
-------
in distal lung dose. Localized damage to lung tissue has been modeled showing variation of O3
dose among anatomically equivalent ventilatory units as a function of path length from the
trachea with shorter paths showing greater damage.
New models have produced some refinements of earlier models such as: (1) the use of
mucus resistance and thickness in describing O3 dosimetry and determining the patterns
of O3-induced lesions; (2) the shape of the dose versus generation plot along any path from the
trachea to alveoli is independent of path, with the tissue dose decreasing with increasing
generation index; (3) simulations sensitive to conducting airway volume but relatively
insensitive to characteristics of the respiratory airspace; (4) the importance of TB region
expansion; (5) the importance of dose received in the PAR both inter-individual differences and
extrapolations based on dose; and (6) revaluation of mass transfer coefficients for conducting
airways. Additionally, more recent data indicate that the primary site of acute cell injury occurs
in the conducting airways and that reactive intermediates in the ELF, rather than O3 itself, are
responsible for pulmonary injury. These data must be considered when developing new models.
4.3 SPECIES HOMOLOGY, SENSITIVITY, AND ANIMAL-TO-HUMAN
EXTRAPOLATION
Basic similarities exist across human and other animals species with regard to basic
anatomy, physiology, biochemistry, cell biology, and disease processes. However, there are
obviously some species differences that have the potential to affect both the patterns of O3
uptake in the RT as well as responses. For instance, primates are oronasal breathers with a
dichotomous branching lung structure, whereas, rodents are obligate nasal breathers with a
monopodial branching lung structure (Miller et al., 1993). Even when comparing nasal
breathing, differences in the nasal structure between primates and rodents can affect both the site
and amount of gaseous uptake in this region (DeSesso, 1993; Morgan et al., 1989). Cellular
profiles also differ between species as a function of location in the RT (Miller et al., 1993;
Plopper et al., 1989; Stone et al., 1992).
4-15
-------
The homology as it exists creates similarities in acute O3-induced effects, especially in the
RT and in lung defense mechanisms. Rodents appear to have a slightly higher tachypneic
response to O3, which is clearly concentration-dependent in most species and shows parallel
exacerbation when hyperventilation (e.g., exercise or CO2) is superimposed. What is not known
is whether this is evidence of pulmonary irritant sensitivity, perhaps as a prelude to toxicity, or
whether tachypnea is a defensive action taken by the respiratory system to minimize distal lung
O3 deposition. Airway or lung resistance in humans is not affected appreciably by acute
exposure to O3, except under conditions of heavy exercise; animals appear to need high-level
exposures or special preparations that bypass nasal scrubbing. Dynamic lung compliance (Cdyn)
has been shown to have small magnitude decreases in response to O3 in some studies across
species, but it is thought that these changes are of little biological significance for ambient
exposures. Spirometric changes, the hallmark of O3 response in humans, occur in rats, but to a
lesser degree. It is unclear, however, the degree to which anesthesia (rat) and the comparability
of hyperventilation induced by CO2 (rat) or exercise (human) may influence this difference in
responsiveness. Collectively, the acute functional response of laboratory animals to O3 appears
quite homologous to that of the human.
When humans are exposed to O3 repeatedly for several consecutive days, lung function
decrements subside, and normal spirometric parameters are regained (see Section 6.6). This
phenomenon of functional attenuation also has been demonstrated in rats, not only in terms of
spirometry, but also in terms of the classic tachypneic ventilatory response. Full or partial
attenuation of some BAL parameters also appears to occur in both rats and humans, but exposure
scenario appears to play a role; other cellular changes do not attenuate (see Section 6.9.4).
Existing epidemiologic studies provide only suggestive evidence that persistent or progressive
deterioration in lung function is associated with long-term oxidant-pollutant exposure (see
Chapter 7). With chronic, repeated exposures to >0.12 ppm O3, however, laboratory animals
demonstrate changes in lung structure, function, and biochemistry that are indicative of airway
irritation and inflammation with the possible development of chronic lung disease (U.S.
Environmental Protection Agency, 1996). Based on the apparent homology of these responses
4-16
-------
between humans and laboratory animals, animal studies appear to provide a means for assessing
such chronic health concerns.
For similarly exposed animals (i.e., same O3 concentration and exposure duration), intra-
and inter-species differences in pulmonary responses are observed as a function of animal age,
ventilation, and antioxidant status. Examination of BAL constituents show that the influx of
inflammatory cells and protein from the serum is influenced by species, but perhaps to less
extent than by ventilation and antioxidant status. Based on lavage protein levels (mg protein
per mL lavage fluid), Hatch et al. (1986) reported that guinea pigs were the most responsive
(to >0.2 ppm); rabbits were the least responsive (2.0 ppm only); and rats, hamsters, and mice
were intermediate (effects at 0.5 tol.O ppm). Recognizing that there are differences in O3 doses
to tissues between species, Miller et al. (1988) examined the relationship protein levels reported
by Hatch et al. (1986) and predicted pulmonary tissue dose (mass O3 per body weight). Miller
et al. (1988) found that protein levels in guinea pigs increased more rapidly with tissue dose than
in rats and rabbits.
A species' susceptibility to the effects of O3 exposure may be due, in part, to biochemical
differences among species. Evidence for this is provided by differences in activity of SD rat and
rhesus monkey CYP moonoxygenases elicited by O3 exposure (Lee et al., 1998). Additional
characterization of species- and region-specific CYP enzymes will create a better understanding
of the differences in response to O3. This will allow more accurate extrapolation from animal
exposures to human exposures and toxic effects.
Antioxidant metabolism varies widely among species, which can greatly influence the
effects of O3 (discussed in greater detail in 5.2.1.3). The guinea pig appears to be the species
most susceptible to O3. Early studies ranked mice > rats > guinea pigs in order of antioxidant
responsiveness to O3 challenge. Guinea pigs have been shown to have lower basal levels of
GSH transferase activity, lower activity of GSH peroxidases, and lower levels of vitamin E
compared to rats. However, differences in the levels of antioxidants between species and
regions of the lung do not appear to be the primary factor determining susceptibility to O3-
induced tissue injury (Duan et al., 1993, 1996). Plopper et al. (1998) concluded that in monkeys
there was a close association between site-specific O3 dose, the degree of epithelial injury, and
4-17
-------
reduced-glutathione depletion. Within a species, antioxidant defenses against O3 can also vary
with animal age (Servias et al., 2005) and exposure history (Duan et al., 1996).
Because cytokine and chemokine responses are so important in an animal's defense
against O3 exposure, comparisons of differences in species expression and activity of these
inflammatory mediators is necessary. Arsalane et al. (1995) compared guinea pig and human
AM recovered in BALF and subsequently exposed in vitro to 0.1 to 1 ppm for 60 min.
Measurement of inflammatory cytokines showed a peak at 0.4 ppm in both species. Guinea pig
AM had an increase in IL-6 and TNFa while human AM had increases in TNFa, IL-lb, IL-6 and
IL-8. This exposure also caused an increase in mRNA expression for TNFa, IL-lb, IL-6 and
IL-8 in human cells. At 0.1 ppm exposures, only TNFa secretion was increased. These data
suggest both qualitatively and quantitatively similar cytokine responses in AM from guinea pigs
and humans. However, these in vitro AM responses can not be extended directly to an in vivo
scenario since similar O3 exposures (concentration and duration) do not give the same O3 doses
in different species (Hatch et al., 1994; Miller et al., 1978).
Species differences in morphological responses to O3 exposure have been characterized by
Dormans et al. (1999), as discussed in previous sections. Dormans et al. (1999) continuously
exposed rats, mice, and male guinea pigs to filtered air, 0.2, or 0.4 ppm O3 for 3, 7, 28, and
56 days. The animals exposed for 28 days were examined at 3, 7, or 28 days PE. Depending
on the endpoint studied, the species varied in sensitivity. Greater sensitivity was shown in the
mouse as determined by biochemical endpoints, persistence of bronchiolar epithelial
hypertrophy, and recovery time. Guinea pigs were more sensitive in terms of the inflammatory
response though all three species had increases in the inflammatory response after three days that
did not decrease with exposure. These data on inflammation are in general agreement with
Hatch et al., (1986), discussed above. In all species, the longest exposure to the highest O3
concentration caused increased collagen in ductal septa and large lamellar bodies in Type II
cells, but that response also occurred in rats and guinea pigs at 0.2 ppm. No fibrosis was seen at
the shorter exposure times and the authors question whether fibrosis occurs in healthy humans
after continuous exposure. The authors do not rule out the possibility that some of these
4-18
-------
differences may be attributable to differences in total inhaled dose or dose actually reaching a
target site. Overall, the authors rated mice as most susceptible, followed by guinea pigs and rats.
Comparisons of airway effects in rats, monkeys and ferrets resulting from exposures of
1.0 ppm O3 for 8 h (Sterner-Kock et al. 2000) demonstrated that monkeys and ferrets had similar
inflammatory responses and epithelial necrosis. The response of these two species was more
severe than that seen in rats. These data suggest that ferrets are a good animal model for O3-
induced airway effects due to the similarities in pulmonary structure between primates and
ferrets. However, the mechanisms of O3 effects at these high concentrations may differ from
those at more realistic levels.
A number of species, including nonhuman primates, dogs, cats, rabbits, and rodents, have
been used to study the effects of O3 exposure on airway bronchoconstriction. A commonly used
model of bronchospasm utilizes guinea pigs acutely exposed to high O3 concentrations (2 to
3 ppm) to induce airway hyperreactivity (AHR). As mentioned earlier, the model is helpful for
determining mechanistic aspects of AHR, but is not really relevant for extrapolation to potential
airway responses in humans exposed to ambient levels of O3. Additionally, guinea pigs have
been shown to have AHR in other studies that is very similar to asthmatic humans, but the utility
of guinea pig data is somewhat limited by their disparity from other animal models.
The rat is a key species used in O3 toxicological studies, but the rat has both behavioral
and physiological mechanisms that can lower core temperature in response to acute exposures,
thus limiting extrapolation of rat data to humans. Iwasaki et al. (1998) evaluated cardiovascular
and thermoregulatory responses to O3 at exposure of 0.1, 0.3, and 0.5 ppm O3 8 hrs/day for
4 consecutive days. A dose-dependent disruption of HR and Tco was seen on the first and second
days of exposure, which then recovered to control values. Watkinson et al. (2003) exposed
rats to 0.5 ppm O3 and observed this hypothermic response, which included lowered HR,
lowered Tco, and increased inflammatory components in BALF. The authors suggested that the
response is an inherent reflexive pattern that can possibly attenuate O3 toxicity in rodents. They
discuss the cascade of effects created by decreases in Tco, which include: (1) lowered metabolic
rate, (2) altered enzyme kinetics, (3) altered membrane function, (4) decreased oxygen
consumption and demand, (5) reductions in minute ventilation, which would act to limit the dose
4-19
-------
of O3 delivered to the lungs. These effects are concurrent with changes in HR which lead to:
(1) decreased CO, (2) lowered BP, and (3) decreased tissue perfusion, all of which may lead to
functional deficits. The hypothermic response has not been observed in humans except at very
high exposures, which complicates extrapolation of effects in rats to humans.
The importance of animal studies derives from their utilization in determining cause-effect
relationships between exposure and health outcome, but the animal data must be integrated with
epidemiological studies and controlled human clinical studies. Animal studies can corroborate
both clinical and epidemiology studies and further provide important data that is impossible to
collect in human studies. Toxic pulmonary and extrapulmonary effects following O3 exposure
have been well-studied in rodents, nonhuman primates, and a few other species; so,
extrapolation, both qualitative and quantitative, to human exposures and consequent health
effects is possible. Quantitative extrapolation, required to determine what specific exposure is
likely to cause an effect in humans, is theoretically founded on the equivalency of mechanisms
across species. At the molecular level, O3 acts on the carbon-carbon double bond in
polyunsaturated fatty acids and on sulfhydryl groups in proteins, both of which are found within
cell membranes in animals and humans. At higher levels of cellular organization, cells affected
in animals (e.g., AMs, Type 1 cells) have similar functions in humans, and organ systems (e.g.,
respiratory system) have major interspecies similarities. However, interspecies differences do
occur and complicate extrapolation.
Quantitative extrapolation, which involves a combination of dosimetry and species
sensitivity, still requires more research before it can be fully realized. Knowledge of dosimetric
animal-to-human extrapolation is more advanced than that of species-sensitivity, but
extrapolation models have not been completely validated, and therefore, significant uncertainties
remain. Mathematical modeling of O3 deposition in the lower RT (i.e., from the trachea to
alveoli) of several animal species and humans shows that the pattern of regional dose is similar,
but that absolute values differ. In spite of structural and ventilatory differences between species,
the greatest predicted tissue dose is to the CAR. Even though the CAR of rats has very
rudimentary respiratory bronchioles, compared to well-developed ones in primates, the CAR of
both rats and nonhuman primates respond similarly to O3.
4-20
-------
Experimental measurement of delivered O3 doses estimate that total respiratory uptake is
-47% in laboratory animals and -87% in exercising humans, while nasopharyngeal removal is
-17% in rats and -40% in humans. The previous O3 AQCD (U.S. Environmental Protection
Agency, 1996) provided the first quantitative animal-to-human extrapolation of morphological
changes in the proximal alveolar region using rat and monkey studies. The extrapolation
predicted that a 9-year-old child would have a 20% or 75% increase in PAR tissue thickness if
their sensitivity to O3 was equal to that of a rat or monkey, respectively. Adults would have
15 or 70% increase, suggesting the potential for chronic effects in humans. In spite of the
significant uncertainties, this extrapolation raises concern about the potential for chronic effects
in humans
Experiments using 2 h exposures to 0.4 ppm 18O3 suggested that exercising (15 min
intervals, rest and exercise at 60 L/min) humans received a 4- to 5-fold higher 18O3
concentrations in BAL than resting rats (Hatch et al., 1994). That level of exposure increased
BAL protein and PMNs in humans, while a concentration of 2.0 ppm in rats was necessary for
similar effects. Caveats in the interpretation of 18O3 studies include: (1) only a very small
portion of the labeled compound is recoverable to assess incorporation; and (2) if species being
compared differ in physiocochemical factors controlling mass transfer and downstream O3
metabolism, it could cause significant differences in the amount of inhaled 18O3 that is detected
during subsequent tissue analysis. Further, species differences in pulmonary anatomy,
ventilation, antioxidants, and susceptibility all influence dose, repair processes, and tolerance to
subsequent O3 exposure. Important differences between exercising humans and resting rats that
can affect tissue O3 dose include: (1) increased ventilation and O3 delivery with exercise;
(2) decreased pulmonary ventilation and body temperature during O3 exposure in rats;
(3) diminished dose received in rats due to their burying their noses in their fur during exposure;
and (4) increased concentration of antioxidants in ELF in rats compared to humans. These
antioxidants are important for converting O3 to inactive products before toxicity occurs (Kari
et al., 1997; Gunnison and Hatch, 1999; Plopper et al., 1998), though this quenching is not
quantitative. These and possibly other differences between rats and humans suggest that a
2 ppm exposure in nonexercising rats approximates a 0.4 ppm exposure in exercising humans.
4-21
-------
Further comparisons of exercising human exposure to 0.1 ppm for 6 hours (Devlin et al., 1991)
and resting rat exposure to 0.3 ppm show inflammatory and permeability changes in humans but
not rats.
4.3.1 Summary and Conclusions: Species Homology, Sensitivity, and
Animal-to-Human Extrapolation
Comparisons of acute exposures in rats and humans suggest that, though both species have
similar qualitative responses to O3 exposure, there are interspecies mechanistic disparities that
necessitate careful comparisons of dose-response relationships. There is no perfect nonhuman
species with which to model O3 toxicity. All have limitations that must be considered when
attempting to extrapolate to human exposures. Awareness of these limitations, even at the level
of subtle strain differences within a test species, is extremely important. The currently available
data suggest that LOELs in resting rats are approximately 4- to 5-fold higher than for exercising
humans for toxicological endpoints including BAL protein and BAL PMNs. Studies comparing
species-specific differences in O3-induced effects showed that guinea pigs were the most
susceptible, rabbits the least susceptible, and rodents intermediate in susceptibility. The recent
work being done utilizing various mouse strains with differing sensitivities to O3 will help us to
understand the extremely complex inter-individual differences in human sensitivity to O3.
4-22
-------
REFERENCES
Arsalane, K.; Gosset, P.; Vanhee, D.; Voisin, C.; Hamid, Q.; Tonnel, A.-B.; Wallaert, B. (1995) Ozone stimulates
synthesis of inflammatory cytokines by alveolar macrophages in vitro. Am. J. Respir. Cell Mol. Biol.
13:60-68.
Asplund, P. T.; Ben-Jebria, A.; Rigas, M. L.; Ultman, J. S. (1996) Longitudinal distribution of ozone absorption in
the lung: effect of continuous inhalation exposure. Arch. Environ. Health 51: 431-438.
Bush, M. L.; Asplund, P. T.; Miles, K. A.; Ben-Jebria, A.; Ultman, J. S. (1996a) Longitudinal distribution of O3
absorption in the lung: gender differences and intersubject variability. J. Appl. Physiol. 81: 1651-1657.
Bush, M. L.; Raybold, T.; Abeles, S.; Hu, S.-C.; Ben-Jebria, A.; Ultman, J. S. (1996b) Longitudinal distribution of
ozone absorption in the lung: simulation with a single-path model. Toxicol. Appl. Pharmacol. 140: 219-226.
Bush, M. L.; Zhang, W.; Ben-Jebria, A.; Ultman, J. S. (2001) Longitudinal distribution of ozone and chlorine in the
human respiratory tract: simulation of nasal and oral breathing with the single-path diffusion model.
Toxicol. Appl. Pharmacol. 173: 137-145.
Cohen-Hubal, E. A.; Kimbell, J. S.; Fedkiw, P. S. (1996) Incorporation of nasal-lining mass-transfer resistance into a
CFD model for prediction of ozone dosimetry in the upper respiratory tract. Inhalation Toxicol. 8: 831-857.
DeSesso, J. M. (1993) The relevance to humans of animal models for inhalation studies of cancer in the nose and
upper airways. Qual. Assur. (San Diego) 2: 213-231.
Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in
the lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
Dormans, J. A. M. A.; VanBree, L.; Boere, A. J. F.; Marra, M.; Rombout, P. J. A. (1999) Interspecies differences in
time course of pulmonary toxicity following repeated exposure to ozone. Inhalation Toxicol. 11: 309-329.
Duan, X.; Buckpitt, A. R.; Plopper, C. G. (1993) Variation in antioxidant enzyme activities in anatomic
subcompartments within rat and rhesus monkey lung. Toxicol. Appl. Pharmacol. 123: 73-82.
Duan, X.; Buckpitt, A. R.; Pinkerton, K. E.; Ji, C.; Plopper, C. G. (1996) Ozone-induced alterations in glutathione in
lung subcompartments of rats and monkeys. Am. J. Respir. Cell Mol. Biol. 14: 70-75.
Gerrity, T. R.; Biscardi, F.; Strong, A.; Garlington, A. R.; Brown, J. S.; Bromberg, P. A. (1995) Bronchoscopic
determination of ozone uptake in humans. J. Appl. Physiol. 79: 852-860.
Gunnison, A. F.; Hatch, G. E. (1999) O3-induced inflammation in prepregnant, pregnant, and lactating rats correlates
withO3 dose estimated by 18O. Am. J. Physiol. 276: L332-L340.
Hatch, G. E.; Slade, R.; Stead, A. G.; Graham, J. A. (1986) Species comparison of acute inhalation toxicity of ozone
and phosgene. J. Toxicol. Environ. Health 19: 43-53.
Hatch, G. E.; Slade, R.; Harris, L. P.; McDonnell, W. F.; Devlin, R. B.; Koren, H. S.; Costa, D. L.; McKee, J. (1994)
Ozone dose and effect in humans and rats: a comparison using oxygen-18 labeling and bronchoalveolar
lavage. Am. J. Respir. Crit. Care Med. 150: 676-683.
Hu, S.-C.; Ben-Jebria, A.; Ultman, J. S. (1994) Longitudinal distribution of ozone absorption in the lung: effects of
respiratory flow. J. Appl. Physiol. 77: 574-583.
Iwasaki, T.; Takahashi, M.; Saito, H.; Arito, H. (1998) Adaptation of extrapulmonary responses to ozone exposure in
conscious rats. Ind. Health 36: 57-60.
Joad, J. P.; Brie, J. M.; Weir, A. J.; Putney, L.; Hyde, D. M.; Postlewait, E. M.; Plopper, C. G. (2000) Effect of
respiratory pattern on ozone injury to the airways of isolated rat lungs. Toxicol. Appl. Pharmacol.
169: 26-32.
Kabel, J. R.; Ben-Jebria, A.; Ultman, J. S. (1994) Longitudinal distribution of ozone absorption in the lung:
comparison of nasal and oral quiet breathing. J. Appl. Physiol. 77: 2584-2592.
Kari, F.; Hatch, G.; Slade, R.; Crissman, K.; Simeonova, P. P.; Luster, M. (1997) Dietary restriction mitigates
ozone-induced lung inflammation in rats: a role for endogenous antioxidants. Am. J. Respir. Cell Mol. Biol.
17: 740-747.
Lee, C.; Watt, K. C.; Chang, A. M.; Plopper, C. G.; Buckpitt, A. R.; Pinkerton, K. E. (1998) Site-selective
differences in cytochrome P450 isoform activities: comparison of expression in rat and rhesus monkey lung
and induction in rats. Drug Metab. Dispos. 26: 396-400.
Miller, F. J.; Menzel, D. B.; Coffin, D. L. (1978) Similarity between man and laboratory animals in regional
pulmonary deposition of ozone. Environ. Res. 17: 84-101.
4-23
-------
Miller, F. I; Overton, J. H.; Gerrity, T. R.; Graham, R. C. (1988) Interspecies dosimetry of reactive gases. In:
Mohr, U.; Dungworth, D.; McClellan, R.; Kimmerle, G.; Stober, W.; Lewkowski, J., eds. Inhalation
toxicology: the design and interpretation of inhalation studies and their use in risk assessment. New York,
NY: Springer-Verlag; pp. 139-155.
Miller, F. J.; Overton, J. H.; Kimbell, J. S.; Russell, M. L. (1993) Regional respiratory tract absorption of inhaled
reactive gases. In: Gardner, D. E.; Crapo, J. D.; McClellan, R. O., eds. Toxicology of the lung. 2nd ed. New
York, NY: Raven Press; pp. 485-525. (Target organ toxicology series).
Morgan, K. T.; Monticello, T. M.; Patra, A. L.; Fleishman, A. (1989) Preparation of rat nasal airway casts and their
application to studies of nasal airflow. In: Crapo, J. D.; Smolko, E. D.; Miller, F. J.; Graham, J. A.; Hayes,
A. W., eds. Extrapolation of dosimetric relationships for inhaled particles and gases. New York, NY:
Academic Press, Inc.; pp. 45-58.
Mudway, I. S.; Kelly, F. J. (2004) An investigation of inhaled ozone dose and the magnitude of airway inflammation
in healthy adults. Am. J. Respir. Crit. Care Med. 169: 1089-1095.
Nodelman, V.; Ultman, J. S. (1999) Longitudinal distribution of chlorine absorption in human airways: a comparison
to ozone absorption. J. Appl. Physiol. 87: 2073-2080.
Overton, J. H.; Graham, R. C. (1995) Simulation of the uptake of a reactive gas in a rat respiratory tract model with
an asymmetric tracheobronchial region patterned on complete conducting airway cast data. Comput.
Biomed. Res. 28: 171-190.
Overton, J. H.; Graham, R. C.; Menache, M. G.; Mercer, R. R.; Miller, F. J. (1996) Influence of tracheobronchial
region expansion and volume on reactive gas uptake and interspecies dose extrapolations. Inhalation
Toxicol. 8: 723-745.
Plopper, C. G.; St. George, J.; Mariassy, A.; Nishio, S.; Heidsiek, J.; Weir, A.; Tyler, N.; Wilson, D.; Cranz, D.;
Hyde, D. (1989) Species differences in airway cell distribution and morphology. In: Crapo, J. D.; Smolko,
E. D.; Miller, F. J.; Graham, J. A.; Hayes, A. W., eds. Extrapolation of dosimetric relationships for inhaled
particles and gases. New York, NY: Academic Press, Inc.; pp. 19-34.
Plopper, C. G.; Hatch, G. E.; Wong, V.; Duan, X.; Weir, A. J.; Tarkington, B. K.; Devlin, R. B.; Becker, S.;
Buckpitt, A. R. (1998) Relationship of inhaled ozone concentration to acute tracheobronchial epithelial
injury, site-specific ozone dose and glutathione depletion in rhesus monkeys. Am. J. Respir. Cell Mol. Biol.
19: 387-399.
Postlethwait, E. M.; Joad, J. P.; Hyde, D. M.; Schelegle, E. S.; Brie, J. M.; Weir, A. J.; Putney, L. F.; Wong, V. J.;
Velsor, L. W.; Plopper, C. G. (2000) Three-dimensional mapping of ozone-induced acute cytotoxicity in
tracheobronchial airways of isolated perfused rat lung. Am. J. Respir. Cell Mol. Biol. 22: 191-199.
Rigas, M. L.; Ben-Jebria, A.; Ultman, J. S. (1997) Longitudinal distribution of ozone absorption in the lung: effects
of nitrogen dioxide, sulfur dioxide, and ozone exposures. Arch. Environ. Health 52: 173-178.
Rigas, M. L.; Catlin, S. N.; Ben-Jebria, A.; Ultman, J. S. (2000) Ozone uptake in the intact human respiratory tract:
relationship between inhaled dose and actual dose. J. Appl. Physiol. 88: 2015-2022.
Santiago, L. Y.; Harm, M. C.; Ben-Jebria, A.; Ultman, J. S. (2001) Ozone adsorption in the human nose during
unidirectional airflow. J. Appl. Physiol. 91: 725-732.
Sarangapani, R.; Gentry, P. R.; Covington, T. R.; Teeguarden, J. G.; Clewell, H. J., III. (2003) Evaluation of the
potential impact of age- and gender-specific lung morphology and ventilation rate on the dosimetry of
vapors. Inhalation Toxicol. 15: 987-1016.
Schelegle, E. S.; Alfaro, M. F.; Putney, L.; Stovall, M.; Tyler, N.; Hyde, D. M. (2001) Effect of C-fiber-mediated,
ozone-induced rapid shallow breathing on airway epithelial injury in rats. J. Appl. Physiol. 91: 1611-1618.
Sterner-Kock, A.; Kock, M.; Braun, R.; Hyde, D. M. (2000) Ozone-induced epithelial injury in the ferret is similar
to nonhuman primates. Am. J. Respir. Crit. Care Med. 162: 1152-1156.
Stone, K. C.; Mercer, R. R.; Gehr, P.; Stockstill, B.; Crapo, J. D. (1992) Allometric relationships of cell numbers and
size in the mammalian lung. Am. J. Respir. Cell Mol. Biol. 6: 235-243.
U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF.
3v. Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available
online at: www.epa.gov/ncea/ozone.htm.
Ultman, J. S.; Ben-Jebria, A.; Arnold, S. F. (2004) Uptake distribution of ozone in human lungs: intersubject
variability in physiologic response. Boston, MA: Health Effects Institute. Available:
http://www.healtheffects.org/Pubs/Ultman.pdf [29 July, 2005].
4-24
-------
Watkinson, W. P.; Campen, M. J.; Wichers, L. B.; Nolan, J. P.; Costa, D. L. (2003) Cardiac and thermoregulatory
responses to inhaled pollutants in healthy and compromised rodents: modulation via interaction with
environmental factors. Environ. Res. 92: 35-47.
Weibel, E. R. (1980) Design and structure of the human lung. In: Fishman, A. P., ed. Pulmonary diseases and
disorders. New York, NY: McGraw-Hill; p. 18-65.
4-25
-------
5. TOXICOLOGICAL EFFECTS OF OZONE AND
RELATED PHOTOCHEMICAL OXIDANTS IN
LABORATORY ANIMALS AND IN VITRO
TEST SYSTEMS
5.1 INTRODUCTION
A wide range of effects of ozone (O3) has been demonstrated in laboratory animals. The
major research findings are that environmentally relevant levels of O3 cause lung inflammation;
decreases in host defenses against infectious lung disease; acute changes in lung function,
structure, and metabolism; chronic lung disease, some elements of which are irreversible; and
systemic effects on target organs (e.g., brain, heart, liver, immune system) distant from the lung.
The research also has served to expand the understanding of mechanisms of O3 toxicity and the
relationships between concentration and duration of exposure.
The framework for presenting the health effects of O3 in animals begins with a presentation
of respiratory tract effects, followed by systemic effects, and then interactions of O3 with other
common co-occurring pollutants. The information discussed in this chapter is founded on a very
wide body of literature on studies in laboratory animals and on in vitro test systems of animal
cell lines and organ systems that may mimic responses in intact animals. The direct effects of O3
in humans are discussed in the following chapter (Chapter 6).
This chapter is not intended to be a compendium of all that is known about toxicologic
effects of O3; rather, it is an update of the toxicology chapter from the last previous O3 criteria
document (U.S. Environmental Protection Agency, 1996), or 1996 O3 AQCD, and other reviews
of the earlier published literature. The historical O3 literature is very briefly summarized in an
opening paragraph of each section or subsection. That paragraph is intended as a very concise
overview of previous work, and the reader is referred to the 1996 O3 AQCD for more detailed
discussion of the literature prior to the early 1990's. Each section then continues with brief
discussions of the key new studies (or somewhat older studies that were not included in the
1996 O3 AQCD). Longer discussions of new studies are included where warranted. Sections are
ended with comparisons of data from the previous AQCD with new data, and basic conclusions
are drawn. Summaries of new studies and results are provided in tables in Annex AX5.
5-1
-------
Except for nitrogen dioxide (NO2), the subject of another criteria document (U.S.
Environmental Protection Agency, 1993), there is very little relevant information on other
photochemical oxidants in the published literature. What is known about the effects of these
other oxidants is also summarized briefly in this chapter.
5.2 RESPIRATORY TRACT EFFECTS OF OZONE
5.2.1 Biochemical Effects
Biochemically detected effects of O3 are integrally involved in effects on both structure
and function (respiratory and nonrespiratory) of the respiratory tract. Changes in xenobiotic
metabolism, antioxidant metabolism and oxygen consumption, lipids and arachidonic acid
metabolism, and collagen metabolism are all observed with O3 exposure, though the mechanisms
and associations are not fully understood.
5.2.1.1 Cellular Targets of Ozone Interaction
Ozone has the potential to interact with a wide range of different cellular components that
include polyunsaturated fatty acids (PUFAs); some protein amino acid residues; and some
low-molecular-weight compounds that include glutathione (GSH), urate, vitamins C and E, and
free amino acids. Early work demonstrated that O3, being a highly reactive compound, does not
penetrate much beyond the epithelial lining fluid (ELF); and reaction/diffusion analyses suggest
that O3, at environmentally-relevant concentrations, diffuses no more than 0.1 to 0.2 jim into the
ELF. However, Miller (1995) points out that throughout the respiratory tract, the ELF varies in
thickness and that the distal conducting airways may have only a patchy lining layer which
allows O3 to react directly with cell membranes. Various models utilizing differing reaction
kinetics and rate constants do not agree as to whether or not inhaled O3 can penetrate the ELF
and reach the epithelial cells. Ozone-induced cell damage results, in part, from its reactions with
PUFAs to form stable but less reactive ozonide, aldehyde, and hydroperoxide reaction products.
These reaction products (Criegee ozonides and hydroxyhydroperoxides) may act as signal
transduction molecules involved in signaling of cellular responses such as inflammation and thus
mediate O3 toxicity. These reactions are summarized in Figure 5-1. Studies published since the
1996 AQCD are listed in Table AX5-1.
5-2
-------
RHC = CH
PUFA
ozone
O O
I I
RHC —CH—
trioxolane
RHC = O — O + RHC = O
carbonyl oxide aldehyde
either in
the —
absence
ofH2O
XO Ox or jn the
RHC CH— presence
^Q/ of H2O
Criegee ozonide
/OH
->• RHC >
\DOH
hydroxyhydroperoxy cpd.
= O +
aldehyde
H2O2
hydrogen
peroxide
Figure 5-1. Schematic overview of ozone interaction with epithelial lining fluid and lung
cells. It should be noted that not all secondary reaction products are shown.
Frampton et al. (1999) demonstrated the ozonation of PUFA to form nonanal and hexanal
in rat bronchioalveolar lavage (BAL) after exposures to 0.22 ppm O3 for 4 h with exercise.
Increases in nonanal were not accompanied by significant changes in lung function, in epithelial
permeability, or in airway inflammation. Hexanal levels did not increase significantly and levels
of both aldehydes returned to baseline by 18 h postexposure (PE). Pryor et al. (1996) exposed
rats to 0.5 to 10 ppm O3 both with and without 5% CO2 to measure the amount of aldehyde
generated in BAL and also the rate of disappearance of aldehydes from the ELF following the O3
exposure. Ozone exposure with CO2 increased the tidal volume and the yield of aldehydes, with
a maximal aldehyde yield at 2.5 ppm for 1 h. Absolute yields were impossible to ascertain in
this system, because absorption of O3 is unknown and aldehyde recovery is not complete due to
loss of aldehyde by volatization and by diffusion into underlying tissue. The data showed that at
0.5 ppm O3 with 5% CO2, levels of hexanal and nonanal increased at 30 min, decreased slightly
from that levels at 60 min, was maximal at 90 min and then dropped to 60 min levels at 120 min.
Heptanal levels did not change appreciably during this time course. Levels of these aldehydes
were dependent on a dynamic relationship between their production and disappearance from the
ELF. The authors stated that O3 is the limiting reagent in this process because the amount of
PUFA far exceeds the amount of O3 on a molar basis. Because of the limitations of measuring
aldehydes in this study protocol, the results are not useful for quantitative dosimetry; however,
5O
-------
the authors suggest that the study does serve to demonstrate the use of aldehydes as biomarkers
of O3 exposure, as nonanal is produced in an O3-specific pathway.
Postlethwait et al. (1998) utilized three biologically relevant models (isolated epithelial
lining fluid, intact lung, and liposome suspensions) to determine the O3-induced production of
heptanal, nonanal and hexanal in an attempt to estimate formation of lipid-derived bioactive
compounds. Exposures used were 0.25 to 1.0 ppm for 30 to 60 min. The results suggest that
PUFAs directly react with O3 and the amount of bioactive lipids produced is inversely related to
ascorbic acid (AA) availability. The authors caution that there are limitations to the use of
measurements of these reaction products in determining O3 dose-response relationships due to
the heterogenous nature of O3 reactions in the epithelial lining fluid. Connor et al. (2004) have
recently examined the reactive absorption of O3 (0.3 to 1.1 ppm for 1 to 2 h) within ELF using
interfacial films composed of dipalmitoylglycero-3-phosphocholine (DPPC) and rat lung lavage
fluid. The films reduced O3-reactive absorption by antioxidants. Further experiments using a
human lung fibroblast cell line exposed to O3 demonstrated that AA produced cell injury, that
high levels of O3 and AA were needed to induce cell injury, and that DPPC films reduced the
amount of cell injury. Based on these data, the authors suggest that O3 reactions with ELF
substrates cause cell injury, that films of active, saturated phospholipids reduce the local dose
of O3-derived reaction products, and that these interfacial phospholipids modulate the
distribution of inhaled O3 and the extent of site-specific cell injury.
Recent studies have examined the formation of ozonation products such as
4-hydroxynonenal (FINE), a toxic aldehyde that reacts with cysteine, histamine, and lysine
amino acid residues and creates protein adducts. Hamilton et al. (1998) demonstrated (see
Chapter 6) using human alveolar macrophages (AMs) exposed to 0.4 ppm O3 for 1 h that
exposure caused apoptosis, an increase in a 32-kDa protein adduct, and an increase in ferritin
and a 72-kDa heat shock protein. By exposing AM to FINE in vitro, all of these effects are
replicated, which the authors interpret to mean that creation of protein adducts and apoptotic cell
death are cellular toxic effects of acute O3 exposure that are mediated, at least in part, by FINE.
These recent reports, combined with observations reported in the previous O3 AQCD (U.S.
Environmental Protection Agency, 1996), suggest that interactions of O3 with cellular
components and ELF generate toxic ozonation products and mediate toxic effects through these
products.
5-4
-------
5.2.1.2 Monooxygenases
Both short- and long-term exposures to O3 have been shown to enhance lung xenobiotic
metabolism, possibly as a result of changes in the number and function of bronchiolar epithelial
Clara cells and alveolar epithelial Type II cells. Studies of the effects of O3 on lung
monooxygenases are listed in Table AX5-2. Early studies showed that exposure to O3 increased
CYP 2B1 (the major CYP isoform in rat lung) content and activity in rat lung. Ozone exposures
also caused hypertrophy and hyperplasia of CYP 2Bl-immunoreactive Clara cells. Comparisons
of rat and rhesus monkey CYP isoforms demonstrated species-specific and region-specific (e.g.,
trachea, parenchyma) differences in the activities of P450 isoforms (Lee et al., 1998).
Watt et al. (1998) found that, with both acute (8 h, 1 ppm) and chronic (90 days, 1 ppm)
exposures in rat, 1 ppm O3 increased CYP 2E1 in a region-specific manner. Paige et al. (2000a)
showed that a long-term exposure (0.8 ppm, 8 h/day for 90 days) increased the activity of CYP
2B in distal lung but not in trachea or intrapulmonary airways. Studies have focused on P450
gene expression to examine possible genetic mechanisms that may explain differential O3-
sensitivity (Mango et al., 1998). Mice (129 strain) deficient in Clara cell secretory protein
(CCSP-/-), which are oxidant-sensitive, were exposed to 1 ppm O3 for 2 hours. The CCSP null
mice demonstrated increases in interleukin-6 (IL-6) and metallothionein (MT) mRNA that
preceded decreases in Clara cell CYP 2F2 mRNA (normally expressed at high levels in mouse
lung) levels. In 129 strain wild-type (WT) mice, RNA levels changed similarly, but to a lesser
degree. These data suggest a protective role against oxidant damage for CCSP, and further, that
genetic susceptibility to oxidant stress may be mediated, in part, by the gene coding for CCSP.
5.2.1.3 Antioxidants, Antioxidant Metabolism, and Mitochondrial Oxygen Consumption
Ozone also undergoes reactions with AA, reduced glutathione (GSH), and uric acid (UA),
all antioxidants present in ELF (see Figure 5-2, A). In vivo experiments have shown that
reactions with O3 occur preferentially with antioxidants compared to proteins and lipids also
present in ELF. This is a protective interaction, but even with environmentally relevant
exposures to O3, the reactivity of O3 is not quantitatively quenched. Antioxidants offer some
protection from O3 exposure but often are not maintained at concentrations sufficient to fully
protect the lung. Thus, O3-induced cell injury occurs in both the lower and upper respiratory
tract. Early work has shown that acute (1 week) exposures to <1 ppm O3 increase antioxidant
5-5
-------
Ozone
i
t
Mucus
||GSH||[~AA~| | UA || ^os
T.J.C.
A.AI.
B.E.C.
Endo.
Figure 5-2. The major cellular targets and proposed mechanisms of ozone toxicity in the
lung. Abbreviations: GSH - reduced glutathione; AA - ascorbate; UA - urate;
a-Toc - alpha-tocopherol; GSH-Px - glutathione peroxidase; EC-SOD -
extracelluar superoxide dismutase; T.I.C. - Type I cells; T.II.C. - Type II cells;
Neut - neutrophils; Lymp - lymphocytes; Epi. - epithelium; Inst. - interstitum;
Endo - endothelium; M.C. - monocyte; C.E.C. - ciliated epithelial cells; A.M. -
alveolar macrophage; B.E.C. - bronchial epithelial cells; ROS - reactive oxygen
species.
Source: Mudway and Kelly (2000), with permission.
metabolism, including levels of cytosolic enzymes glucose-6-phosphate dehydrogenase (G6PD),
6-phosphogluconate dehydrogenase (6PGD), glutathione reductase (GR), and glutathione
peroxidase (GSHPx). Reexposure after a recovery period causes increases equivalent to first-
time exposures; thus, previous exposure does not appear to be protective.
Increases in enzyme activity appear to increase as a function of age, suggesting that O3
exposure can cause greater lung injury in the older animal. This has been attributed to
differences in dose reaching lung target sites, differing base levels of antioxidants and
antioxidant enzymes, and differences in cellular sensitivity. Species differences exist in
antioxidant metabolism, with guinea pigs being very sensitive to O3 due to their diminished
increases in antioxidants and antioxidant enzymes. Chronic exposures of rats to urban patterns
5-6
-------
of O3 (daily peaks of 0.25 ppm) caused increases in GSHPx and GR, but not superoxide
dismutase (SOD). The enzyme changes could be accounted for by differences in the steady-state
cell population or in cellular antioxidant capacity. More recent studies examining antioxidants
and O3 exposure are listed in Table AX5-3.
Ozone induced both site- and cell-specific changes in copper-zinc (Cu-Zn) and manganese
(Mn) SOD in rats exposed to 1.0 ppm O3 for up to 3 months (Weller et al., 1997). CuZnSOD
labeling was decreased in epithelial cells in airways and parenchyma. Mn-SOD labeling was
increased in both AM and epithelial type II cells of the centriacinar region (CAR), which the
authors suggest may allow these cells to tolerate further O3 exposure. This work is in agreement
with earlier work suggesting a role of SOD in protection of cells against oxidative stress.
Freed et al. (1999) evaluated the role of antioxidants in O3-induced oxidant stress in dogs
(exposed to 0.2 ppm for 6 h) by inhibiting the antioxidant transport using probenecid (an anion-
transport inhibitor). Blocking antioxidant transport caused heterogeneously distributed increases
in peripheral airway resistance (Raw) and reactivity, supporting the hypothesis that in the lung
periphery, endogenous antioxidants moderate the effects of O3 and that this exposure is a
subthreshold stimulus for producing effects on peripheral Raw and reactivity in dogs. The
authors further found that treatment with probenecid also inhibited O3-induced neutrophilic
inflammation, providing evidence for a dissociation between airway function and inflammation.
This suggests that O3-induced inflammation and airway hyperreactivity (AHR) are independent
phenomena operating through multiple mechanistic pathways.
Mudway and Kelly (1998) modeled the interactions of O3 with ELF antioxidants using a
continually mixed, interfacial exposure set up with O3 concentrations ranging from 0.1 to
1.5 ppm for durations ranging from 30 to 720 min. Uric acid was ranked the most O3-reactive,
AA the second most reactive, and GSH the least reactive. Thus, they concluded that GSH is not
an important substrate for O3, while UA appeared to be the most important reactive substrate,
which confers protection from O3 by removing it from inhaled air and limiting the amount that
reaches the distal lung. By providing a substrate for O3 reactions in the ELF, UA effectively
reduces the diffusion resistance of O3 (see Bush et al., 2001) in the tracheobronchial airways and
thus may serve to limit the amount of O3 reaching the distal lung. The authors acknowledge
limitations in extrapolating these data to in vivo O3 exposures due to the absence of surfactant
lipids and airway mucus in this in vitro model.
5-7
-------
5.2.1.4 Lipid Metabolism and Content of the Lung
One of the major postulated molecular mechanisms of action of O3 is peroxidation of
mono- and polyunsaturated fatty acids and unsaturated neutral lipids in the lung. Because all of
these lipids appear both in cell membranes and as secretions in the ELF, it is difficult to ascertain
which lipid pool contributes to the formation of lipid ozonation products. As mentioned, O3 can
penetrate only a short distance into the ELF; and, therefore, it reacts with epithelial cell
membranes only in regions of distal lung where ELF is very thin or absent. The inflammatory
cascade (shown in Figure 5-3) initiated by O3 generates a mix of secondary reactants (e.g.,
aldehydes), which then are likely to oxidize lipids and proteins in cell membranes.
In both acute and short-term studies, a variety of lung lipid changes occur, including an
increase in arachidonic acid. Metabolism of arachidonic acid produces a variety of biologically
active mediators that can, in turn, affect host defenses, lung function, the immune system, and
other functions. The protein A component of surfactant is also a primary target of O3 interaction.
During the first few days of O3 exposure, the changes in lung lipid biosynthesis can be accounted
for by the alveolar epithelial proliferative repair. With longer exposures (e.g., 0.12 ppm for
90 days) an increase in PUFAs and a decrease in cholesterol-esters are seen, indicative of
long-term alterations of surfactant lipid composition.
Several new studies listed in Table AX5-4 examined the effects of O3 exposure on
phospholipids in lung tissue. Ozonation of PUFAs has been shown to generate other aldehydes
such as nonanal and hexanal in rats (Pryor et al., 1996; Frampton et al., 1999). These aldehydes
are short-lived and found to not affect lung function (Frampton et al., 1999). These observations
suggest that levels of these aldehydes are dependent on a dynamic relationship between their
production and their disappearance from the ELF.
Pryor et al. (1995) proposed a cascade mechanism whereby ozonation products cause
activation of specific lipases, which then trigger the activation of second messenger pathways
(e.g., phospholipase A2 or phospholipase C). This group (Kafoury et al., 1999) showed that
exposure of cultured human bronchial epithelial cells to the lipid ozonation product 1-palmitoyl-
2-(9-oxononanoyl)-sn-glycero-3-phosphocholine elicited release of platelet-activating factor
(PAF) and PGE2, but not IL-6. The lipid ozonation product 1-hydroxy-l-hydroperoxynonane
caused release of PAF and IL-6 in these cells, but not PGE2. These results suggest to the authors
-------
Airway
Epithelial
Lining
Fluid
Superoxide
+ H2O
Reactive Oxygen
rmediate
Lipase Activation
I
Phospholipases
I
ArachidonicAcid
\
Leukotrienes Prostanoids
Inflammatory Cells
e.g.,AMS, PMNs,
ood monocytes,
ast cells
Proteolytic
Enzymes
111
u CL *-
JE IJJ 0)
Airway
Hyperreactivity
Airway
Smooth
Muscle
Adapted from: Pryoretal. (1995); Krishna et al. (1998); Bhallaetal. (1999)
Figure 5-3. Mechanisms of ozone toxicity. The inflammatory cascade starts by interactions of O3
with lipids in the epithelial lining fluid and in cell membranes. Lipid ozonation
reaction products include aldehydes, hydroxyhydroperoxides, and H2O2. These
initiate lipase activation in epithelial cells resulting in the downstream production of
phopholipases, arachidonic acid, leukotrienes, and prostanoids. Leukotrienes and
prostanoids alter lung function. Phosphoiipases initiate the production of NO,
inflammatory cytokines, and PAF. NO, IL-6, and IL-8 then participate in the
recruitment and activation of inflammatory cells (AMs, PMNs, blood monocytes, and
mast cells), which, in turn, can start a second cycle of inflammatory responses. PAF
increases epithelial permeability and may have thrombolytic effects. Release of
substance P by neurons causes altered lung function and increases vascular
permeability. Production of NOS-1 by these neurons induces airway hyperreactivity.
5-9
-------
that O3-induced production of lipid ozonation products causes release of proinflammatory
mediators that then generate an early inflammatory response.
Very new work (Ballinger et al., 2005) has shown that O3-induced membrane oxidation is
augmented by antioxidants present in ELF. They utilized a red cell membrane model exposed to
0.8 ppm O3 for 30 min. The monolayer of cells was intermittently covered by an aqueous film
consisting of rat bronchioalveolar lavage fluid (BALF) or BALF plus added antioxidants.
Ascorbic acid and GSH induced dose-dependent oxidative damage to the cell membrane proteins
and lipids via secondary oxidant formation. The authors concluded that early in O3 exposure,
ELF antioxidants are high enough to drive reactive absorption of O3 into the ELF and to
concurrently quench secondary reaction products, thus limiting cell injury. With continued
exposure, levels of antioxidants decrease such that unreacted O3 and cytotoxic products can
diffuse to the cell membranes, causing injury. Limitations of this in vitro study include possible
differences in chemical species and mechanisms compared to in vivo systems.
Uhlson et al. (2002) reacted O3 with calf lung surfactant which resulted in the production
of l-palmitoyl-2-(9'-oxo-nonanoyl)-glycerophosphocholine (16:Oa/9-al-GPCho). The biological
activity of this oxidized phospholipid included: (1) decreased macrophage viability,
(2) induction of apoptosis in pulmonary epithelial-like A549 cells, (3) and release of IL-8 from
A549 cells. Exposures to 0.125 ppm O3 for 2 to 4 h in this system were capable of generating
biologically active phospholipids that were capable of mediating toxic effects of O3.
In addition to PUFA, cholesterol, the most abundant neutral lipid present in ELF, is also a
target of O3. Pulfer and Murphy (2004) demonstrated the ozonolysis of cholesterol in an in vitro
system using BALF isolated from rats that had been exposed to 2.0 ppm O3 for 4 h. Production
of S-hydroperoxy-.g-homo-e-oxa-cholestan-SJa-diol, 5p,6p-epoxycholesterol, and 3p-hydroxy-
5-oxo-5,6-seco-cholestan-6-al was shown. Additionally, both 5p,6p-epoxycholesterol and its
most abundant metabolite, cholestan-6-oxo-3p,5a-diol, were demonstrated to be cytotoxic to
16-HBE cells and to inhibit cholesterol synthesis. Studies (Pulfer et al., 2005) in C57BL/6J mice
exposed to 0.5, 1.0, 2.0, or 3.0 ppm O3 for 3 h demonstrated that these oxysterols were produced
in vivo also. The authors suggested that this may be an additional mechanism of O3 toxicity.
Though these oxysterol reaction products have not been fully characterized, they may be
involved in O3-induced inflammation by disrupting cellular membranes or altering signaling
5-10
-------
between cells. Similar oxysterols have been implicated in the inflammatory cascade associated
with atherosclerosis.
Thus, new work has attempted to elucidate the mechanisms by which reactions of O3 with
lipids create phospholipids that then mediate downstream toxic effects. It is uncertain whether
these described changes in lipid content and/or metabolism lead to significant changes in surface
tension or compliance properties of the lung.
5.2.1.5 Ozone Interactions with Proteins and Effects on Protein Synthesis
Epithelial lining fluid contains proteins arising from airway secretions and from blood.
Ozone can react with four amino acid residues (cysteine, histidine, methionine, and tryptophan)
and can cause oxidation of functional groups on proteins, including aldehydes, alcohol, amines
and sulfhydryls. A number of enzymes have been shown to be inhibited by O3, including
cholinesterase, al-antiproteinase, and prostaglandin synthetase. Additionally, O3 decreases the
inhibitory activity of al-proteinase inhibitor, which is implicated in development of emphysema.
Surfactant protein A (SP-A) is a target for O3 toxicity by modulation of SP-A self association,
vesicle aggregation, phospholipid secretion, and stimulation of AM superoxide anion generation
(see Section 5.2.2.3). Further, O3 is thought to interfere in SP-A's homeostatic role in surfactant
release from alveolar Type II cell lamellar bodies and its subsequent uptake by Type II cells
and AMs.
Lung collagen, collagen synthesis, and prolyl hydroxylase activity associated with
fibrogenesis have been shown to increase in rodents with O3 exposure of >0.45 ppm. Some
studies have shown that this increase persists after exposure stops and that there is an influence
of exposure pattern on the response. The increased collagen has been correlated with structural
changes in the lung. Rats exposed to an urban pattern of O3 with daily peaks of 0.25 ppm for
38 weeks displayed extracellular matrix thickening. Increased levels of collagen in CAR were
demonstrated in female rats exposed to 0.5 to 1.0 ppm O3 for 6 h/day for 20 months and in
monkeys exposed to 0.61 ppm for 1 year. Both increased age and health status (e.g.,
emphysemic) were implicated in the increased collagen formation in response to O3 exposure.
A time-course study (van Bree et al., 2001; Table AX5-5) evaluating the lung injury and
changes in collagen content in rats exposed acutely or subchronically to 0.4 ppm O3
demonstrated CAR thickening of septa, which progressed from 7 through 56 days of exposure.
5-11
-------
Though collagen content decreased with PE recovery, the structural fibrotic changes in ductular
septa and respiratory bronchioles persisted, suggesting that subchronic O3 exposures in rats
creates a progression of structural lung injury that can evolve to a more chronic form, which
includes fibrosis. The biological relevance and adverse health effects of altered protein synthesis
and collagen accumulation are uncertain.
5.2.1.6 Differential Gene Expression
Gohil et al. (2003) examined differential gene expression in C57BL/6 mice exposed to
1 ppm O3 for three consecutive nights for 8 h (see Table AX5-6). Ozone exposure induced
changes in expression of 260 genes (80% repressed and 20% induced). Differentially expressed
genes included those involved in progression of the cell cycle, such as ^-adenosyl methionine
decarboxylase 3 (SAMD), ribonucleotide reductase (RR), and clusterin. Increased transcription
of these genes suggests O3-induced activation of the cell cycle, with subsequent cellular
proliferation. This is in accord with the finding of increased epithelial proliferation with acute
O3 exposure, as discussed in studies in Sections 5.2.4.1 and 5.2.5.1. Several nuclear factor-KB
(NF-KB)-induced genes were upregulated, including serum amyloid protein, topoisomerase Hoc,
monocyte chemoattractant protein and platelet-derived growth factor, an inhibitor of apoptosis.
Upregulation of these genes suggests to the authors that they may account for O3-induced
proliferation of nonciliated cells and Clara cells. Downregulation of transcripts for isoforms of
myosins and actins were also observed, which may explain, in part, a mechanism of O3-induced
vascular permeability. Several members of the CYP family were downregulated, including 2a4,
2el, and 2f2, as were aryl-hydrocarbon receptor and several glutathione transferases.
Metallothionein 1 and 2 and lactotransferrin were upregulated, indicative of their function as
antioxidants and anti-inflammatory agents. Ozone-induced suppression of immune function is
suggested by downregulation of transcripts encoding major histocompatibility complex genes,
lymphocyte-specific proteins, and immunoglobulins. Section 5.2.2.3 discusses the effects of O3
exposure on the immune system.
Quinlan et al. (1994) have reviewed the regulation of antioxidant enzymes in lung after
oxidant injury. A comparison of alterations in gene expression in rat following O3 or hyperoxia
exposure, both of which induce reactive oxygen species (ROS) and injury to vascular endothelial
cells and cells of the alveoli, show that both ~1 ppm O3 and 85 to 95% O2 increase expression of
5-12
-------
CuZnSOD, GSHPx, and catalase. Johnston et al. (1998) also found changes in gene expression
indicative of inflammation and epithelial injury that occur with hyperoxia in mice (95% O2) that
are analgous to similar injury that occurs following O3 exposure.
5.2.1.7 Summary and Conclusions—Biochemical Effects
Ozone has been shown to interact with a wide range of different cellular components
including PUFAs, amino acid residues, and some low-molecular-weight compounds (GSH,
urate, vitamins C and E). As O3 does not penetrate much beyond the ELF, damage likely results
from its PUFA ozonation products (mostly hydroxyhydroperoxides) involvement in signaling of
cellular responses such as inflammation. New work has shown that ozonation of PUFA also
forms the aldehydes nonanal, heptanal, and hexanal, the production of which are dependent on
AA availability. Saturated phospholipids are thought to reduce the local dose and limit site-
specific cell injury from O3 exposure. Another ozonation product, FINE, creates protein adducts
that have been linked to apoptosis and heat shock proteins in vitro.
Both short- and long-term exposures to O3 have been shown to enhance lung xenobiotic
metabolism, possibly as a result of changes in the number and function of bronchiolar epithelial
Clara cells and alveolar epithelial Type II cells. This modulation is both species- and region-
specific and includes the isoforms CYP 2B1, CYP 2E1. CCSP is also involved in inflammatory
responses to O3 exposure. Mouse strains with differing sensitivities to O3 show that responses in
protein, lactate dehydrogenase (LDH), and inflammatory cell influx are due to CCSP levels and
changes in lung epithelial permeability.
Reactions of O3 with AA, GSH, and UA (all antioxidants present in ELF) are a protective
mechanism. But even with environmentally relevant exposures, the reactivity of O3 is not
quantitatively quenched, and cell injury occurs in both the lower and upper respiratory tract.
Early work has shown that short-term exposures to <1 ppm O3 increase antioxidant metabolism.
Reexposure after a recovery period causes increases equivalent to first-time exposures,
suggesting that previous exposure is not protective. Elevations in enzyme activity appear to
increase as a function of age, suggesting that O3 exposure can cause greater lung injury in the
older animal. Long-term urban patterns of exposure to O3 (daily peaks of 0.25 ppm) caused
increases in GSHPx and GR, but not SOD. Recent work has suggested that endogenous
antioxidants moderate the effects of O3 and that this exposure is a subthreshold stimulus for
5-13
-------
producing effects on peripheral Raw and reactivity, thus indicating a dissociation between
airway function and inflammation.
In both acute and short-term studies, a variety of lung lipid changes occur with O3
exposure, including an increase in AA. With longer exposures (e.g., 0.12 ppm for 90 days),
an increase in PUFAs and a decrease in cholesterol-esters are seen, indicative of long-term
alterations of surfactant lipid composition. Whether these changes in lipid content and/or
metabolism lead to significant changes in surface tension or compliance properties of the lung
remains unknown. New studies evaluating O3-induced alterations in lipid metabolism have not
been completed.
Collagen, a structural protein involved in fibrosis, increases with O3 exposure, and some
studies have shown that this increase persists after exposure stops. Urban patterns of exposure
(daily peaks of 0.25 ppm O3 for 38 weeks) created extracellular matrix thickening. Increases in
centriacinar collagen were demonstrated in female rats exposed to 0.5 to 1.0 ppm O3 for 6 h/day
for 20 months and in monkeys exposed to 0.61 ppm for 1 year. New work examining the time
course of lung injury and changes in collagen content in rats exposed acutely or subchronically
to 0.4 ppm O3 showed centriacinar thickening of septa. Collagen content decreased with PE
recovery but not the structural fibrotic changes in ductular septa and respiratory bronchioles,
which suggests that subchronic O3 exposures in rats creates a progression of structural lung
injury that can evolve to a more chronic form, including fibrosis.
5.2.2 Lung Host Defenses
Defense mechanisms, including the mucociliary clearance system, AMs, and humoral- and
cell-mediated immune system, exist in the lung to protect it from infectious and neoplastic
disease and inhaled particles. Summaries of key new animal studies examining the effects of O3
on lung host defenses are presented in Table AX5-7 of Annex AX5. Acute human exposures
to O3 result in similar effects on AMs (see Chapter 6).
5.2.2.1 Clearance
Early studies of O3 effects on the mucociliary escalator showed morphological damage to
ciliated epithelial cells of the tracheobronchial tree at O3 doses of <1 ppm. Functionally, O3
slowed particle clearance in rats at doses of 0.8 ppm for 4 h and in rabbits at 0.6 ppm for 2 h
5-14
-------
exposures. Acute exposures at 0.5 ppm O3 in sheep caused increased basal secretion of
glycoproteins, while longer exposures reduced tracheal glycoprotein secretions, both of which
can alter the effectiveness of the mucociliary escalator. Early postnatal exposures of sheep to
1 ppm O3 caused retardation of normal morphologic development of the tracheal epithelium,
decreased epithelial mucosa density, decreased tracheal mucous velocity, and delayed
development of carbohydrate composition. Conversely, alveolar clearance in rabbits after
acute O3 exposure (0.1 ppm, 2 h/day, for 1 to 4 days) is increased. Longer exposures showed no
effect, and increased O3 (1.2 ppm) slowed clearance. This pattern of altered clearance also
occurs in rats. A study using rat tracheal explants exposed to O3 for 10 min (Churg et al., 1996)
showed that uptake of titanium dioxide (TiO2) and asbestos was enhanced at 0.01 and 0.1 ppm
O3, respectively. The authors attribute the increased uptake as a direct effect of O3, suggesting
mediation by H2O2 or hydroxyl radical. Studies of the clearance of the radiolabled chelate 99mTc
diethylenetriamine pentaacetic acid (99mTc-DTPA) have shown that clearance is significantly
increased following a 3 h exposure to 0.8 ppm O3 in Sprague-Dawley (SD) rats (Pearson and
Bhalla,1997). Examination of regional clearance of 99mTc-DTPA in dogs following a 6 h isolated
sublobar exposure to 0.4 ppm O3 or air showed that O3 decreased the clearance half-time by 50%
at 1 day following exposure (Foster and Freed, 1999). Clearance was still elevated at 7 days PE
but had recovered by 14 days. Thus, a single local exposure to O3 increases transepithelial
clearance but without any influence on contralateral segments, i.e., only for epithelia directly
exposed to O3.
Alveolar clearance is slower than tracheobronchial clearance and involves particle
movement through interstitial pathways to the lymphatic system or movement of particle-laden
AMs to the bottom of the mucociliary escalator. Exposures of rabbits to 0.1 ppm accelerated
clearance, whereas 1.2 ppm O3 slowed clearance. A chronic exposure has been shown to slow
clearance. New evaluations of O3 effects on alveolar clearance have not been performed.
5.2.2.2 Alveolar Macrophages
A primary function of AMs is to clear the lung of infectious and noninfectious particles by
phagocytosis, detoxification, and removal. Further, AMs secrete cellular mediators that recruit
and activate inflammatory cells in the lungs (see Figure 5-3). Ozone has been shown to inhibit
phagocytosis at 0.1 ppm for 2 h in rabbits. This inhibition returns to control levels if exposures
5-15
-------
are repeated for several days. The production of superoxide anion radicals and the activity of
AM lysosomal enzymes (both involved in bactericidal activity) are inhibited by 3 h exposures to
0.4 and 0.25 ppm O3 in rodents and rabbits, respectively. Production of interferon-y (IFN-y)
was decreased in rabbit AM by 1 ppm O3 for 3 h.
New studies have shown that O3 affects AM chemotaxis, cell adhesion, and surface
expression of cell adhesion molecules (Bhalla, 1996). AM from SD rats exposed to 0.8 ppm O3
for 3 h showed greater mobility and greater adhesion than air exposed controls. This increased
mobility and adhesion were attenuated by CD16b and intercellular adhesion molecule-1
(ICAM-1) antibodies, suggesting that these adhesion molecules modulate O3-induced
inflammation. Antibodies to tumor necrosis factor-a (TNF-a) and IL-la also mitigated AM
adherence, suggesting further that the inflammatory response to O3 is mediated by these
cytokines (Pearson and Bhalla, 1997). Cohen et al. (1996) showed that 1 ppm O3 for 4 h reduces
binding of INF-y to AM in WEHI-3 cells and, additionally, reduces phagocytic activity,
production of reactive oxygen intermediates, and elevation of intracellular Ca2+.
Cohen et al. (2001, 2002) exposed male Fisher 344 (F344) rats to either 0.1 or 0.3 ppm O3
for 4 h/day, 5 days/week or either 1 or 3 weeks. In this study, superoxide anion production was
increased at 1 week. Hydrogen peroxide (H2O2) production was reduced at both exposure
concentrations and durations and was further reduced with INF-y stimulation, suggesting that
one effect of O3 is compromised killing of bacteria by AM due to the reduction in H2O2
production.
Ozone treatment (2 ppm O3, 3 h in female SD rats) caused a time-dependent increase in
nitric oxide (NO) levels in both AM and type II epithelial cells that was correlated with
increased expression of inducible nitric oxide synthase (iNOS) mRNA and protein (Laskin et al.,
1998). Inhibition of NF-KB, caused a dose-dependent inhibition of NO and iNOS production.
Additionally, O3 caused a time-dependent increase in NF-KB binding activity in the nucleus of
both cell types. The authors hypothesize that O3 exposure causes the cytokines TNF-a and
IL-lpa to bind to surface receptors and initiate intracellular signaling pathways in AM, leading
to activation of NF-KB, its entry into the nucleus, and its binding to the regulatory sequences of
genes such as iNOS to allow their transcription. Additional studies (Laskin et al., 2002), using
AM isolated from C57B16xl29 mice with a targeted disruption of the gene for iNOS, showed no
toxicity to 0.8 ppm O3 for 3 h, as measured by BALF protein levels and nitrotyrosine staining of
5-16
-------
the lung. Additionally, mice overexpressing human CuZnSOD and mice with a targeted
disruption of p50 NF-KB were resistant to O3 toxicity. Wild-type mice exposed to O3 showed an
increase in expression of STAT-1, a protein that binds to the regulatory region of iNOS. Taken
together, these results suggest to the authors that a number of proteins, including NF-KB,
phosphoinoside 3-kinase, and STAT-1, that bind to and regulate expression of iNOS are
modulated by O3 exposure. The same iNOS knockout mice strain exposed to 0.8 ppm O3for 3 h
(Fakhrzadeh et al., 2002) showed no increase in AM superoxide anion and prostaglandin. These
data provide further evidence that NO and its reactive oxidative product, peroxynitrite, are
important in O3-induced lung injury. Further discussions of the role of nitric oxide
synthase/reactive nitrogen and cytokines/chemokines in O3-induced inflammation are provided
in Section 5.2.3.
5.2.2.3 Immune System
Other than by natural protection (e.g., opsonizing antibody, nonspecific phagocytosis by
AM), the immune system defends the lung by mounting three major waves of response: natural
killer (NK) cells (nonspecific lymphocytes that kill viruses, bacteria, and tumor cells), followed
by cytotoxic T lymphocytes (TCTL, lymphocytes that lyse specifically recognized microbial and
tumor-cell targets), followed by antigen-specific antibodies. These T-cell types are involved
with other immunologically active cells (e.g., B cells and AM), which in a complex manner,
interact in immunological defense. To date, only a few of these mechanisms have been
investigated in the context of their role in O3 susceptibility. The effects of O3 on the immune
system are complex and depend on the exposure parameters and observation periods. T-cell-
dependent functions appear to be more affected than B-cell-dependent functions. Generally,
there is an early immunosuppressive effect that can, with continued exposure, either return to
normal or actually enhance immunity. Changes in immune cell population occur with O3
exposure, including T:B cell ratios in the mediastinal lymph node. Natural killer cell activity
increases with 1 week exposures of 0.2 to 0.4 ppm O3 but decreases with exposures to 0.82 ppm.
Ozone exposure has also been shown to be responsible for enhancement of allergic sensitization
at levels of 0.5 to 0.8 ppm for 3 days. Studies of the effects of O3 on the immune system are
summarized in Table AX5-7.
5-17
-------
Garssen et al. (1997) have studied the effects of O3 on non-IgE-mediated pulmonary
hyper-immune reactions induced by picryl chloride (PCI). BALB/c mice sensitized with PCI,
both actively and passively (by adoptive transfer of lymphoid cells from pre-sensitized mice),
were then challenged with picryl sulfonic acid (PSA). The mice were exposed for 12 h to 0.2,
0.4, or 0.8 ppm O3 during one night, at 4 days or 7 days after skin sensitization done either just
before or just after PSA challenge (i.e., during the induction or effector phase). Nonsensitized
mice showed no changes in tracheal reactivity to carbacol with O3 exposure. Sensitized mice
were hyperreactive to carbachol 48 h after PSA challenge, whereas sensitized mice exposed to
all concentrations of O3 showed no significant tracheal hyperreactivity to carbachol. The
sensitized mice also showed a suppressed inflammatory reaction (polymorphonuclear leukocyte,
PMN) with the 0.8 ppm O3 exposure. Ozone exposure following PSA challenge also caused a
suppression of tracheal hyperresponsiveness. In a separate experiment wherein mice were
exposed to O3 before sensitization and then lymphoid cells from these mice were injected into
nonexposed mice, the recipients also manifested an inhibition of the induction of hyperreactivity.
These results are opposite to the effect on type I (IgE-mediated) allergic reactions, which the
authors suggest is due to activation of T-derived lymphocyte helper 2 (Th2) cell-dependent
reactions that are possibly potentiated by O3 or to a direct effect by O3 on Thl cells or other cells
crucial for the tracheal hyperreactivity and inflammation seen in this mouse model.
Kleeberger et al. (2000, 200la) have demonstrated a potential interaction between the
innate and acquired immune system with O3 exposure. Using O3-susceptible (C57BL/6J)
and O3-resistant (C3H/HeJ) mice, they identified a candidate gene on chromosome 4, Toll-like
receptor 4 (Tlr4). Ozone exposure (0.3 ppm for 24 to 72 h) of C3H/HeJ and C3H/HeOuJ mice,
the latter differing from the O3-resistant strain by a polymorphism in the coding region of Tlr4,
demonstrated greater protein concentrations in the OuJ strain. The two strains exhibited
differential expression of Tlr4 mRNA with O3 exposure. Thus, a quantitative trait locus on
chromosome 4 appears to be responsible for a significant portion of the genetic variance in
O3-induced lung hyperpermeability. In these mouse strains, lavageable protein concentration
was lowered by inhibition of iNOS and by targeted disruption ofNos2. Comparisons of
C3H/HeJ and C3H/HeOuJ O3 exposures demonstrated reduced Nos2 and Tlr4 mRNA levels in
the O3-resistant C3H/HeJ mice. These data are consistent with the hypothesis that O3-induced
lung hyperpermeability is mediated by iNOS. These studies suggest a role for the Toll-like
5-18
-------
receptor 4 (TLR4) in the host response to O3 similar to the role it has demonstrated in
lipopolysaccharide (LPS) sensitivity (Schwartz 2002; Wells et al. 2003). TLR4 signaling is
thought to be critical to linking the innate and acquired immune system through antigen
presenting cells and Thl/Th2 differentiation.
Ozone exposure has been shown to affect antibody responses both in vitro and in mice.
Becker et al. (1991) demonstrated changes in IgG production in cultured human lymphocytes
with O3 exposures of 1.0, 0.5, and 0.1 ppm for 2 h. Subsequent to O3 exposure, cells were
stimulated with pokeweed mitogen (PWM, a T-cell-dependent stimulus) or Staphylococcus
aureus Cowan 1 strain (SAC, a T-cell-independent stimulus). Both B and T cells were affected
by O3. T cells also demonstrated an increase in IL-6 and a decrease in IL-2, which suggested to
the authors that O3 may have direct effects on IgG producing cells and concurrently an effect that
is mediated by altered production of T cell immunoregulatory molecules. Responses to
repeated O3 (0.08 to 0.25 ppm) and ovalbumin (OVA, 1%) exposures were compared in "IgE-
high responder" (BALB/c) and "IgE-low responder" (C57BL/6) mice (Neuhaus-Steinmetz et al.,
2000). Ozone appeared to shift the immune response toward a Th2-like pattern in the two mouse
strains, with differing potentials for developing allergic reactions.
Another study (Depuydt et al., 2002) demonstrated that O3 (0.1 ppm for 2 h) increases
allergen-induced airway inflammation in previously sensitized mice but has no effect on the
sensitization process itself. This study used OVA-pulsed dendritic cells instead of systemic
adjuvant, which the authors consider a more relevant model of sensitization, as it clearly
separates the immune response from the challenge and does not obscure regulatory processes as
does intraperitoneal (i.p.) injections of OVA. They further suggest that dendritic cells, the
principal antigen-presenting cells in the airway, are an important component of O3-induced
eosinophilic airway inflammation.
Surfactant proteins A and D were shown to create an inflammatory feedback loop with
perturbations in lung immune defenses (reviewed in Hawgood and Poulain, 2001). Earlier
studies suggested that SP-A is a target for O3 toxicity by causing inhibition of SP-A self-
association and SP-A-mediated lipid vesicle aggregation. Further, O3 reduced the ability of
SP-A to inhibit phospholipid secretion by alveolar type II cells and reduced the capacity of SP-A
to induce superoxide anion production and enhance phagocytosis of herpes simplex virus.
Bridges et al. (2000) reported that both SP-A and SP-D directly protect surfactant phospholipids
5-19
-------
and macrophages from oxidative damage by blocking accumulation of thiobarbituric acid
reactive substances (TEARS) and conjugated dienes.
Eight human variants of SP-A in Chinese hamster ovary (CHO) cells exposed to O3 (1 ppm
for 4 h) showed decreased ability to stimulate cytokine (TNF-a and IL-8) production in THP-1
cells, a macrophage-like cell line (Wang et al., 2002). Each variant had a unique time- and
dose-dependent pattern of stimulation of cytokine production with O3 exposure, which the
authors attribute to possible differences in susceptibility to O3 oxidation. Targeted disruption of
mouse SP-A and SP-D (Hawgood et al., 2002) caused increases in BAL phospholipid,
macrophage, and protein through 24 weeks of age. Further, the deficient mice developed patchy
lung inflammation and air space enlargement consistent with emphysema. Future experiments
using these null mice will help to establish the role of SP-A and SP-D in pulmonary host defense
to O3 exposure.
5.2.2.4 Interactions with Infectious Microorganisms
Ozone-induced dysfunction of host defense systems results in enhanced susceptibility to
bacterial lung infections. Acute exposures of 0.08 ppm (3 h) O3 can overcome the ability of
mice to resist infection (by decreasing lung bactericidal activity) with Streptococcal bacteria,
resulting in mortality. Changes in antibacterial defenses are dependent on exposure regimens,
species and strain of test animal, species of bacteria, and age of animal, with young mice being
more susceptible to the effects of O3. The effect of O3 exposure on antibacterial host defenses
appears to be concentration- and time-dependent. Early studies using the mouse "infectivity
model," consisting of exposure to clean air or O3 followed by exposure to an aerosolized
microorganism, showed that the difference in mortality between O3-exposed groups and controls
is concentration-related. Chronic exposures (weeks, months) of 0.1 ppm do not cause greater
effects on infectivity than short exposures, due to defense parameters becoming reestablished
with prolonged exposures.
More recent studies of O3-induced modulation of cell-mediated immune responses showed
effects on the onset and persistence of infection. Cohen et al. (2001, 2002) exposed male F344
rats subchronically to either 0.1 or 0.3 ppm O3 for 4 h/day 15 days/week, for 1 or 3 weeks.
Subsequent exposure with viable Listeria monocytogenes demonstrated no observed effect on
cumulative mortality but did show a concentration-related effect on morbidity onset and
5-20
-------
persistence. These data suggest that O3 may cause a possible imbalance between Thl and Th2
cells, which can subsequently lead to suppression of the resistance to intracellular pathogens.
Effects of O3 on viral infections are dependent on the temporal relationship between O3
exposure and viral infection. Only high concentrations (1.0 ppm O3, 3 h/day, 5 days, mice)
increased viral-induced mortality. No detrimental effects were seen with a 120-day exposure to
0.5 ppm O3 on acute lung injury from influenza virus administered immediately before O3
exposure started. But there were O3-enhanced postifluenzal alveolitis and lung parenchymal
changes. As O3 does not affect lung influenza viral liters, it apparently does not impact antiviral
clearance mechanisms. In general, the evidence suggests that O3 can enhance both bacterial and
viral lung infections, but the key mechanisms have not yet been identified. New studies on the
interactions of O3 and viral infections have not been published.
5.2.2.5 Summary and Conclusions—Lung Host Defenses
New data on lung host defenses support earlier work which suggests that mucociliary
clearance is affected in most test species at just under 1 ppm, with lower levels (-0.1 ppm)
increasing clearance and somewhat higher levels decreasing clearance. These data also tend to
suggest mechanisms whereby O3 affects clearance, which include uptake being a direct effect of
O3, but modulated by ROS and hydroxyl radicals.
Alveolar macrophage function is disrupted by O3, as shown by several studies showing
inhibition of phagocytosis at concentrations ranging from 0.1 to 1.2 ppm. This inhibition returns
to control levels if O3 exposures are repeated for several days. Two new studies corroborate
earlier findings of increases in AM number in that same exposure range. In this environmentally
relevant exposure range, new studies support older findings of decreased resistance to microbial
pathogens, as shown by those endpoints examining superoxide radical formation, altered
chemotaxis/motility, decreased INF-y levels, decreased lysosomal activity, increased
prostaglandin E (PGE) levels, and increased NO mRNA and protein.
New research evaluating the effects of O3 on immune function advances previous work that
has shown that exposures can enhance or suppress immune responsiveness, depending on the
species studied, concentration of O3, route of exposure of allergen, and timing of exposure.
Continuous exposure to O3 impairs immune responses for the first several days of exposure,
followed by an adaptation to O3 that allows a return of normal immune responses. Most species
5-21
-------
show little effect of O3 exposures prior to immunization, but show a suppression of responses to
antigen in O3 exposures post-immunization. Evaluation of mouse strains with genetically-
determined differential sensitivity or resistance to O3 indicated a possible interaction between the
innate and acquired immune systems and, further, that O3 may shift the immune response toward
a Th2-like pattern. Work has also focused on the deleterious effects of O3 exposure on SP-A and
SP-D and their immunomodulatory function in protecting against oxidative stress.
Findings from several new studies evaluating the effects of O3 exposures on infectious
microorganisms are in concurrence with those from previous studies which showed, in general,
increased mortality and morbidity, decreased clearance, increased bacterial growth, and
increased severity of infection at exposure levels of 0.1 to 1 ppm O3 for 1 week.
5.2.3 Inflammation and Lung Permeability Changes
The normal lung has an effective barrier function that controls bidirectional flow of fluids
and cells between the air and blood compartments. Ozone disrupts this function, resulting in two
well-characterized effects of O3 exposure, lung inflammation and increased permeability, which
are distinct events controlled by independent mechanisms. Ozone initiates inflammation of lung
tissue by reactions with antioxidants and lipids in ELF (discussed in Section 5.2.1, see
Figure 5-2). Secondary reaction products generated in this process then cause changes in cell
membranes, disruption of the lung barrier leading to leakage of serum proteins, influx of PMNs,
release of bioactive mediators, and movement of compounds from the air spaces into the blood.
This increased permeability allows accumulation of co-occurring pollutants into the lung tissue.
The framework for presenting this stereotypical response to O3 consists of discussions covering:
(1) the time course of these changes, (2) concentration x time (C x T) relationships,
(3) susceptibility factors, (4) mediators of inflammation, and (5) NO and reactive nitrogen.
Rats appear to be more resistant to Os-induced inflammation than humans (see Chapter 4).
With comparable exposure protocols, both species have similar observed inflammatory and
permeability changes, i.e., controlled human exposure studies discussed in Chapter 6 indicate
that the majority of acute responses in humans are similar to those observed in animals.
Ozone also increases the permeability from the air to the blood compartment. Ozone
(0.8 ppm for 2 h) caused a 2-fold increase of the transport of labeled DTPA from the rat tracheal
lumen to the blood. This coincided with a 2-fold increase in the number of endocytic vesicles in
5-22
-------
epithelial cells that contained intraluminally instilled horseradish peroxidase (HRP) as a tracer.
These studies also suggest an uneven disruption of tight junctions and alternative transport via
endocytotic mechanisms. In studies aimed at detecting the effects of O3 exposure on regional
permeability, O3 increased the transmucosal transport of diethylenetriamine pentaacetic acid
(DTPA) and bovine serum albumin (BSA) more in the trachea and bronchoalveolar zone than in
the nose. These changes in barrier integrity may allow increased entry of antigens and other
bioactive compounds (e.g., bronchoconstrictors) into lung tissue. Data from analyses at regular
intervals PE indicate that maximal increases in BALF protein, albumin and number of PMNs
occur 8 to 18 h (depending on the study) after an acute exposure to 0.5 to 1.0 ppm O3 ceases.
Increases in permeability and inflammation have been observed at levels as low as
0.1 ppm O3 for 2 h/day for 6 days in rabbit and 0.12 ppm in mice (24 h exposure) and rats
(6 h exposure). After acute exposures, the influence of the time of exposure increases as the
concentration of O3 increases. The exact role of inflammation in the etiology of lung disease is
not known; nor is the relationship between inflammation and changes in lung function.
Table AX5-8 in Annex AX5 summarizes new key studies describing the potential for O3
exposure effects on lung permeability and inflammation.
5.2.3.1 Time Course of Inflammation and Lung Permeability Changes
A study of OVA-sensitized male Dunkin-Hartley guinea pigs exposed to 1.0 ppm O3 for
3 h showed that levels of PMN significantly increased at 3 h PE, but BAL protein levels did not,
suggesting a lack of correlation between the two endpoints (Sun et al., 1997). Increased PMN
without a concordant increase in BAL protein levels were found when the guinea pigs were
exposed to 1.0 ppm O3 for 1 h and evaluated 24 h PE. The first group also had an increase in
AHR, but not the second group, which suggests a dissociation between PMN levels and AHR.
Earlier work demonstrated that O3 exposures of 0.8 to 1 ppm in rat and guinea pig
transiently increase the permeability from the air to the blood compartment. This permeability is
greatest in tracheal and bronchoalveolar zones, and may allow increased entry of antigens and
other bioactive compounds (e.g., bronchoconstrictors) into lung tissues. The time course of the
influx of PMNs into the lung and the BALF fluid levels of macrophage inflammatory protein-2
(MIP-2) were found to be roughly similar to that for proteins (Bhalla and Gupta, 2000).
Adherence of neutrophils to pulmonary vascular endothelium is maximal within 2 h after
5-23
-------
exposure and returns to control levels by 12 h PE (Lavnikova et al., 1998). Cheek et al. (1995)
cultured monolayers of rat alveolar type II cells and exposed them to 0.1 or 0.5 ppm O3 for 0.5 h
to evaluate the effects of O3 on permeability. Permeability increased dose-dependently and the
higher exposures elicited greater numbers of injured epithelial cells. Exposure to 0.1 ppm O3
was thought to expedite the restoration of epithelial barrier functions, while at higher exposures,
neutrophils exacerbated the O3-induced injury. Vesely et al. (1999a) have demonstrated that
neutrophils contribute to the repair process in O3-injured airway epithelium and they may play a
role in removal of O3-injured cells.
Exposures of 3 to 7 days have been found to cause increases in BALF protein and PMNs
that typically peak after a few days (depending upon species tested and exposures) and return
toward control even with continuing exposure. Van Bree et al. (2002) observed lower
BALF levels of protein, fibronectin, IL-6, and inflammatory cells in rats exposed for 5 days to
0.4 ppm O3 than in rats exposed for 1 day, suggesting adaptation to O3 exposure. Postexposure
challenge with single O3 exposures at different time points showed recovery of susceptibility
to O3. McKinney et al. (1998) observed differences in IL-6 levels due to repetitive exposures
and demonstrated a role of IL-6 in the adaptive response induced by repeated O3 exposures of
0.5 ppm for 4 h.
5.2.3.2 Concentration and Time of Exposure
The relative influence of concentration and duration of exposure (i.e., C x T) has been
investigated extensively in rats, using BALF protein as an endpoint. Earlier work utilizing
concentrations of 0.1 to 2 ppm O3 and durations of 1 to 8 h has shown that the interaction
between C and T is complex. At these levels of exposure, concentration generally dominates the
response. The smallest C x T product causing an increase in protein was 0.52 ppm • h (0.13 ppm
x 4 h). C x T studies using the endpoints of changes in lung protein or cell type showed that
acute damage is a function of cumulative dose. The impact of T is C-dependent (at higher Cs,
the impact of T is greater); at the lowest C and T values, this dependence appears to be lost. The
controlled human exposure data described in Chapter 6 concur with most animal data, showing
that (a) concentration of O3 is an important factor determining O3 responses and (b) duration of
exposure and ventilation rate also exert substantial influences.
5-24
-------
New studies evaluating C x T relationships in animal models have not been completed.
However, a full understanding of C x T relationships in ambient exposures must include the
recognition that 'real world' exposures are cyclic in nature, due to the daily and seasonal
variations in O3 levels. The concentration of O3, the duration of the exposure, and duration
between exposures are all relevant to determining the type and level of O3-induced injury.
5.2.3.3 Susceptibility Factors
Factors that have been studied for potential impact on the effects of O3 exposure include
age, gender, nutritional status, exposure to copollutants, exercise, and genetic variability. A full
characterization of the effects of age on O3 responses has not been completed. Data available
indicate that effects of age on O3 responses are endpoint-dependent, with young mice, rats, and
rabbits having greater prostaglandin levels with exposure and senescent rats having greater IL-6
and N-acetyl-p-D-glucosaminidase levels with exposure.
A study (Johnston et al., 2000a) compared gene expression of chemokines and cytokine in
newborn and 8-week-old C57BL/6J mice exposed to 1.0 or 2.5 ppm for 4, 20, or 24 h. The
newborn mice displayed increased levels of MT mRNA only, while the 8-week-old mice had
increases in MlP-la, MIP-2, IL-6, and MT mRNA. Comparisons were made with mice of the
same age groups with exposures to endotoxin (10 ng/mouse for 10 min). Both age groups
displayed similar cytokine/chemokine profiles with endotoxin exposure. This suggested to the
authors that the responses to endotoxin (which does not cause epithelial injury) and the
responses to O3 (which does) demonstrate that differences in inflammatory control between
newborn and adult mice is secondary to epithelial injury.
Pregnancy and lactation increased the susceptibility of rats to acute O3, but no clear effects
of gender have been identified. The effects of vitamin C deficiency on O3 responses are unclear.
Ascorbate-deficient guinea pigs exposed to O3 demonstrated only minimal effects on injury and
inflammation (Kodavanti et al., 1995). Utilizing a diet-restricted (20% of the freely fed diet) rat
model, Elsayed (2001) demonstrated higher survivability upon exposure to higher O3 (0.8 ppm
continuously for 3 days) compared to freely fed rats. Preexposure to sidestream cigarette
smoke had been found to cause increased lung injury (Yu et al., 2002). In vitro studies on the
macrophages from smoke + O3- exposed animals responded by a greater release of TNF-a
5-25
-------
following LPS stimulation when compared to macrophages exposed to air, smoke, or O3
(0.5 ppm for 24 h) alone.
Lines of evidence illustrate that genetic background is an extremely important determinant
of susceptibility to O3. Earlier studies using inflammation-prone (susceptible) C57BL/6J (B6)
and inflammation-resistant C3H/HeJ (C3) mouse strains and high doses of O3 (2 ppm for 3 h)
identified Inf-2 as a locus controlling susceptibility. Further studies in these two strains of mice
using more relevant exposures (0.3 ppm for 72 h) identified that the acute and subacute
exposures are controlled by two distinct genes, referred to as Inf-\ and Inf-2, respectively
(Tankersley and Kleeberger, 1994). Exposures to 0.3 ppm O3 for 48 or 72 h, when repeated
14 days after the initial exposures, caused a smaller increase in BALF protein and numbers of
macrophages, lymphocytes, and epithelial cells in both strains, but PMN number was greater in
both strains compared to initial exposure (Paquette et al., 1994). Kleeberger et al. (1997) also
identified another potential susceptibility gene, tumor necrosis factor (Tnf, which codes for the
pro-inflammatory cytokine TNF-a ) on a qualitative trait locus on mouse chromosome 17.
By neutralizing the function of TNF-a with a specific antibody, they were able to confer
protection against O3 (0.3 ppm, 48 h) injury in susceptible mice. They then demonstrated a role
for TNF receptor 1 and 2 (TNFR1 and TNFR2, respectively) signaling in subacute (0.3 ppm for
48 h) O3-induced pulmonary epithelial injury and inflammation (Cho et al., 2001). TNFR1 and
TNFR2 knockouts were less sensitive to subacute O3 exposure than were WT C57BL/6J mice.
An integrated and more comprehensive effort to identify the genetic basis for the
susceptibility to O3-induced lung injury was reported by Savov et al. (2004). In this report, acute
lung injury in response to high O3 dose (2 ppm for 3 h) was assessed and integrated with
physiological, biochemical, and genetic observations using nine inbred mouse strains. This work
indicated the presence of genetic loci on chromosomes 1, 7, and 15 associated with phenotypic
characteristics for resistance to acute O3-induced lung injury. They identified C3Ft/HeJ and
A/J as consistently O3-resistant; C57BL/6J and 129/SvIm as consistently O3-vulnerable; and
CAST/Ei, BTBR, DBA/2J, FVB/NJ, and BALB/cJ as intermediate in response to O3.
Ozone-induced changes in CCSP (called CC16 by this group) expression were evaluated in
five inbred mouse strains: C57BL/6J and CBA both considered sensitive to acute O3-induced
inflammation, C3H/HeJ and AKR/J both considered resistant, and SJL/J considered intermediate
(Broeckaert et al., 2003). Two exposure protocols, 1.8 ppm O3 for 3 h or 0.11 ppm O3, 24 h/day
5-26
-------
for up to 3 days, were used; and BALF and serum were assayed immediately after exposure or at
6 h PE. Both exposure levels caused a transient increase in CC16 in serum that correlated with
BALF changes in protein, LDH, and inflammatory cells. There was an inverse relationship
between preexposure levels of CC16 in BALF and epithelial damage, based on serum CC16
levels and BALF markers of inflammation. There was also an inverse relationship between
preexposure levels of albumin in BALF and lung epithelium damage. Based on these results, the
authors concluded that a major determinant of susceptibility to O3 is due to differences in the
basal permeability of the airway epithelium. As all of the mouse strains had similar levels of
preexposure CC16 mRNA, they explored the possible role of CC16 isozymes in differences
among strains. The CC16 monomer, a 7kD protein, exists in two isoforms with differing pi
values, CC16a (4.9) and CC16b (5.2). To evaluate the role of CC16 isoform profiles in
permeability differences between C57BL/6J and C3H/HeJ, this group evaluated the CC16
protein profiles in BALF of the strains before and after O3 exposure following two-dimensional
protein electrophoresis analysis. C57BL/6J mice had lower levels of CC16a (the more acidic
form) than C3FI/HeJ. But both the strains had similar levels of CC16b. Based on these
observations, Broeckaert et al. (2003) concluded that greater epithelial permeability observed in
C57BL/6J mice may be due to differences in the expression of CC16a and, possibly, other
antioxidant/anti-inflammatory proteins.
Wattiez et al. (2003) examined BALF protein from C57BL/6J (O3-sensitive) and
C3FI/HeJ (O3-resistant) mice exposed to filtered air, using a two-dimensional polyacrylamide
gel approach to analyze the protein profiles. C3H/HeJ mice expressed 1.3 times more Clara cell
protein!6 (CC16) than C57BL/6J mice and, further, expressed more of the acidic isoform of
CC16. Strain-specific differential expression of isoforms of the antioxidant protein 2 (AOP2),
the isoelectric point 5.7 isoform in C3H/HeJ and isoelectric point 6.0 isoform in C57BL/6J, were
observed. These data suggest a protective role for CCSP against oxidative damage and, further,
that genetic susceptibility to oxidant stress may be moderated, in part, by the gene coding for
CCSP. Taken together, these mouse studies of genetic susceptibility are useful for helping to
understand underlying mechanisms potentially leading to O3-induced effects. However, at this
point, corresponding human polymorphisms have not yet been identified as being associated
with differing human sensitivities to O3.
5-27
-------
5.2.3.4 Mediators of Inflammatory Response and Injury
Ozone reacts with lipids in the ELF or epithelial cell membranes, creating ozonation
products which then stimulate airway epithelial cells, AMs, and PMNs to release a host of
pro-inflammatory mediators, including cytokines, chemokines, ROS, eicosanoids, and PAF (see
Figure 5-3). While neutrophils in the lung characterize an inflammatory response to O3, the
release of chemotactic mediators by inflammatory cells indicates their state of activation and
their role in continued inflammation and injury. At O3 exposures of > 1 ppm in rodents, these
mediators recruit PMNs and increase expression of MIP-2 mRNA or BALF levels of MIP-2
(Driscoll et al., 1993; Haddad et al., 1995; Bhalla and Gupta, 2000). The increased mRNA
expression was associated with an increased neutrophilia in the lung. Harkema et al. (1987)
reported neutrophilic rhinitis in monkeys exposed to 0.15 ppm O3 for 6 days. Zhao et al. (1998)
showed that 0.6 ppm O3 exposure for 2 h in mice and rats causes an increase in monocyte
chemotactic protein-1 (MCP-1).
Fibronectin, an extracellular matrix glycoprotein, is thought to have a role in lung
inflammation and inflammatory disorders, and has been shown to be increased with exposure to
1 ppm O3 for 14 days. Gupta et al. (1998) observed an increase in both fibronectin protein and
mRNA expression in the lung of rats exposed to 0.8 ppm O3 for 3 h. A mechanistic role of
fibronectin in O3-induced inflammation and injury was suggested on the basis of comparability
of temporal changes in BALF protein, fibronectin, and alkaline phosphatase activity with
exposures of 1 ppm for 3 h (Bhalla et al., 1999). Studies have reported an effect of O3 on other
cytokines and inflammatory mediators. An increase occurred for cytokine-induced neutrophil
chemoattractant (CINC) and NF-KB expression in vivo (Koto et al., 1997); for IL-8 in vivo and
in vitro (Chang et al., 1998); TNF-a, fibronectin, IL-1 and CINC release by macrophages
ex vivo (Pendino et al., 1994; Ishii et al., 1997); and NF-KB and TNF-a (Nichols et al., 2001; see
Section 6.9.2 of Chapter 6). An increase in lung CINC mRNA occurred within 2 h after the end
of a 3 h exposure of rats to 1 ppm O3. The CINC mRNA expression was associated with
neutrophilia at 24 h PE. Exposure of guinea pig AMs recovered in BALF and exposed in vitro
to 0.4 ppm O3 for 1 h produced a significant increase in IL-6 and TNF-a (Arsalane et al., 1995).
An exposure of human AMs to an identical O3 concentration increased TNF-a, IL-lp, IL-6, and
IL-8 and their mRNAs. Ozone exposure (0.3 to 2.5 ppm, 1 to 48 h) of mice caused an increase
in IL-6, MlP-la, MIP-2, eotaxin, and MT abundance (Johnston et al., 1999a). The IL-6 and MT
5-28
-------
increase was enhanced in mice deficient in CCSP, suggesting a protective role of Clara cells and
their secretions (Mango et al., 1998). CCSP deficiency, also increased sensitivity of mice to O3,
as determined by an increase in abundance of MlP-la and MIP-2 following a 4 h exposure to
1.0 ppm O3 (Johnston et al., 1999b).
Mast cells, which are located below the epithelium, release pro-inflammatory mediators
and have been shown to contribute to O3-induced epithelial damage. Greater increases in
lavageable macrophages, epithelial cells and PMNs were observed in mast cell-sufficient mice
than in mast cell-deficient mice exposed to 0.26 ppm O3 for 8 h per day, 5 days per week
(Kleeberger et al., 2001b). Increases in inflammatory cells were also observed in mast
cell-deficient mice, however, O3-induced permeability changes were similar between genotypic
groups exposed to 0.26 ppm. When an RBL-2H3 mast cell line was exposed to 0.1 to 1.0 ppm
O3 for 1 h, spontaneous release of serotonin and modest generation of prostaglandin D2 (PGD2)
occurred only under conditions that caused cytotoxicity (Peden and Dailey, 1995). Additionally,
O3 inhibited IgE- and A23187-induced degranulation. Mast cells recovered from O3-exposed
peripheral airways of ascaris-sensitive dogs released significantly less histamine and PGD2
following in vitro challenge with ascaris antigen or calcium ionophore (Spannhake, 1996).
Ozone (0.4 ppm, 5 weeks) exposure also promoted eosinophil recruitment in the nose and
airways in response to instillation of OVA or OVA-pulsed dendritic cells and aggravated allergy
like symptoms in guinea pigs (lijima et al., 2001).
The role of PMNs and cellular mediators in lung injury and epithelial permeability has
been investigated using antibodies and inhibitors of known specificity to block inflammatory cell
functions and cytokine activity. Treatment of rats with cyclophosphamide prior to O3 exposure
(0.8 ppm, 48 h) resulted in a decreased recovery of PMNs in the BALF and attenuated
permeability induced by O3 (Bassett et al., 2001).
Pretreatment of animals with antiserum against rat neutrophils abrogated PMN
accumulation in the lung, but did not alter permeability increase produced by O3. Studies
utilizing antibodies to selected pro- or anti-inflammatory cytokines suggest a role of TNF-a,
IL-10, and IL-lp in O3-induced changes in permeability, inflammation and cytokine release
(Ishii et al., 1997; Reinhart et al., 1999; Bhalla et al., 2002) in exposures of ~ 1 ppm for 3 to 6 h.
An attenuation of O3-induced increases in permeability and inflammation was also observed in
mice treated either before or after exposure with UK-74505, a PAF receptor antagonist
5-29
-------
(Longphre et al., 1999). These results were interpreted to indicate that O3-induced epithelial and
inflammatory changes are mediated in part by activation of PAF receptors.
Ozone exposure stimulates macrophage motility toward a chemotactic gradient, and
macrophages isolated from rats exposed to 0.8 ppm O3 for 3 h adhered to epithelial cells
(ARL-14) in culture to a greater extent than macrophages from air-exposed controls (Bhalla,
1996). Both macrophage motility and chemotaxis were attenuated by antibodies to cell adhesion
molecules CD-I Ib and ICAM-1, suggesting a role for cell adhesion molecules in O3-induced
cellular interactions. This may also explain the increased tissue localization and reduced
recovery of macrophages in BALF (Pearson and Bhalla, 1997) following O3 exposure (0.8 ppm,
3 h). Studies investigating the mechanisms of PMN recruitment in the lung have explored the
role of cell adhesion molecules that mediate PMN-endothelial interactions. An exposure of
female rats to O3 (1 ppm, 2 h) had an attenuating effect on CD-18 expression on AMs and
vascular PMNs, but the expression of CD62L, a member of selection family, on vascular PMNs
was not affected (Hoffer et al., 1999). In monkeys, O3-induced (0.8 ppm, 8 h) inflammation was
blocked by treatment with a monoclonal antibody to CD 18, suggesting dependence of PMN
recruitment on this adhesion molecule (Hyde et al., 1999). Treatment of monkeys with CD18
antibody also reduced tracheal expression of the P6 integrin (Miller et al., 2001), suggesting that
lung epithelial cell expression of this adhesion molecule is associated with sites of neutrophil
recruitment. A single 3 h exposure of rats to 1 ppm O3 caused an elevation in concentration of
ICAM-1, but not CD-18, in the BALF (Bhalla and Gupta, 2000). Takahashi et al. (1995a) found
an increase in tissue expression of ICAM-1 in mice exposed to 2 ppm O3 for 3 h, noting a
temporal correlation of inflammatory activity and ICAM-1 expression, which varied in different
regions of the lung. A comparable pattern of time-related changes in total protein, fibronectin
and alkaline phosphatase activity in the BALF of rats exposed to 0.8 ppm O3 for 3 h was also
noted by Bhalla et al. (1999). Together, these studies support the role of extracellular matrix
protein and cell adhesion molecules in the induction of lung inflammation and injury.
5.2.3.5 The Role of Nitric Oxide Synthase and Reactive Nitrogen in Inflammation
Nitric oxide is a messenger molecule involved in many biological processes, including
inflammation (see Figure 5-3). Cells in the respiratory tract (including mast cells, neutrophils,
epithelial cells, neurons, and macrophages) produce three different forms of nitric oxide synthase
5-30
-------
(NOS), the enzyme that catalyzes the formation of NO. NOS-l(neuronal) and NOS-3
(endothelial) are constitutively expressed, whereas NOS-2 (also referred to as iNOS) is
inducible, commonly by pro-inflammatory cytokines. Macrophages isolated from O3-exposed
(0.8 ppm for 3 h) mice produced increased amounts of NO, superoxide anion, and PGE2,
but production of these mediators by macrophages from NOS knockout mice was not
elevated (Fakhrzadeh et al., 2002). Additionally, mice deficient in NOS or mice treated
with N -monomethyl-L-arginine, an inhibitor of total NOS, were protected from O3-induced
permeability, inflammation, and injury, suggesting a role for NO in the production of O3 effects
(Kleeberger et al., 2001a; Fakhrzadeh et al., 2002). These results contrast with a study showing
that O3 exposure (of 1 ppm for 8 h/night for 3 nights) produced greater injury, as determined by
measurement of MTP-2, matrix metalloproteinases, total protein, cell content, and tyrosine
nitration of whole lung protein, in iNOS knockout mice than in wild-type mice (Kenyon et al.,
2002). This group suggests that protein nitration is related to inflammation and is not dependent
on iNOS-derived NO. They point out the possible experimental differences, such as O3
concentration, for inconsistency between their results and those of Kleeberger et al. (2001a).
Rats that were pretreated with ebselen (a potent anti-inflammatory, immunomodulator, and
NO/peroxynitrite scavenger) and then were exposed to 2 ppm O3 for 4 h had decreased numbers
of neutrophils, lowered albumin levels, and inhibited nitration of tyrosine residues in BALF 18 h
PE, though macrophage iNOS expression was not changed (Ishii et al., 2000a). These results
suggest an iNOS-independent mechanism for O3-induced inflammation. Jang et al. (2002)
showed dose-dependent increases in nitrate (indicative of in vivo NO generation) with O3
exposure (0.12, 0.5, 1, or 2 ppm for 3 h). Functional studies of enhanced pause (Penh)
demonstrated increases with O3 exposure which were also dose-dependent. Western blot
analysis of lung tissue showed increases in NOS-1 but not in NOS -3 or iNOS isoforms. These
results suggest that, in mice, NOS-1 may induce airway responsiveness by a neutrophilic airway
inflammation. The literature regarding the effects of O3 exposure on NOS activity is complex
and conflicting. Similarly, the issue of protein nitration as it relates to cell injury due to O3
exposure is somewhat controversial.
5-31
-------
5.2.3.6 Summary and Conclusions—Inflammation and Permeability Changes
Figure 5-3 depicts many of the inflammatory and permeability changes that occur with O3
exposure. The figure also demonstrates links between inflammatory/permeability responses and
altered lung function (discussed in Section 5.2.5), ciliary motility (Section 5.2.2.1), AHR
(Section 5.2.5.3), and possible thrombolytic effects (Section 5.3.3). Airway mucosa in the
normal lung serves as an effective barrier that controls bidirectional flow of fluids and cells
between the air and blood compartments. Ozone disrupts this function, resulting in a cascade of
effects that includes an increase in serum proteins, bioactive mediators, and PMNs in the
interstitium and air spaces of the lung. Damaged epithelial cells release cytokines, which
function to recruit and activate AMs and PMNs. PMN recruitment into the lung is maximal at
several hours PE. PMN recruitment is followed by recruitment of blood monocytes that enter
the lung and enlarge to become AMs. The AMs persist for days to weeks, phagocytizing injured
cells. Activated PMNs and AMs continue the cascade of effects by further releasing
inflammatory mediators, which serve to amplify the initial effects of O3. Generally, the
initiation of inflammation is an important component of the defense process; however, its
persistence and/or repeated occurrence can result in adverse health effects. Activation of this
inflammatory cascade takes several hours. Chemical mediators released early in the cascade
contribute to effects on pulmonary function. Events occurring later in the cascade, by
which time O3-induced alterations in pulmonary function have attenuated, are related to
sustained inflammation. Further, mechanistic separation of inflammation, permeability, and
AHR is suggested by the temporal disparities of their increases. The O3-induced disruption of
the tight junctions between epithelial cells also increases the permeability between the air and
blood compartments. This disruption, occurring with exposures of 0.8 ppm for 2 h, is greater in
the trachea and bronchoalveolar zone than in the nose and allows entry of particles, including
bioactive compounds, into lung tissue.
For environmentally-relevant exposures to O3, exposure concentration dominates the
response. Studies evaluating C x T relationships have not been published recently. Other
factors that have been studied for potential impact on the effects of O3 include age, gender,
nutritional status, genetic variability, exercise and exposure to copollutants. The effects of age
on lung inflammation are not well known. After an acute exposure to 0.8 or 1 ppm, young mice,
rats, and rabbits had greater changes in prostaglandins in BALF, but there were no
5-32
-------
age-dependent effects on BALF protein or cell number. Comparisons of male and female
animals, and ascorbate deficiency did not reveal significant differences in the effects of O3, but
exercise during exposure increased susceptibility.
Important new work has revealed that susceptibility to O3 is, in part, genetically
determined. Mouse strains with differing sensitivities to O3 have identified genes on separate
loci controlling various aspects of inflammation, providing additional evidence for the
mechanistic separation of responses to O3. Such research is summarized in Figure 5-4.
Kleeberger's group has identified Inf-l, which modulates acute inflammatory responses; Inf-2,
which modulates responses to subacute exposures; and TNF-a and TNF receptors, which are
involved in inflammatory responses. Other research groups have identified loci linked to other
endpoints. This line of research provides a groundwork for understanding the underlying
mechanisms of O3-induced injury and can shed light on genes responsible for human
susceptibility to O3.
• AOP2
oxidative stress
(Wattiez et al., 2003)
1 C
0)
O
to
O
O
O
0)
S 15C
O
11
6h AHR
24 h inflammation
(Savovet a I., 2004)
„,—>Toll-like receptor-4
(Kleebergeretal.,2000)
—> 8 h inflammation
6 h protein
(Savoy et al., 2004)
PMN influx
(Kleebergeretal., 1997)
17
19C
CC16 •
oxidant stress
(Mango et al., 1998)
epithelial permeability
(Broeckaert et al.. 2003)
-> Tnfa
(Kleebergeretal., 1997)
Figure 5-4. Mouse chromosomes on which genes or gene loci have been identified that
modulate responses to ozone.
5-33
-------
Recent studies have placed a major focus on mediators released from inflammatory cells to
understand the mechanisms of O3-induced inflammation and injury. Cytokines and chemokines
have been shown to be released as a result of stimulation or injury of macrophages, epithelial
cells, and PMNs. Exposure of guinea pig AMs recovered in BALF and exposed in vitro to
0.4 ppm O3 produced a significant increase in IL-6 and TNF-a. An exposure of human AMs
to an identical O3 concentration increased TNF-a, IL-lp, IL-6, and IL-8. The expression
of MIP-2 mRNA or BALF levels of MIP-2 increased in mice and rats exposed to O3
concentrations > 1 ppm. An increase after O3 exposure has also been reported for other cytokines
and inflammatory mediators, including CINC and fibronectin. The CINC mRNA expression
was associated with neutrophilia at 24 h PE. Ozone exposure of mice also caused an increase in
IL-6, MlP-la and eotaxin in mice. Further understanding of the role of mediators has
come from studies utilizing antibodies and inhibitors of known specificity. In these studies,
treatment of rats with an anti-IL-6-receptor antibody prior to a nighttime exposure to O3
abolished O3-induced cellular adaptive response following a subsequent exposure. Studies
utilizing antibodies to selected pro- or anti-inflammatory cytokines suggest a role for TNF-a,
IL-10 and IL-lp in O3-induced changes in permeability, inflammation, and cytokine release.
Studies investigating the mechanisms of PMN recruitment in the lung have explored the
role of cell adhesion molecules that mediate PMN-endothelial cell interactions. An increase in
tissue expression of ICAM-1 occurred in mice exposed to 0.8 ppm O3. A comparable pattern of
time-related changes in total protein, fibronectin, and alkaline phosphatase activity in the BALF
was observed in rats exposed to 1 ppm O3. In monkeys, the O3-induced inflammation was
blocked by treatment with a monocolonal antibody to CD 18, suggesting dependence of PMN
recruitment on this adhesion molecule. Together, these studies support the role of extracellular
matrix protein and cell adhesion molecules in lung inflammation and injury.
Ozone exposure also affects macrophage functions and, consequently, their role in lung
inflammation. Macrophages isolated from O3-exposed mice produced increased amounts of
NO, superoxide anion, and PGE2, but production of these mediators by macrophages from
NOS knockout mice was not elevated. Additionally, mice deficient in NOS or mice treated
with NG-monomethyl-L-arginine, an inhibitor of total NOS, were protected from O3-induced
permeability, inflammation, and injury. These findings suggest a role for NO in the production
of O3 effects.
5-34
-------
5.2.4 Morphological Effects
Most mammalian species show generally similar morphological responses to <1 ppm O3
that differ only by region, cell type, exposure parameters, and length of time between exposure
and examination. Constant low exposures to O3 create an early bronchoalveolar exudation that
declines with continued exposure and drops sharply in the PE period. Epithelial hyperplasia also
starts early, increases in magnitude for several weeks, plateaus with continuing exposure, and
declines slowly during PE. Interstitial fibrosis has a later onset, continues to increase throughout
the exposure, and can continue to increase after the exposure ends. A schematic comparison of
these duration-response profiles is presented in Figure 5-5 (Dungworth, 1989). Nonhuman
primates respond more than rats at this concentration, due to differences in antioxidants, the
CAR (predicted to receive the highest dose of O3), the presence of respiratory bronchioles, acinar
volume, and differences in the nasal cavity's ability to "scrub" the O3. Ciliated epithelial cells of
the airway, Type I epithelial cells of the gas-exchange region, and ciliated cells in the nasal
cavity are the cells most affected by O3. Ozone-damaged ciliated cells are replaced by
nonciliated cells, which are unable to provide clearance function, and Type I cells are replaced
by Type II cells, which are thicker and produce more lipids. Inflammation also occurs,
especially in the CAR, wherein the tissue is thickened as collagen accumulates. At exposures of
0.25 ppm O3 (8 h/day, 18 months) in monkeys, the distal airway was found to be remodeled, as
bronchiolar epithelium replaces the cells present in alveolar ducts. In both rodents and monkeys,
it appears that the natural seasonal patterns of O3 exposure alter morphology more than
continuous exposures; thus, long-term animal studies with uninterrupted exposures may
underestimate morphological effects.
5.2.4.1 Acute and Subchronic Exposure Effects
Morphological effects of key acute and subchronic exposure studies are summarized in
Table AX5-9. Harkema et al. (1997a) reviewed toxicological studies of the nasal epithelial
response to short-term O3. New information regarding the effects of O3 in this region include
results indicating that the topical anti-inflammatory corticosteriod fluticasone propionate
prevents inflammation and mucous cell metaplasia in rats after cumulative O3 exposure (0.5 ppm
O3, 8 h/day, for 3 or 5 days) (Hotchkiss et al., 1998). Exposure to bacterial endotoxin, a
common ambient air toxicant, can potentiate mucous cell metaplasia in the nasal transitional
5-35
-------
CO
Exposure
Postexposure
6 1 mo
days
6 mo
Time
12 mo
Figure 5-5. Schematic comparison of the duration-response profiles for epithelial
hyperplasia, bronchoalveolar exudation, and interstitial fibrosis in the
centriacinar region of lung exposed to a constant low concentration of ozone.
Source: Dungworth(1989).
epithelium of rats caused by a previous 3 day 0.5 ppm O3 exposure (Fanucchi et al., 1998). Male
F344/N Hsd rats were intranasally instilled with endotoxin after exposure to filtered air (FA) or
0.5 ppm O3, (8 h/day for 3 days). Mucous cell metaplasia was not found in the air/endotoxin
group but was found in the O3/saline group and was most severe in the O3/endotoxin group. A
similar synergistic effect was demonstrated by Wagner et al. (2001a,b) with exposure of Fischer
rats to O3 and/or endotoxin for 8 h per day for 3 days. Ozone alone created epithelial lesions in
the nasal transitional epithelium, while endotoxin alone caused lesions in the epithelium of the
nose and conducting airways. The enhanced O3-induced mucous cell metaplasia was related to
neutrophilic inflammation.
Pre-metaplastic responses, such as mucin mRNA upregulation, neutrophilic inflammation,
and epithelial proliferation, were shown to be responsible for O3-induced mucous cell metaplasia
in the transitional epithelium of rats (Cho et al., 1999a, 2000). Male F344/N rats exposed to O3
(0.5 ppm, 8 h/day for 1, 2, or 3 days) showed a rapid increase in an airway-specific mucin gene
5-36
-------
mRNA after exposure to O3, both before and during the onset of mucous cell metaplasia.
Neutrophilic inflammation coincided with epithelial DNA synthesis and upregulation of
rMuc-5 AC but was resolved before the development of epithelial metaplasia. The mucous cell
metaplasia was neutrophil-dependent, whereas O3-induced epithelial cell proliferation and mucin
gene upregulation were neutrophil-independent.
Dormans et al. (1999) compared the extent and time course of fibrotic changes in mice,
rats, and guinea pigs exposed to 0.2 or 0.4 ppm O3 for 3, 7, 28, and 56 days. They found a
concentration-related centriacinar inflammation in all three species, with a maximum after
3 days of exposure and total recovery within 3 days after exposure. Repair of O3-induced
damage by removal of injured epithelial cells is enhanced by the influx of neutrophils (Hyde
et al., 1999; Veseley et al., 1999b; Miller et al., 2001; see Section 5.2.3). A study examining the
kinetics of early cellular responses to O3 utilized bromodeoxyuridine to label S-phase cells
(Hotchkiss et al., 1997). Labeling indices for rat nasal transitional epithelial cell DNA were
greatest 20 to 24 h after O3 (0.5 ppm for 8 h) exposure, and remained greater than control at
36 h PE.
Very few published studies have explicitly explored susceptibility factors such as species,
gender, age, antioxidant defense, acute and chronic airway disease, and exercise. Most typical
laboratory species studied have qualitatively similar effects associated with O3 exposure.
Dormans et al. (1999) compared morphological, histological, and biochemical effects in the rat,
mouse, and guinea pig following O3 exposure and recovery in clean air. Wistar RIV:Tox male
rats, specific-pathogen free (SPF)-bred National Institutes of Health (NIH) male mice, and
Hartley Crl:(HA)BR male guinea pigs were continuously exposed to FA, 0.2 or 0.4 ppm for 3, 7,
28, and 56 days. Recovery from 28 days of exposure was studied at intervals of 3, 7, and 28
days PE. The mouse was the most sensitive as shown by a concentration and exposure
time-dependent persistence of bronchiolar epithelial hypertrophy, elevated lung enzymes, and
slow recovery from exposure. Exposure to the high dose for 56 days in both rats and guinea pigs
caused increased amounts of collagen in ductal septa and large lamellar bodies in Type II cells.
The inflammatory response was greater in the guinea pig. Overall, the authors rated mice as
most susceptible, followed by guinea pigs and rats. However, this ranking was done without
first adjusting for differences in species-specific delivered dose of O3.
5-37
-------
Ferrets, monkeys, and rats were exposed to O3 (1.0 ppm, 8 h) to compare airway effects
(Sterner-Kock et al., 2000). The ferrets and monkeys had similar epithelial necrosis and
inflammation that was more severe than that found in rats. Because ferrets have a similar
pulmonary structure as humans (e.g., well-developed respiratory bronchioles and submucosal
glands), the authors concluded that the ferret would be a better model than rodents
for O3-induced airway effects. Age susceptibility is dependent on the endpoint examined
(see Chapter 4 for discussions of age-related differences in O3 dosimetry). One study (Dormans
et al., 1996) demonstrated that O3-induced centriacinar lesions are larger in younger rats than in
older rats with exposures to 0.4 ppm O3 for 1 to 7 days.
New studies have examined O3-induced morphological effects in compromised laboratory
animals. Rats with endotoxin-induced rhinitis were more susceptible to mucous cell metaplasia
in the nasal transitional epithelium caused by a 3 day exposure to 0.5 ppm O3 (Cho et al.,1999b).
Wagner et al. (2002) reported a similar O3-induced enhancement of inflammatory and epithelial
responses associated with allergic rhinitis. Brown Norway rats were exposed to 0.5 ppm O3,
8 h/day for 1 day or 3 consecutive days and then immediately challenged intranasally with either
saline or OVA. Multiple exposures to O3 caused greater increases in mucosubstances produced
in the nose by allergen challenge.
Recent research has focused on the concept of O3-susceptible and -nonsusceptible sites
within the respiratory tract, including in situ antioxidant status and metabolic activity. Plopper
et al. (1998) examined whether the variability of acute epithelial injury to short-term O3
exposure within the tracheobronchial tree is related to local tissue doses of O3 or to local
concentrations of GSH. Adult male rhesus monkeys exposed to O3 (0.4 or 1.0 ppm for 2 h)
demonstrated significant cellular injury at all sites, but the most damage, along with increased
inflammatory cells, occurred in the proximal respiratory bronchiole. A significant increase in
GSH was found in the proximal bronchus at 0.4 ppm O3 and a decrease in the respiratory
bronchiole at 1.0 ppm O3. A significant decrease in the percentage of macrophages, along with
significant increases in the percentage of neutrophils and eosinophils, and a doubling of total
lavage protein, were found after exposure to 1.0 ppm O3 only. The authors concluded that the
variability of local O3 dose in the respiratory tract was related to inhaled O3 concentration and
was closely associated with local GSH depletion and with the degree of epithelial injury.
5-38
-------
Plopper and colleagues (e.g., Watt et al., 1998; Paige et al., 2000a) explored the site-
specific relationship between epithelial effects of O3 exposure and the metabolism of
bioactivated compounds within the respiratory tract of rats. The distribution of CYP 2E1-
dependent activity, measured with a selective substrate (p-nitrocatechol), was found to be
highest in the distal bronchioles and minor daughter airways and lower in the lobar bronchi and
major daughter airways. Short-term O3 exposure (1 ppm for 8 h) increased CYP 2E1 activity in
the lobar bronchi/major daughter airways only; however, long-term O3 exposure (1 ppm for
90 days) decreased CYP 2E1 activity in the major and minor airways, further complicating the
interpretation of O3 effects based on concentration and duration of exposure and recovery. Rats
treated i.p. with 1-nitronaphthalene, a pulmonary toxicant requiring metabolic activation, and
exposed to 0.8 ppm O3, 8 h/day for 90 days showed greater histopathologic and morphometric
effects in the CAR of the lung (Paige et al., 2000b). Despite reported tolerance to oxidant stress
after long-term O3 exposure, there was increased severity of ciliated cell toxicity.
5.2.4.2 Summary of Acute and Subchronic Morphological Effects
Short-term exposures to O3 cause similar alterations in lung structure in a variety of
laboratory animal species. Cells in the CAR are the primary targets of O3, but ciliated epithelial
cells in the nasal cavity and airways and Type I epithelial cells in the gas exchange region are
also targets. New work has shown that a topical anti-inflammatory corticosteroid can prevent
these effects in nasal epithelia, while exposure to bacterial endotoxin can potentiate the effects.
Ozone-induced fibrotic changes in the CAR are maximal at 3 days of exposure and recover
3 days PE with exposures of 0.2 ppm in rodents. New studies of susceptibility factors
demonstrated that ferrets and monkeys have similar inflammatory and necrotic responses to
1 ppm O3, which differs from lesser injury seen in rats. Rats with induced allergic rhinitis are
more susceptible to 0.5 ppm O3 than are control animals. Important new work has demonstrated
variability of local O3 dose and subsequent injury in the respiratory tract due to depletion of
GSH. The proximal respiratory bronchiole receives the most acute epithelial injury from
exposures < 1 ppm, while metabolic effects were greatest in the distal bronchioles and minor
daughter airways. Table 5-1 summarizes new studies evaluating the morphological effects of a
single acute O3 exposure.
5-39
-------
Table 5-1. Summary of Studies that Evaluated Morphological Effects
of a Single Acute O3 Exposure.
0.25 20-90
min
0.5
1
1 90 min
0.5 8hr
0.5 8hr
0.5 8hr
1 8hr
1 8hr
1 8hr
0.4 2hr
1
0.8 8hr
1 8hr
1 8hr
Rat nr nr t 1 1
" nr nr It t
nr nr It! t
Rat nr nr It! It
Rat It! nr nr nr
Rat ttt nr nr nr
Rat It nr nr nr
Rat nr nr t nr
Rat It nr t t
Rat nr nr nr nr
Monkey nr -> t t
nr t t t
Monkey nr nr nr nr
Monkey nr nr nr nr
Ferret nr nr nr nr
nr t nr nr Postlethwait et al.
(2000)
nr t nr nr Postlethwait et al.
(2000)
nr t nr nr Postlethwait et al.
(2000)
t t nr nr Joad et al. (2000)
nr nr nr nr Hotchkiss et al.
(1997)
nr nr nr nr Cho et al.
(1999a,2000)
nr nr nr nr Wagner et al.
(2002)
nr ttt nr nr Schelegle et al.
(2001)
nr T nr nr Vesely et al.
(1999a)
nr T T nr Sterner-Kock
et al. (2000)
nr It nr nr Plopper et al.
(1998)
nr It nr nr Plopper et al.
(1998)
nr It nr nr Hyde etal. (1999)
nr It nr nr Sterner-Kock
et al. (2000)
nr It nr nr Sterner-Kock
et al. (2000)
Symbols: -> = no change; t = small increase in indices reflecting O3-induced morphological changes in specified upper
or lower respiratory region; 11 = relatively moderate morphological changes induced by O3 exposure; 111 = marked
morphological changes or remodeling due to O3 exposure; nr = changes not reported. The number of arrows in each study is
based on the authors' description of injury or from graphs in the paper comparing O3-exposed animals to air-exposed controls.
As endpoints for the studies varied among papers and information about injury was largely descriptive, the table does not
attempt a quantitative comparison, but only a subjective illustration of the current literature. Specific types of changes are
discussed in text of Section 5.2.4.
5-40
-------
5.2.4.3 Subchronic and Chronic Exposure Effects
Summaries of new studies of morphological effects of subchronic and chronic exposures
are listed in Table AX5-10 in Annex AX5. In general, as the duration of exposure lengthens,
there is no concomitant linear increase in the intensity of effect for any given endpoint. Rather,
as exposure proceeds past 1 week to 1 year, Type I cell necrosis and inflammatory responses
generally decrease to near control values, and hyperplastic and fibrotic changes remain elevated.
After long-term exposure ended, some indicies of fibrosis persisted and, in some cases, became
more severe during PE periods in clean air.
Effects of O3 on the upper respiratory tract of F344 rats exposed to O3 (0.12, 0.5, or
1.0 ppm for 20 months) included marked mucous cell metaplasia in the rats exposed to 0.5 and
1.0 ppm O3, but not at 0.12 ppm O3 (Harkema et al., 1997a). In a follow-up study, hyperplasia
was found in the nasal epithelium of rats exposed to 0.25 and 0.5 ppm, 8 h/day, 7 days/week, for
13 weeks (Harkema et al., 1999). The mucous cell metaplasia, and associated intraepithelial
mucosubstances, induced by 0.5 ppm O3 persisted for 13 weeks after exposure. An acute (8 h)
exposure to 0.5 ppm O3 13 weeks after the chronic exposure induced an additional increase of
mucosubstances in the nasal epithelium of rats but not in rats chronically exposed to 0 or
0.25 ppm O3. The persistent nature of the O3-induced mucous cell metaplasia in rats reported in
this study suggests that O3 exposure may have the potential to induce similar long-lasting
alterations in the airways of humans.
No significant changes in nasal tissue were seen in rats continuously exposed for 49 days
to the ambient air of Mexico City, Mexico (Moss et al., 2001). A rat study using 6-month
exposures to ambient air of Sao Paulo, with a different pollutant composition than that of
Mexico City, demonstrated development of secretory hyperplasia in rats (Lemos et al., 1994).
However, without information on specific differences in ambient pollution composition in the
two cities, the studies cannot be compared; nor can the observed changes be confidently
attributed to O3 versus the overall oxidant-contaminated ambient air mix. Because of the
persistent nature of these changes in the controlled studies with rats and the fact that the upper
airways of humans are probably more sensitive (like in the monkey) the authors suggested that
long-term exposure to ambient levels of O3 could induce significant nasal epithelial lesions that
may compromise the upper respiratory tract defense mechanisms of exposed human populations.
5-41
-------
Rats exposed to 0.5 ppm O3 for 1 month exhibited Bcl-2 in protein extracts of nasal
epithelium (Tesfaigzi et al., 1998). Further, after 3 and 6 months of exposure, the number of
metaplastic mucous cells in the transitional epithelium was indirectly related to the percentage of
cells that were Bcl-2 positive. Cells from rats exposed to FA did not express any Bcl-2. This
study suggests that apoptosis regulators like Bcl-2 may play a role in the development and
resolution of mucous cell metaplasia in the nasal airway.
A spectrum of lesions was reported (Herbert et al., 1996) in the nasal cavity and
centriacinar lung of male and female mice exposed to 0.5 or 1.0 ppm of O3 for 2 years; the
lesions persisted with continued exposure for 30 months. These lesions included bone loss in the
maxilloturbinates, mucosal inflammation, mucous cell metaplasia in the nasal transitional
epithelium and increased interstitial and epithelial thickening in the proximal alveolar region.
In the CAR, there were increased numbers of nonciliated cells. However, changes in other
endpoints including lung function and lung biochemistry were not evident. The investigators'
interpretation of the entire study is that rodents exposed to the two higher O3 concentrations had
some of the structural hallmarks of chronic airway disease in humans.
A chronic study using a simulated, seasonal O3-exposure pattern was reported by Plopper
and colleagues (Evans et al., 2003; Schelegle et al., 2003a; Chen et al., 2003; Plopper and
Fanucchi, 2000). Infant rhesus monkeys (30 days old) were exposed to FA, house dust mite
allergen aerosol (HDMA), or O3 + HDMA. The 0.5 ppm O3 exposures were 8 h/day for 5 days,
every 14 days for a total of 11 O3 episodes. Half of the monkeys were sensitized to HDMA
(Dermatophagoides farinae) at 14 and 28 days of age. The sensitized monkeys were exposed to
HDMA for 2 h/day on Days 3 to 5 of the FA or O3 exposures. The lungs were removed during
the last FA exposure and the right and left cranial and right middle lobes were separately
inflation fixed. Microdissection and morphometric analyses were performed on the conducting
airways to the level of the most proximal respiratory bronchiole. Repeated exposures to O3 or O3
+ HDMA over a 6-month period resulted in an atypical development of the basement membrane
zone (BMZ) of airways in nonsensitized developing monkeys. Remodeling in the distal
conducting airways was found in the sensitized monkeys as a result of the damage and repair
processes occurring with repeated exposure (Evans et al., 2003; Schelegle et al., 2003a). Lung
function changes in these monkeys (Schelegle et al., 2003b), and associated adaptation of the
respiratory motor responses (Chen et al., 2003), are described in Section 5.2.5.2. Collectively,
5-42
-------
these findings provide a plausible pathophysiologic basis for changes in airway function
described in children growing up in polluted metropolitan areas (e.g., Tager, 1999)
(see Chapter 7).
Necropsy of the left caudal lobe of these infant monkeys showed accumulation of mucous
cells and eosinophils within the combined epithelial and interstitial compartments in the
conducting airways and in the terminal/respiratory bronchioles (Schelegle et al., 2003a). House
dust mite sensitization and HDMA challenge alone, or combined with O3 exposure, resulted in
significantly greater eosinophil accumulation in the conducting airways when compared to FA-
and O3-only exposures. A significant accumulation of eosinophils was found in the
terminal/respiratory bronchioles of the sensitized monkeys challenged with HDMA when
compared to monkeys exposed to FA, O3, and HDMA + O3. The mean mass of mucous cells
increased in the fifth generation conducting airways of sensitized animals challenged with
HDMA alone and when combined with O3 exposure and in the terminal bronchioles of sensitized
animals exposed to HDMA + O3. The tracheal basement membrane of HDMA-sensitized
monkeys exposed to HDMA or to HDMA + O3 was significantly increased over controls;
however, there were no significant changes in the airway diameter of proximal and mid-level
airways. Exposures of sensitized young monkeys to HDMA alone, or to O3 alone, resulted in
eosinophilia of the mid-level conducting airways and the terminal/respiratory bronchioles, but
without alterations in airway structure or function. The authors interpreted these findings to
indicate that the combination of cyclic O3 exposure and HDMA challenge in HDMA-sensitized
infant monkeys acts synergistically to produce an allergic-reactive airway phenotype
characterized by significant eosinophilia of mid-level conducting airways, transmigration of
eosinophils into the lumen, and an altered structural development of conducting airways that is
associated with increased Raw and nonspecific airway reactivity (see Section 5.2.5).
Examination of development of the tracheal BMZ in these monkeys (Evans et al., 2003)
showed that with exposures to either O3 or HDMA + O3, BMZ development was affected.
Abnormalities in the BMZ included: (1) irregular and thin collagen throughout the BMZ;
(2) perlecan depleted or severely reduced; (3) fibroblast growth factor receptor-1 (FGFR-1)
immunoreactivity reduced; (4) fibroblast growth factor-2 (FGF-2) immunoreactivity absent in
perlecan-deficient BMZ, but present in the lateral intercellular space (LIS), in basal cells, and in
attenuated fibroblasts; (5) syndecan-4 immunoreactivity increased in basal cells. The authors
5-43
-------
interpret these data as suggesting that O3 target cells are associated with synthesis of epithelial
BMZ perlecan. The absence of FGF-2, normally stored in the BMZ, could affect downstream
signaling in airway epithelium and could be responsible for the abnormal development of the
airway seen in this study and thus be an important mechanism modulating O3-induced injury.
Mid-level bronchi and bronchioles from these monkeys (Larson et al., 2004) demonstrated
decrements in the density of epithelial nerves in the axial path between the sixth and seventh
airway generations in exposures to O3. Combined O3 + HDMA exposures exacerbated this
reduction. They attribute this loss of nerve plexuses to neural regression or stunted nerve
development, the latter corroborated by the Evans et al. (2003) finding of decreased growth
factors following O3 exposure. Additionally, they found streaks or clusters of cells that were
immunoreactive for protein gene product 9.5 (PGP 9.5, a pan-neuronal marker) and negative for
calcitonin gene-related peptide. The functional significance of this is unknown but suggests to
the authors a possible injury-repair process induced by O3.
Remodeling of the distal airways and CAR is one of the most disturbing aspects of the
morphological changes occurring after subchronic and chronic exposure to O3. Recently,
bronchiolization was reported in rats exposed to 0.4 ppm O3 for only 56 days (van Bree et al.,
2001). They also found collagen formation progressively increased with increasing O3 exposure
and persisted into PE recovery. In addition to centriacinar remodeling, Pinkerton et al. (1998)
reported thickening of tracheal, bronchial, and bronchiolar epithelium after 3 or 20 months
exposure to 1 ppm O3, but not to 0.12 ppm. Although some older literature had reported that
chronic exposures to < 1.0 ppm O3 cause emphysema, none of the more recent literature supports
this hypothesis.
5.2.4.4 Summary and Conclusions—Subchronic and Chronic Morphological Effects
The progression of effects during and after a chronic O3 exposure across a range of 0.5 to
1.0 ppm is complex, with inflammation peaking over the first few days of exposure, then
dropping, next plateauing, and then finally largely disappearing. Epithelial hyperplasia follows a
somewhat similar pattern. Effects of 0.5 ppm O3 for 20 months on the nasal mucosa include
atrophy of nasal turbinates, epithelial hyperplasia, and mucous cell metaplasia, which persisted
long after the exposure ceased. Fibrotic changes in lung tissue increase very slowly over months
of exposure; and, after exposure ceases, the changes sometimes persist or increase. The pattern
5-44
-------
of exposure in this same concentration range determines effects, with 18 months of daily
exposure causing less morphologic damage than exposures on alternating months. This is
important, given that environmental O3 exposure is typically seasonal. Plopper and colleagues'
long term study of infant rhesus monkeys exposed to simulated, seasonal O3 (0.5 ppm 8 h/day
for 5 days, every 14 days for 11 episodes) demonstrated: (1) remodeling in the distal airways;
(2) abnormalities in tracheal basement membrane; (3) eosinophil accumulation in conducting
airways; and (4) decrements in airway innervation. These findings advance earlier information
regarding possible injury-repair processes occurring with seasonal O3 exposures.
5.2.5 Effects on Pulmonary Function
5.2.5.1 Acute and Subchronic Exposure Effects on Pulmonary Function
Numerous pulmonary function studies of the effects of acute O3 exposure (defined here
as < 1 week of exposure) in several animal species have been conducted and generally show
responses similar to those of humans (e.g., increased breathing frequency, decreased tidal
volume, increased resistance, decreased forced vital capacity [FVC] and changes in the
expiratory flow-volume curve). These effects are seen at 0.25 to 0.4 ppm O3 for several hours in
a number of species. At concentrations of > 1 ppm, breathing mechanics (compliance and
resistance) are affected. The breathing pattern returns to normal after O3 exposure. In rats
exposed to 0.35 to 1 ppm O3 for 2 h/day for 5 days, there was a pattern of attenuation of
pulmonary function responses similar to that observed in humans. Concurrently, there was no
attenuation of biochemical indicators of lung injury or of morphological changes.
Work demonstrating attenuation of pulmonary function effects (see Table AX5-11) was
completed by Wiester et al. (1996), who exposed male Fischer 344 rats to 0.5 ppm O3 for either
6 or 23 h/day over 5 days. Ozone-induced changes in lung volume were attenuated during the
5 exposure days and returned to control levels after 7 days recovery. The responses to repeated
O3 exposure in rats were exacerbated by reduced ambient temperature, presumably as a result of
increased metabolic activity.
In an attempt to model susceptible populations, researchers have utilized inbred mouse
strains with varying ventilatory responses to O3. As differences were seen in inflammatory
responses to acute O3 exposures in C57BL/6J and C3H/HeJ mice, comparisons were also made
of their ventilatory responses (Tankersley et al., 1993). Following an exposure of 2 ppm O3 for
5-45
-------
3 h, breathing frequency (fB), tidal volume (VT), and minute ventilation (VE) were measured 1
and 24 h in both normocapnia (or air at -0% CO2) and hypercapnia (5 or 8% CO2). They
demonstrated that acute O3 exposures caused altered hypercapnic ventilatory control, which
varied between strains. This suggested to the authors that O3-induced alterations in ventilation
are determined, at least in part, by genetic factors. A caveat regarding studies such as this using
high exposure concentrations is that events observed at high concentrations may differ from
those observed at near-ambient O3 levels.
Paquette et al. (1994) measured ventilatory responses in C57BL/6J and C3H/HeJ mice
given repeated acute exposures of 0.3 ppm for 48 and 72 h. The two strains had differing
responses to both normocapnia and hypercapnia. Normocapnic VE was greater following
subacute O3 exposure in C57BL/6J mice than in C3H/HeJ mice, due to increased fB and
reduced VT, respectively. This suggests that the increased VT in C57BL/6J mice may contribute
to the increased susceptibility to lung injury due to a greater dose of O3 reaching the lower lung.
Hypercapnic ventilatory responses following subacute O3 exposures demonstrated reduced VE
(due to decreased VT) in the C57BL/6J mice only. Evaluations of O3 dosimetry were performed
in these two strains using 18O3-labeled ozone (2 ppm for 2-3 h) (Slade et al., 1997). Immediately
after exposures of 2 ppm 18O3 for 2 to 3 h, C3H/HeJ mice had 46% less 18O in lungs and 61%
less in the trachea than C57BL/6J. Additionally, C3H/HeJ mice had a greater body temperature
decrease following O3 exposure than C57BL/6J mice, suggesting that the differences in
susceptibility to O3 are due to differences in the ability to decrease body temperature and,
consequently, decrease the dose of O3 to the lung.
Tracheal transepithelial potential has also been shown to differ in eight mouse strains 6 h
after exposure to 2 ppm O3 for 3 h (Takahashi et al., 1995b). AKR/J, C3H/HeJ, and CBA/J were
identified as resistant strains; and 129/J, A/J, C57BL/6J, C3HeB/FeJ, and SJL/J were identified
as susceptible strains. The authors noted that strains' responses in this parameter did not show
concordance with inflammatory responses, suggesting to the authors that the two phenotypes are
not controlled by the same genetic factors.
Savov et al. (2004) characterized ventilatory responses in nine mouse strains exposed to O3
(2.0 ppm O3 for 3 h). The C57BL/6J strain was hyporeactive to methacholine (MCh) prior to O3,
but was very responsive to MCh following O3. Conversely, C3H/HeJ mice had an intermediate
5-46
-------
baseline Penh and a small response to MCh following O3 exposure. This study corroborates the
evidence of no consistent relationship between respiratory Penh and inflammation.
5.2.5.2 Summary and Conclusions—Acute and Subchronic Effects on
Pulmonary Function
Early work has demonstrated that during acute exposure of-0.2 ppm O3 in rats, the most
commonly observed alterations are increased frequency of breathing and decreased tidal volume
(i.e., rapid, shallow breathing). Exposures of-1.0 ppm O3 affect breathing mechanics
(compliance and resistance). Additionally, decreased lung volumes are observed in rats with
acute exposures at levels of 0.5 ppm. New work utilizing inbred mouse strains with varying
ventilatory responses to O3 has suggested that: (1) control of the ventilatory response is
determined, at least in part, by genetic factors; (2) increased VT in some strains may contribute
to lung injury due to a greater dose of O3 reaching the lower lung; (3) the ability to reduce body
temperature in some strains may account for their decreased O3-induced lung injury; and
(4) tracheal transepithelial potential is determined, in part, by genetic factors. Importantly, the
genetic loci that appear to be modulating various aspects of pulmonary responses to O3 differ
from each other and from loci controlling inflammatory responses.
Exposures of 2 h/day for 5 days create a pattern of attenuation of pulmonary function
decrements in both rats and humans without concurrent attenuation of lung injury and
morphological changes, indicating that the attenuation did not result in protection against all the
effects of O3. Chronic O3 exposure studies evaluating pulmonary function are not available.
Earlier work has demonstrated that repeated daily exposure of rats to an episodic profile of O3
caused small, but significant, decrements in lung function that were consistent with early
indicators of focal fibrogenesis in the proximal alveolar region without overt fibrosis.
5.2.5.3 Ozone Effects on Airway Responsiveness
Effects of O3 on airway reactivity have been observed in a variety of species across an
exposure range of 0.5 to 1 ppm. Many of the new studies on pulmonary function in laboratory
animals allow a better prediction of the effects of O3 exposure on the exacerbation of asthma
symptoms and the risk of developing asthma in humans. However, it is necessary to understand
the factors that determine airway responsiveness across different mammalian species, as is
discussed in Chapter 4.
5-47
-------
Traditional studies of airway responsiveness require sedation in both infants and laboratory
animals. Laboratory animal studies employ intravenous agonist challenges as well as inhalation
challenges, though inhaled agonist challenges are preferred in humans. Exercise testing is
not possible with sedation unless exercise is "simulated" by increasing ventilation using
elevated CO2, and the need for artificial ventilation in laboratory animal studies may cause
breathing patterns that affect O3 deposition. load et al. (2000) reported that when 1 ppm O3 for
90 min is administered to isolated rat lung at either 2.4 mL/40 breaths/minute (bpm) or
1.2 mL/80 bpm, the more rapid breathing pattern elicits less epithelial cell injury than the slower
breathing pattern. Though this study design does not really model the rapid, shallow breathing
elicited in the intact animal, it shows greater reduction in injury in the proximal axial airway
compared to its adjacent airway branch and terminal bronchiole. The rapid, shallow breathing
pattern protects the large conducting airways of rats but causes a more even distribution of
epithelial cell injury to the terminal bronchioles (Schelegle et al., 2001). Postlethwait et al.
(2000) demonstrated that the conducting airways are the primary site of acute cytotoxicity from
O3 exposure. Three-dimensional mapping of the airway tree in SD rat isolated lung exposed to
0, 0.25, 0.5, or 1.0 ppm O3 showed a concentration-dependent increase in injured cells. Injury
was evident in proximal and distal conduction airways, lowest in terminal bronchioles, and
highest in the small side branches downstream of bifurcations. These exposure levels did not
concurrently elicit changes in LDH activity or total protein in BALF, suggesting that the
mapping technique is a more sensitive measure of injury and is useful in dosimetry studies.
Whole-body plethysmography of unanesthetized, unrestrained rodents has been used to
indirectly measure pulmonary resistance (RL) (Shore et al., 2002; Goldsmith et al., 2002; Jang
et al., 2002). However, these indices of inspiratory/expiratory pressure differences, including
Penh, may be less sensitive than direct measurements of lung airflow resistance (Murphy, 2002).
Changes in airway structure caused by viral infections also must be considered when evaluating
laboratory animal studies. Animals with acute viral illness have morphological evidence of
inflammatory cell infiltration, bronchiolar wall edema, epithelial hyperplasia, and increased
airway mucous plugs that can cause airway narrowing, air trapping, and serious functional
changes in the lung (Folkerts et al., 1998).
Exercise-induced bronchoconstriction in humans appears to be mediated by changes in the
tonicity of the airway lining fluid (Anderson and Daviskas, 2000). Brannan et al. (1998) suggest
5-48
-------
that a test in laboratory animals based on the inhalation of mannitol aerosol (hyperosmolar) may
be feasible and provide information similar to that from exercise challenges in cooperative
children and adults. Unfortunately, there have been few reports of mannitol or adenosine
monophosphate challenges in laboratory animals; most studies have utilized histamine, MCh,
acetylcholine (ACh), or carbachol to determine outcome. In active humans with asthma,
adenosine monophosphate challenges appear to better reflect ongoing airway inflammation than
histamine or MCh challenges (Polosa and Holgate, 1997; Avital et al, 1995a,b) and may be
useful in identifying mechanisms of asthma in laboratory animals and their responsiveness to
environmental pollutants.
The increased responsiveness to bronchoconstrictor challenge in asthma is thought to result
from a combination of structural and physiological factors that include increased inner-wall
thickness, smooth-muscle responsiveness, and mucus secretion. These factors also are likely to
determine a level of innate airway responsiveness that is genetically influenced. Chapter 6
(Section 6.8) discusses cellular and biochemical changes that have been identified in human
asthmatics. These studies suggest that the mechanisms involved in AHR are multifactoral, with
general agreement that there is an inconsistent relationship between AHR and markers of
inflammation.
A large database of laboratory animal research has been collected on the role of O3 in
producing an increase in AHR (see Table AX5-12). Exposure levels (> 1 ppm for >30 min)
in many of these studies are not environmentally relevant, but information may be obtained
regarding the mechanisms of action of O3 concerning: O3 concentration and peak response time,
inhaled versus intravenous challenge with nonspecific bronchoconstrictors, neurogenic
mediation, neutrophilic inflammation, and interactions with specific biological agents (e.g.,
antigens and viruses). However, as with other toxicants, high-dose and low-dose mechanisms
may differ, so interpretation of results must take this into consideration.
Many species of laboratory animals have been used to study the effects of O3 on airway
bronchoconstriction. Ozone-induced AHR in guinea pigs has been used to model human
bronchospasm (van Hoof et al., 1996; 1997a,b; Matsubara et al., 1997a,b; Sun and Chung, 1997;
Aizawa et al., 1999a,b; Tsai et al., 1998; Nakano et al., 2000). Because these studies were done
at 2 to 3 ppm O3, these results are not directly relevant for extrapolation to potential airway
responses in humans exposed to ambient levels of O3. Humans with reactive airway disease
5-49
-------
(e.g., asthma) appear to be sensitive to ambient levels of O3 (see Chapters 6 and 7) and the
current understanding is that O3 exacerbates airway responsiveness to specific allergens,
presumably by nonspecifically increasing AHR.
Shore et al. (2000, 2002) have shown that O3-induced AHR is reduced in immature rats and
mice. SD rats exposed to 2 ppm O3 at ages 2, 4, 6, 8, or 12 weeks and A/J mice exposed to 0.3 to
3 ppm for 3 h at age 2, 4, 8, or 12 weeks had similar concentration-related decreases in VE
except at the youngest ages. This smaller decrement in VE suggested a delivered dose that was
much greater in the younger animals. This group (Shore et al., 2003) has also recently shown
that obese mice have greater ventilatory responses to O3. Exposures of 2.0 ppm O3 for 3 h to
lean, WT C57BL/6J and ob/ob mice (mice with a genetic defect in the coding for leptin, the
satiety hormone) showed that the ob/ob mice had enhanced AHR and inflammation compared to
the WT mice. These data correlate with epidemiological data showing increased incidence of
asthma in overweight children.
Increased AHR to various nonspecific bronchoconstrictive agents (e.g., ACh, MCh,
histamine, carbachol) given by inhalation or intravenous routes has been previously shown in
laboratory animals exposed to O3 concentrations < 1.0 ppm. Dye et al. (1999) showed
hyperresponsiveness to MCh in rats 2 h after exposure to 2 ppm O3 for 2 h. AHR can be induced
by specific antigens as well as O3. The most commonly used laboratory animal model is the
OVA-sensitized guinea pig. Animals sensitized with OVA have been shown to have similar
responses to nonspecific bronchoconstrictors as control animals.
OVA-sensitized guinea pigs (Sun et al., 1997) and mice (Yamauchi et al., 2002) were used
to determine the enhancement of antigen-induced bronchoconstriction by acute, high-level O3
(1.0 ppm O3 for 1 h). Male Dunkin-Hartley guinea pigs were sensitized by i.p. injection of OVA
and exposed to O3 alone, OVA aerosol, or O3 + OVA. Ozone exposure alone increased
bronchial responsiveness to ACh at 3 h, but not 24 h, whereas OVA alone had no effect.
Combined exposure to O3 and OVA (1 ppm for 1 h, then 3 min OVA) increased bronchial
responsiveness to ACh 3 h after O3 exposure. At 24 h following O3 exposure, AHR increased
when OVA challenge was performed at 21 h, suggesting that O3 preexposure can potentiate
OVA-induced AHR. Neutrophil counts in the BALF increased at 3 and 24 h after O3 exposure
alone but were not further increased when O3 exposure was combined with OVA airway
challenge; however, protein content of the BALF increased at 3 and 24 h in the O3 and OVA
5-50
-------
groups. Thus, this study also indicates that high-ambient O3 exposure can augment antigen
(OVA)-induced AHR in guinea pigs.
Yamauchi et al. (2002) sensitized male C57BL/6 mice by i.p. injection of OVA and then
exposed them to O3. The sensitized mice had AHR to MCh. Ozone exposure caused significant
decreases in dynamic lung compliance (Cdyn), VE, and PaO2 in OVA-sensitized mice but not in
controls. A marker of inflammation (soluble intercellular adhesion molecule-1 [sICAM-1]) was
elevated in the BAL fluid of OVA-sensitized mice, but sICAM-1 levels were not significantly
changed by O3 exposure, indicating that the O3-induced AHR to MCh was not caused by O3-
induced inflammation.
Ozone-induced AHR may be temporally associated with inflammatory cells stimulated by
cytokines (Koto et al., 1997), mast cells (Igarashi et al., 1998; Noviski et al., 1999), or by oxygen
radicals (Takahashi et al., 1993). One study, however, has shown that inflammation is not a
prerequisite of AHR (Koto et al., 1997), and it has been suggested that O3-induced AHR may be
epithelium-dependent (McGraw et al., 2000). For example, neonatal rats pretreated with
capsaicin, which will permanently destroy C-fibers and prevent O3-induced (1 ppm, 8 h) release
of neuropeptides (Vesely et al., 1999a), and then exposed to O3 when adults, showed a marked
increase in airway responsiveness to inhaled aerosolized MCh (Jimba et al., 1995). Takebayashi
et al. (1998) has shown that depletion of tachykinins by capsaicin treatment, or by a specific
tachykinin receptor antagonist, can block the induction of AHR by O3. The seemingly disparate
responses in laboratory animals may be due to species- or strain-specific differences in inherent
reactivity to bronchoconstrictors or to inherent differences in susceptibility to O3-induced
inflammation (Zhang et al., 1995; Depuydt et al., 1999; Dye et al., 1999).
Studies that may be potentially relevant to ambient levels of O3 were conducted in vivo, in
an isolated perfused lung model, and in ex vivo lung segments using multihour and repeated
multihour exposures with ambient levels of O3. A study on the relationship between O3-induced
AHR and tracheal epithelial function was conducted in New Zealand White rabbits by Freed
et al. (1996). Rabbits exposed to O3 (0.2 ppm for 7 h) demonstrated significantly decreased
tracheal transepithelial potential difference but no changes in lung resistance. Changes in the
compartmentalized lung resistance, measured in response to ACh challenge before and after
bilateral vagotomy, were not significantly different in air-exposed rabbits; however, bilateral
vagotomy enhanced peripheral lung reactivity in O3-exposed rabbits. The ACh-induced 140%
5-51
-------
increase in lung resistance with O3 exposure was two times higher than with air exposure,
indicating that ambient-level O3 exposure affects tracheal epithelial function in rabbits and
increases central airway reactivity, possibly through vagally-mediated mechanisms.
Pulmonary mechanics and hemodynamics were studied in the New Zealand White rabbit
isolated perfused lung model that allowed partitioning of the total pressure gradient into arterial,
pre- and postcapillary, and venous components (Delaunois et al., 1998). Exposures to O3
(0.4 ppm for 4 h) were followed by evaluation of airway responsiveness to ACh, substance P
(SP), or histamine immediately or 48 h later. Ozone inhibited pulmonary mechanical reactivity
to all three bronchoconstrictors that persisted for 48 h and modified vasoreactivity of the
vascular bed, but only at 48 h PE. Arterial segmental pressure, normally insensitive to ACh and
SP, was significantly elevated by O3; precapillary segmental pressure decreased in response to
Ach, suggesting that O3 can induce direct vascular constriction, but the vascular responses are
variable and depend on the agonist used and on the species studied.
Airway responsiveness to the same three compounds was evaluated by Segura et al. (1997)
in guinea pigs exposed to O3 (0.15, 0.3, 0.6, or 1.2 ppm for 4 h). Ozone did not cause AHR to
ACh or histamine, except at the highest concentration (1.2 ppm O3) for histamine. However, O3
caused AHR to SP at >0.3 ppm, suggesting that O3 destroys neutral endopeptidases (responsible
for SP inactivation) in airway epithelial cells. Vargas et al. (1998), in a follow-up study,
demonstrated that guinea pigs chronically exposed to 0.3 ppm O3 for 4 h/day became adapted to
SP-induced AHR. Ozone caused increased sensitivity to SP after 1, 3, 6, 12, and 24 days of
exposure that was associated with airway inflammation; however, after 48 days of exposure, the
increased sensitivity to SP was lost.
This study is in accordance with Szarek et al. (1995) who demonstrated that AHR
associated with acute O3 exposures does not persist during long-term exposure to near-ambient-
levels of O3 (<1 ppm). Fischer 344 rats, exposed to 0.0, 0.12, 0.5, or 1.0 ppm O3, 6 h/day,
5 days/week for 20 months, demonstrated significantly reduced responses to bethanechol, ACh,
and electrical field stimulation in eighth generation airway segments. This suggests that some
adaptation had taken place during long-term exposure, possibly due to increased inner wall
thickness.
It is well known that the changes in breathing pattern and lung function caused by O3 are
attenuated with repeated daily exposures for at least 3 to 5 days. But guinea pigs exposed to
5-52
-------
0.5 ppm O3, 8 h/day for 7 days showed enhancement of responsiveness of rapidly adapting
airway receptors (load et al., 1998). Repeated exposure increased receptor activity to SP, MCh,
and hyperinflation; there were no significant effects on baseline or SP- and MCh-induced
changes in lung compliance and resistance, suggesting that the responsiveness of rapidly
adapting receptors was enhanced.
Male and female Hartley guinea pigs exposed to O3 (0.1 and 0.3 ppm, 4 h/day, 4 days/week
for 24 weeks) were evaluated for airway responsiveness following ACh or OVA inhalation
challenges (Schlesinger et al., 2002a,b). Ozone exposure did not cause AHR in nonsensitized
animals but did exacerbate AHR to both ACh and OVA in sensitized animals that persisted for
4 weeks after exposure. The effects of O3 on airway responsiveness were gender independent
and were concentration-related for the ACh challenges.
Schelegle et al. (2003a) evaluated airway responsiveness in infant rhesus monkeys exposed
to a 5-day O3 episode repeated every 14 days over a 6-month period. Half of the monkeys were
sensitized to HDMA at 14 and 28 days of age before exposure to a total of 11 episodes of O3
(0.5 ppm, 8 h/day for 5 days followed by 9 days of FA), HDMA, or O3 + HDMA. Baseline Raw
was significantly elevated after 10 exposure episodes in the HDMA + O3 group compared to the
FA, HDMA, and O3 exposure groups. Aerosol challenge with HDMA at the end of the 10th
episode did not significantly affect Raw, VT, fB, or SaO2. Aerosol challenge with histamine was
not significantly different after 6 episodes; however, the EC 150 Raw for the HDMA + O3 group
was significantly reduced after 10 episodes when compared to the FA, HDMA, and O3 exposure
groups, indicating the development of AHR in this group sometime between episodes 6 and 10.
The results are consistent with altered structural development of the conducting airways.
During repeated episodic exposures to O3, respiratory responses are first altered to a rapid,
shallow breathing pattern, which has long been considered protective, especially to the deep
lung. This dogma has been discounted recently as discussed above (Schelegle et al., 2001).
Alfaro et al. (2004) examined the site-specific deposition of 18O (1 ppm 2 h) at breathing
frequencies of 80, 120, 160, or 200 bpm. At all frequencies, parenchymal areas had a lower
content of 18O than trachea and bronchi. As breathing frequency increased from 80 to 160 bpm,
the deposition showed a reduction in mid-level trachea and an increase in both mainstream
bronchi. At this frequency there was also an increase in deposition in parenchyma supplied by
short (cranial) airway paths, consistent with the results of Schelegle et al., (2001).
5-53
-------
At 200 bpm, 18O deposition in trachea increased, concurrent with increases in right cranial and
caudal bronchi regions. Right cranial parenchymal content decreased at 200 bpm, whereas right
caudal parenchymal levels did not change at any breathing frequency. The authors list some
limitations of this study, such as the possible effect on regional distribution of ventilation by use
of the negative-pressure ventilator, the effect of paralysis on airway geometry, and possible
translocation of 18O during the 2 h exposure period. These two studies provide evidence
that O3-induced rapid, shallow breathing creates a more evenly distributed injury pattern,
with possibly greater protection from focal injury to the large conducting airways, including the
trachea and the left mainstem bronchus.
Another study of the adaptive phenomena in SD rats used an exposure protocol consisting
of 5 days of daily 8 h, 1 ppm O3 exposures followed by 9 days of recovery in FA (Schelegle
et al., 2003b). This O3/FA pattern was repeated for four cycles and demonstrated that the O3-
induced rapid shallow breathing pattern was followed by adaptation that occurred with each
cycle. However, the release of SP from the trachea, the neutrophil content, and cell proliferation
became attenuated after the first cycle, suggesting a disconnect from the rapid, shallow breathing
response. Hypercellularity of the CAR epithelium and thickening of the CAR interstitium, not
linked to changes in cell proliferation, were also found. The authors suggest mechanism(s) of
injury from repeated O3 exposures, consisting of diminished neutrophilic inflammation/and or
release of mitogenic neuropeptides, depressed cell proliferative response, and cumulative distal
airway lesion.
Following the initial response of a rapid, shallow breathing pattern, animals eventually
adapt with continued episodic exposure despite the continued presence of epithelial damage,
altered structural development, and inflammation of the airways. Chen et al. (2003) used a
subset of the monkeys from the Schlegele et al. (2003a) study to demonstrate that the attenuation
of O3-induced rapid, shallow breathing and lung function changes typically seen with
repeated O3 exposure may be caused by the adaptation of the respiratory motor responses. This
episodic O3 exposure appeared to create neuroplasticity of the nucleus tractus solitarius (NTS, a
region of the brainstem which controls respiration), including increased nonspecific excitability
of the NTS neurons, an increased input resistance, and an increased spiking response to
intracellular injections of depolarizing current.
5-54
-------
5.2.5.4 Summary and Conclusions—Effects on Airway Responsiveness
Ozone-induced AHR has been reported in a number of laboratory species at an exposure
range of 0.5 to 1.0 ppm and in human asthmatics at ambient levels. In asthmatics, O3 is thought
to exacerbate AHR to specific allergens by nonspecifically increasing AHR. New studies have
demonstrated that AHR in asthmatics is due in part to chronic inflammation and airway
remodeling. Animal studies have shown that O3 exposure can augment OVA-induced AHR.
Importantly, there is a temporal relationship between inflammatory cell influx and O3-induced
AHR, but inflammation is not a prerequisite of AHR. Repeated O3 exposures enhance AHR,
possibly by modulating rapidly adapting airway receptors or by altering the structure of
conducting airways.
Currently reported investigations on AHR with repeated O3 exposure to nonsensitized
laboratory animals have shown equivocal results, especially at the most relevant ambient O3
concentrations of <0.3 ppm. The few available studies in sensitized laboratory animals are
consistent with the O3-induced exacerbation of AHR reported in atopic humans with asthma (see
Chapter 6), but the results are difficult to extrapolate because of interindividual and interspecies
differences in responsiveness to bronchoprovocation and possible adaptation of airway
responsiveness with long-term, repeated O3 exposures. Therefore, further studies in laboratory
animals are needed to investigate responses to the different challenges in relation to
measurements of airway inflammation and the other physiological and structural factors known
to contribute to airway responsiveness in human subjects.
Important new information indicates that rapid, shallow breathing in response to O3 causes
a more evenly distributed injury pattern rather than protection from injury. New insights into the
mechanisms of O3-induced AHR suggest that: (1) exercise-induced bronchoconstriction may be
mediated by changes in tonicity of the bronchial smooth muscles; (2) vagally mediated
mechanisms may affect tracheal epithelial function and increase central airway reactivity;
(3) O3 may induce direct vascular constriction; (4) O3 may destroy neural endopeptidases in
airway epithelial cells, thus preventing the inactivation of SP; and (5) repeated O3 exposures may
diminish neutrophilic inflammation, depress cell proliferation, and cause cumulative distal
airway lesions.
5-55
-------
5.2.6 Genotoxicity Potential of Ozone
There has been an historical interest in the ability of ground-level pollution to cause cancer,
especially lung cancer. This interest has been amplified in recent years by results of an
epidemiologic study that suggest association of increased risks of incident lung cancer with
elevated long-term ambient concentrations of O3, coarse particulate matter (PM10), and sulfur
dioxide (SO2) in nonsmoking California males (Beeson et al., 1998; Abbey et al., 1999).
However, another larger, nationwide American Cancer Society study (Pope et al., 2002) showed
no significant effect of O3 on mortality risk but showed positive associations between
warm-season (July-September) O3 concentrations and cardiopulmonary mortality. Studies of
children and young adults of southwest metropolitan Mexico City, repeatedly exposed to high
levels of O3, particulate matter (PM), nitrogen oxides (NOX), aldehydes, metals, and other
components in a complex ambient mixture, also report DNA damage in blood leukocytes and
nasal epithelial cells (Valverde et al., 1997; Calderon-Garciduefias et al., 1999) and abnormal
nasal biopsies (Calderon-Garciduefias et al., 2001).
A number of experimental studies have been done to explore the mutagenic/carcinogenic
potential of O3. In vitro studies are difficult to interpret due to very high exposure levels and
culture systems that allowed the potential formation of artifacts. Some recently published in
vivo exposure studies (see Table AX5-13) found increased DNA strand breaks in respiratory
cells from guinea pigs (Ferng et al., 1997) and mice (Bornholdt et al., 2002) but, again, only with
exposure to high doses of O3 (1 ppm for 72 h and 1 or 2 ppm for 90 min, respectively).
Exposing the A/J mouse strain (known to have a high incidence of spontaneous pulmonary
adenomas) to 0.12, 0.50, and 1.0 ppm O3 for 6 h/day, 5 days/week for up to 9 months, Witschi
et al. (1999) did not find O3 exposure-related differences in lung tumor multiplicity or incidence.
Similarly, in a subchronic exposure study (B6C3FJ mice to 0.5 ppm O3 for 6 h/day, 5 days/week
for 12 weeks), Kim et al. (2001) did not find statistically significant increases in the incidence of
lung tumors. Significant differences in mean body weight as well as mean absolute and relative
weights of several organs (e.g., liver, spleen, kidney, testes, and ovary) were observed between
O3-exposed and air-exposed mice. Histopathologic examination of major organs revealed
oviductal carcinomas in 3/10 O3-exposed female mice.
5-56
-------
5.2.6.1 Summary and Conclusions—Genotoxicity Potential of Ozone
The weight of evidence from new experimental studies does not appear to support
ambient O3 as a pulmonary carcinogen in laboratory animal models. These new data are in
agreement with a study of carcinogenicity of O3 from the National Toxicology Program (NTP)
study (National Toxicology Program, 1994; Boorman et al., 1994), which was negative in male
and female rats, ambiguous in male mice, and positive only in female mice at high
concentrations of O3 (i.e., 1.0 ppm). As none of the new experimental studies of genotoxicity
provided lifetime exposure durations such as those used in NTP cancer studies, the observation
of no effects must be tempered by consideration of the limited duration of the exposure. Overall,
then, the new animal studies are inconclusive, as are the epidemiologic studies discussed in
Chapter 7, which may be due to significant species differences in this health endpoint. Also, O3
could act as a co-carcinogen functioning to stimulate hyperplasia. In epidemiology studies,
exposures typically consist of mixtures of copollutants, some of which are known carcinogens
(see Section 5.4.3).
5.3 SYSTEMIC EFFECTS OF OZONE EXPOSURE
Ozone indirectly affects organs beyond the respiratory system due to O3 reaction products
entering the bloodstream and being transported to target sites. Extrapulmonary effects could
also be due to the exposure-related production of mediators, metabolic products, and cell
trafficking. Although systemic effects are of interest and indicate a very broad array
of O3 effects, they are of limited influence and difficult to interpret. By protecting from
respiratory tract effects, these systemic effects will likely be protected against also. Systemic
effects are only summarized briefly here and in Table AX5-14.
5.3.1 Neurobehavioral Effects
Animal behavior, both motor activity and operant behavior, has been shown to be
suppressed by acute O3 exposures (3 to 6 h) of 0.12 ppm. There is a dose dependent decrease in
activity with increasing exposure levels. Additionally, these lowered activity levels tend to
attenuate with longer exposure periods. New studies in adult laboratory animals confirm that
environmentally relevant O3 concentrations from 0.2 to 1.0 ppm can decrease motor activity and
5-57
-------
affect short- and long-term memory, as tested by passive avoidance conditioning in 4 h
exposures in rats (Rivas-Arancibia et al., 1998; Avila-Costa et al., 1999; Dorado-Martinez et al.,
2001) or water-maze learning tasks in mice following a 30-day exposure (Sorace et al., 2001).
The effects have been attributed to reactive oxygen/nitrogen species and/or ozonation products.
The memory deficits could be blocked by administration of a-tocopherol (Guerrero et al, 1999)
or taurine (Rivas-Arancibia et al., 2000). Increased freezing and decreased exploratory
behaviors were accompanied by decreased serotonin levels and increased levels of NO,
glutamate, dopamine, and striatal lipoperoxidation in rats exposed to 1 ppm of O3 for 4 h (Rivas-
Arancibia et al., 2003). The O3-exposed animals also demonstrated neuronal cytoplasm and
dendrite vacuolation and dilation of rough endoplasmic reticulum cisterns, which the authors
interpret as a neurodegenerative process resulting from the oxidative stress of acute O3 exposure.
Nino-Cabrera et al. (2002) demonstrated that a 0.7 ppm O3 exposure for 4 h can induce
ultrastructural alterations in the hippocampus and prefrontal cortex in aged rats. These are areas
of the brain where degenerative age-related changes in learning and memory functions have been
reported (Bimonte et al., 2003).
Paz (1997) reviewed a series of studies that demonstrated significant alterations of
electroencephalographic (EEG) patterns during sleep in animals acutely exposed to O3 (0.35 to
1.0 ppm). Rats and cats both showed loss of paradoxical sleep time after 2 to 8 h of O3 exposure
(Paz and Bazan-Perkins, 1992; Paz and Huitron-Resendiz, 1996). Increased total wakefulness,
alterations in circadian rhythm, and a permanent 50% loss of paradoxical sleep time were shown
in rat pups born to dams exposed to 1.0 ppm O3 during gestation (Haro and Paz, 1993). Effects
on sleep patterns were associated with alterations in brain neurotransmitter levels (Huitron-
Resendiz et al., 1994; Gonzalez-Pifia and Paz, 1997) thought to be caused by O3 reaction
products or prostaglandins (Koyama and Hayaishi, 1994). The permanent effects in pups caused
by high O3 exposure during gestation were attributed to the diminished antioxidant capability of
fetal tissue (Giinther et al., 1993).
High, nonambient levels of O3 (e.g., >1.0 ppm) affect visual and olfactory neural pathways
in the rat. For example, Custodio-Ramierez and Paz (1997) reported a significant delay in visual
evoked potentials recorded in the visual cortex and the lateral geniculate nucleus of male Wistar
rats acutely exposed to high levels of O3 (1.5 and 3.0 ppm for 4 h). Colin-Barenque et al. (1999),
using the same strain, reported cytological and ultrastructural changes in the granular layer of the
5-58
-------
olfactory bulb after a 4 h exposure to 1 to 1.5 ppm O3. Although these neural effects are thought
to be caused by O3 reaction products, especially free radicals, the studies do not add much to an
understanding of the underlying mechanisms.
5.3.2 Neuroendocrine Effects
Early studies suggested an interaction of O3 with the pituitary-thyroid-adrenal axis, because
thyroidectomy, hypophysectomy, and adrenalectomy protected against the lethal effects of high
concentrations of O3. Concentrations of 0.7 to 1.0 ppm O3 for a 1 day exposure in male rats
caused changes in the parathyroid; thymic atrophy; decreased serum levels of thyroid stimulating
hormone, triiodothyronine (T3), thyroxine (T4), free T4, and protein binding; and increased
prolactin. In more recent studies, increased toxicity to O3 was reported in hyperthyroid rats by
Huffman et al. (2001) and T3 supplementation was shown to increase metabolic rate and
pulmonary injury in the lungs of O3-treated animals (Sen et al., 1993).
The mechanisms by which O3 affects neuroendocrine function are not well understood.
Cottet-Emard et al. (1997) examined catecholamine activity in rat sympathetic efferents and
brain areas of prime importance to adaptation to environmental stressors. Exposures of
0.5 ppm O3 for 5 days caused inhibition of norepinephrine turnover in heart (-48% of the
control level) but not in lungs and failed to modify the tyrosine hydroxylase activity in superior
cervical ganglia and the catecholamine content in the adrenal glands. In the CNS, O3 inhibited
tyrosine hydroxylase activity in noradrenergic brainstem cell groups and decreased
catecholamine turnover in the cortex (-49%) and striatum (-18%) but not in the hypothalamus.
This suggests that high ambient levels of O3 can produce marked neural disturbances in
structures involved in the integration of chemosensory inputs, arousal, and motor control, effects
that may be responsible for some of the behavioral effects seen with O3 exposure.
5.3.3 Cardiovascular Effects
Studies of the effects on hematological parameters and blood chemistry in rats have shown
that erythrocytes are a target of O3. Exposures to 1.0 ppm O3 for 3 h have been found to
decrease heart rate (HR), mean arterial pressure (MAP), and core temperature (Tco) and to
induce arrhythmias with some exposures in rats. These effects are more pronounced in adult and
awake rats than in younger or sleeping animals. Exposures of 0.2 ppm for 48 h have been shown
5-59
-------
to cause bradycardia, while exposures of 0.1 ppm O3 for 3 days have been shown to cause
bradyarrhythmia in these animals.
A more recent study of rats exposed to FA for 6 h, followed 2 days later by a 5 h exposure
to 0.1 ppm O3, 5 days later by a 5 h exposure to 0.3 ppm O3, and 10 days later by a 5 h exposure
to 0.5 ppm O3, used the head-out plethysmograph for continuous measurements (Arito et al.,
1997). Each of the O3 exposures was preceded by a 1 h exposure to FA. Transient rapid,
shallow breathing with slightly increased HR appeared 1 to 2 min after the start of O3 exposures
and was attributed to an olfactory response. Persistent rapid, shallow breathing with a
progressive decrease in FIR occurred with a latent period of 12 h. During the last 90-min of
exposure, averaged values for relative VE tended to decrease with the increase in O3
concentration for young (4 to 6 months) but not old (20 to 22 months) rats.
Studies by Watkinson et al. (1995, 2001) and Highfill and Watkinson (1996), that utilized
radiotelemetry transmitters in unanesthetized unrestrained rats, demonstrated that when FIR was
reduced during a 5-day, 0.5 ppm O3 exposure, Tco and activity levels also decreased. The
decreases in Tco and BP reported in these studies and by Arito et al. (1997) suggest that the
changes in ventilation and HR are mediated through physiological and behavioral defense
mechanisms in an attempt to minimize the irritant effects of O3 inhalation. Decreased activity
was previously reported in laboratory animals during exposure to O3 (see above).
Similar cardiovascular and thermoregulatory responses in rats to O3 were reported by
Iwasaki et al. (1998). Repeated exposure to 0.1, 0.3, and 0.5 ppm O3 8 h/day for 4 consecutive
days caused disruption of circadian rhythms of HR and Tco on the first and second exposure
days that was concentration-dependent. The decreased HR and Tco recovered to control values
on the third and fourth days of O3 exposure.
The thermoregulatory response to O3 was further characterized by Watkinson et al. (2003).
Male Fischer-344 rats were either exposed to 0.0 ppm for 24 h/day (air), 0.5 ppm for 6 h/day
(intermittent), or to 0.5 ppm for 23 h/day (continuous) at 3 temperatures, 10 °C (cold), 22 °C
(room), or 34 °C (warm). Another protocol examined the effects of O3 exposure (0.5 ppm) and
exercise (described as rest, moderate, or heavy) or CO2-stimulated ventilation. Both intermittent
and continuous O3 exposure caused decreases in HR and Tco and increases in BALF
inflammatory markers. Exercise in FA caused increases in HR and Tco while exercise in O3
caused decreases in those parameters. Carbon dioxide and O3 induced the greatest deficits in HR
5-60
-------
and Tco. Several factors were suggested that may modulate the hypothermic response, including
dose, animal mass, and environmental stress.
Laboratory animals exposed to relatively high O3 concentrations (>0.5 ppm) demonstrate
tissue edema in the heart and lungs. This may be due to increased circulating levels of atrial
natriuretic factor (ANF), which is known to mediate capillary permeability, vasodilation, and BP
(Daly et al., 2002). Increased levels of ANF were reported in the heart, lungs, and circulation of
rats exposed to 0.5 ppm O3 for 8 h ( Vesely et al., 1994a,b,c).
Earlier work demonstrated O3-induced release of functionally active PAF from rodent
epithelial cells and the presence of PAF receptors on AMs. New work examining lipid
metabolism (Section 5.2.1.4) and mediators of inflammatory response and injury
(Section 5.2.3.4) confirm earlier findings indicating that PAF (Kafoury et al., 1999) and PAF
receptors (Longphre et al., 1999) are involved in responses to O3. In addition to the role of PAF
in pulmonary inflammation and hyperpermeability, this potent inflammatory mediator may have
clotting and thrombolytic effects, though this has not been demonstrated experimentally (see
Figure 5-2). This cardiovascular effect may help explain, in part, some limited epidemiologic
findings suggestive of possible association of heart attack and stroke with ambient O3 exposure
(see Chapter 7). The findings of Pulfer and Murphy (2004) and Pulfer et al. (2005), as discussed
in Section 5.2.1.4, which characterize the in vitro and in vivo production of two biologically
active oxysterols, are also suggestive of a mechanism whereby O3 exposure might be implicated
in the increased risk of cardiopulmonary disease.
5.3.4 Reproductive and Developmental Effects
Early studies of pre- and postnatal exposure to O3 were performed at relatively high
concentrations. Teratogenic effects were not observed with intermittent exposures of 0.44 to
1.97 ppm O3 during any part of gestation. Continuous exposure during mid-gestation increased
the resorption of embryos while exposures during late gestation delayed some behavioral
developments (e.g., righting, eye opening). There were no effects on neonatal mortality up to
1.5 ppm O3, whereas some transient effects on weight gain were observed at exposures of
0.6 ppm O3.
More recent studies tend to confirm previous conclusions that prenatal exposures to O3
concentrations <1.0 ppm do not cause major or widespread somatic or neurobehavioral effects in
5-61
-------
the offspring of laboratory animals. These studies generally add some weight toward a negative
interpretation of the importance of contributions of low, ambient O3 to lower birth weights and
gross development defects reported in neonates born to women exposed to typical ambient
pollution (e.g., Renner, 2002; Chen et al., 2002; Ritz and Yu, 1999). Some postnatal O3
exposure studies continue to find a few subtle or borderline somatic and behavioral deficits that
will require further research to better assess potential risk to developing humans.
Studies of somatic and neurobehavioral development in female CD-I mice exposed during
pregnancy (days 7 to 17) to O3 (0, 0.4, 0.8, or 1.2 ppm) failed to show any O3 effects on
reproductive or behavioral performance (Bignami et al., 1994). The study did find significant
decreases in body weight gain and delayed eye opening in pups in the 1.2 ppm exposure group.
The lack of effect on behavioral performance contrasts with earlier findings, which may be due
to the use of different species, differing exposure durations, cross-fostering used in the latter
study, different species, and exposure durations during pregnancy. A second study using CD-I
mice exposed in utero from conception through day 17 of pregnancy to 0, 0.2, 0.4, and
0.6 ppm O3 found no significant deficits in reproductive performance, postnatal somatic and
neurobehavioral development, or adult motor activity (Petruzzi et al., 1995). A third study by
the same group (Petruzzi et al., 1999), using O3 exposures (0.3, 0.6, or 0.9 ppm) that continued
postnatally until weaning, showed subtle changes in handedness and morphine reactivity.
Exposures to 0.6 ppm O3 caused a reduced preference for the right paw in adulthood. Exposures
to 0.9 ppm O3 altered hot plate avoidance after i.p. treatment with morphine in adulthood.
CD-I mice exposed to 0.6 ppm O3 from birth through weaning demonstrated no
impairment of navigational performance during acquisition and only subtle changes during
reversal (DeH'Omo et al., 1995a). Additionally, there were no O3-induced effects on
reproductive performance, but offspring showed a significant reduction in body weight. Effects
on neurobehavioral development with this exposure were minor, with some attenuation of
activity responses and impairment of passive avoidance acquisition (Dell'Omo et al. (1995b).
The offspring of CD-I mice continuously exposed from 30 days prior to the formation of
breeding pairs until postnatal day 17 to 0.0, 0.3, or 0.6 ppm O3 showed only small and selective
effects on somatic and sensorimotor development (Sorace et al., 2001).
Morphological changes were found in the anterior cerebellar lobe of rat pups born to dams
exposed during the entire gestation period to very high (1.0 ppm) O3 concentrations for 12 h/day
5-62
-------
(Rivas-Manzano and Paz, 1999). Additionally, the dams displayed significantly fewer
implantations, increased rate of reabsorptions, a high incidence of spontaneous abortion, and
offspring with low birth weight, as noted by previous investigators.
5.3.5 Effects on the Liver, Spleen, and Thymus
Early investigations of the effects of O3 on liver centered on xenobiotic metabolism, and
the prolongation of sleeping time, which was observed at 0.1 ppm O3. In some species, only
adults and especially females were affected. In rats, high (1.0 to 2.0 ppm for 3 h) acute O3
exposures caused increased production of NO by hepatocytes and enhanced protein synthesis
(Laskin et al., 1994; 1996). The O3-associated effects shown in the liver are thought to be
mediated by inflammatory cytokines or other cytotoxic mediators released by activated
macrophages in the lungs (Vincent et al., 1996; Laskin et al., 1998; Laskin and Laskin, 2001).
Except for the earlier work on xenobiotic metabolism, the responses occurred only after very
high acute O3 exposures.
Examinations of the effects of O3 on spleen and thymus have shown that O3 primarily
affects T-cell mediated systemic immunity. As with the O3-associated effects shown in the liver,
most of the statistically significant changes occurred after acute exposures to very high O3
concentrations and relate to systemic oxidative stress. Using more relevant ambient urban O3
exposure patterns, effects were not found on systemic immune function of rats.
5.3.6 Effects on Cutaneous and Ocular Tissues
Ozone exposure not only affects various organ systems, when inhaled, but also has direct
effects on the exposed skin and eyes. The outermost layer of the skin (SC, stratum corneum)
may be oxidized, which can lead to compromise of the skin barrier and an epidermal pro-
inflammatory response (Weber et al., 2001; Thiele, 2001). These effects are found only at very
high concentrations (>1 to 5 ppm) and have not been shown at more relevant ambient levels of
exposure. The skin possesses a well-developed defense system against oxidative stress, utilizing
nonenzymatic (e.g., vitamin C, GSH, UA, a-tocopherol) and enzymatic (e.g., SOD, catalase, GR,
and GSHPx) antioxidants (Cross et al., 1998). Ocular tissues have similar antioxidant protective
functions as the skin but are not as well studied (Mucke, 1996; Rose et al., 1998). Effects of
ground-level smog on the eyes have been reported but generally are attributed to related
5-63
-------
photochemical oxidants like peroxyacetyl nitrate (PAN) (Vyskocil et al., 1998) or possibly to
atmospheric O3 precursors or reaction products like aldehydes. As in other tissues, O3 may have
disparate high-dose and low-dose mechanisms of effect on skin and eyes; so, results must be
interpreted in this light.
Hairless mice (SKH-1) exposed to O3 (0.8 to 10 ppm for 2 h) were used to demonstrate
that O3 depletes the low molecular weight antioxidants (e.g., a-tocopherol, vitamin C, GSH, UA)
in the SC at >1.0 ppm and causes increased MDA at >5 ppm (Weber et al., 1999, 2000, 2001).
Valacchi et al. (2000) demonstrated that preexposure to 0.5 O3 for 2 h followed by low-dose
ultraviolet (UV) radiation (0.33 MED) caused depletion of a-tocopherol. This suggests that
combined low doses of UV radiation and near-ambient levels of O3 may cause oxidative stress
on the SC. Prolonged exposure to 0.8 ppm O2 for 6 h also induces cellular stress responses that
included the formation of HNE protein adducts, HSP27, and heme-oxygenase-1 in the deeper
cellular layers of the skin that continued for up to 18 h after O3 exposure, followed by repair
processes (Valacchi et al., 2003).
The importance of O3 and UV-induced cellular protein oxidation found in murine skin
models to possibly similar environmentally-induced changes in human SC keratins was
identified by Thiele et al. (1998, 1999) and Thiele (2001). Using the presence of carbonyl
groups in proteins as a marker of reactive oxygen-mediated protein oxidation, they reported
higher carbonyl levels in the upper SC from the tanned skin of humans and in the skin of healthy
human volunteers exposed to model chemical oxidants (e.g., hypochlorite, benzoyl peroxide)
that were inversely correlated with a-tocopherol levels. The environmentally-induced oxidative
damage identified in human SC represents an early pathophysiological stage in the development
of barrier disruption and inflammation, and possibly has implications for the process of
desquamation. The relevance of potentiation of environmental oxidative stress by O3 exposure
of human skin needs further study.
5.3.7 Summary and Conclusions—Systemic Effects of Ozone
Neurobehavioral effects of O3 at concentrations of 0.2 to 1.0 ppm include decreased motor
activity, short- and long-term memory deficits, increased freezing behavior, and decreased
exploratory behaviors. These effects have been associated with reactive oxygen/nitrogen
species, ozonation products, altered neurotransmitter levels, morphological changes in several
5-64
-------
brain regions, and altered EEG patterns during sleep. Neuroendocrine effects of O3 include
morphological and hormonal changes in the pituitary-thyroid-adrenal axis at concentrations
of-0.75 ppm and alterations of visual and olfactory neural pathways at concentrations >1 ppm.
Mechanisms underlying these effects are not understood at this time. Cardiovascular effects
of O3 at concentrations of 0.3 to 0.5 ppm include decreased HR, Tco, and BP, which have been
termed a hypothermic response. Concentrations of O3 >0.5 ppm cause tissue edema (possibly
mediated by ANF).
Prenatal exposures to O3 concentrations <1.0 ppm did not cause noticeable somatic or
neurobehavioral effects in offspring, while concentrations of 1.0 to 1.5 ppm caused varying
effects on neonatal mortality. Some studies have shown an effect of O3 on liver xenobiotic
enzymes at concentrations as low as 0.1 ppm, while other studies have shown no alterations in
metabolic enzymes at even 1 ppm, with the effects appearing to be highly-species specific.
Effects on spleen and thymus appear to only occur at high O3 concentrations (>1.0 ppm), while
relevant ambient urban exposures have no effect on systemic immune function in rats. Effects
of O3 on cutaneous and ocular tissue are only seen at high, nonrelevant concentrations.
5.4 INTERACTIONS OF OZONE WITH OTHER CO-OCCURRING
POLLUTANTS
Ozone is part of a complex mixture of air pollutants with a composition and pattern that
varies geographically and temporally (by hour of the day, day of the week, and season). Health
effects caused by the complex mixture are undoubtedly different (either subtly or significantly)
from the additive effects of a few of the hundreds of compounds present. The only disciplinary
approach that can evaluate a "real-world" complex mixture is epidemiology (Chapter 7). Still,
because of the difficulty in evaluating causative factors and quantitative relationships in
epidemiology studies, it is useful to consider animal toxicological studies of mixtures. Such
studies can be divided into three categories: (1) ambient air mixtures, (2) laboratory-generated
complex mixtures (e.g., gasoline combustion mixtures having UV-irradiation, other reaction
mixtures with O3 and several other components), and (3) binary mixtures. In most cases,
experimental designs in the first two classes did not have an O3-only group, making it difficult or
impossible to discern specifically the influence of O3 per se. The more recent mixture studies
5-65
-------
that are discussed here typically have been with NO2, sulfuric acid (H2SO4), or ammonium
sulfate ((NH4)2SO4).
Interpreting the mixture studies in terms of real-world risk is difficult because laboratory
exposure patterns do not always represent real-world exposure patterns. As shall be seen, all
interaction possibilities have occurred, depending upon the composition of the mixture, the
endpoint examined, and the exposure regimen. In some cases, no interaction was found. Most
often, additivity (the effects of the mixture are equal to the sum of the effects of the individual
components) or synergism (the effects of the mixture are greater than the sum of the effects of
the individual components) was observed. Antagonism (the effects of the mixture are less than
the sum of the individual components) was rarely found.
5.4.1 Ozone and Nitrogen Oxides
The most commonly studied copollutant in binary mixtures with O3 is NO2. Both earlier
studies and more recent work indicate that, although interaction may occur between these two
pollutants, in general, O3 often masked the effects of the NO2 or accounted for most of the
response, due to the greater toxicity of O3. Very generally, additivity occurred after acute
exposure and synergism occurred with prolonged exposure. Interpreting the O3/NO2 mixture
studies is challenging, because laboratory exposure patterns rarely simulate real-world exposure
patterns. NO2 typically peaks before O3, with a mixture occurring between the individual gas
peaks, but most laboratory exposures used mixtures only. Also, most studies of O3 and NO2
mixtures used ambient levels of O3 and levels of NO2 high above ambient. Table AX5-15 lists
more recent studies evaluating coexposures to NO2 and O3.
Chronic exposures of rats to O3 (0.8 ppm) and NO2 (14.4 ppm) for 6 h/day caused
development of respiratory insufficiency and severe weight loss. Half of these animals died after
55 to 78 days of exposure due to severe fibrosis (Farman et al., 1997). Increased total lung
collagen and elastin were observed, with loss of mature collagen, suggesting breakdown and
remodeling of the lung parenchyma. Morphological examination following these coexposures
demonstrates a sequence of events starting with increasing inflammatory and mild fibrotic
changes for the first 3 weeks of exposure stabilized or even reduced changes after 4 to 6 weeks,
and severe increases over 7 to 9 weeks of exposure (Farman et al., 1999). This suggests that
repair processes occurring during the middle 4 to 6 weeks of exposure become overwhelmed,
5-66
-------
leading to progressive fibrosis after 7 to 8 weeks of exposure. When the coexposure was
extended for 90 days, lesions were noted far into the acinus, but the extent of tissue involvement
was the same after 7, 78, and 90 days of exposure. At the end of exposure, high levels of
procollagen types I and III mRNA were observed within central acini in the lungs from the
combined exposure group but not in lungs from the rats exposed to O3 or NO2 alone.
Sprague-Dawley rats exposed to 0.3 ppm O3 and the combined exposure of O3 and
1.2 ppm NO2 for 3 days demonstrated significant DNA single-strand breaks in AMs (Bermudez
et al., 1999). No changes were caused by NO2-only exposure. The same exposures stimulated
the activity of poly (adenosinediphosphate-ribose) (polyADPR) synthetase, suggesting a
response to lung cellular DNA repair caused by oxidant-induced lung injury (Bermudez, 2001).
The laboratory animal model of progressive pulmonary fibrosis, utilizing long-term combined O3
(0.4 to 0.8 ppm) and high-level NO2 (7 to 14 ppm) exposure, causes an initial acute pulmonary
inflammation, followed by adaptation and repair, and eventually causing pulmonary fibrosis
after 6 to 13 weeks of exposure (Ishii et al., 2000a; Weller et al., 2000). Unfortunately, this
model is not very useful for understanding potential interactive effects of ambient concentrations
ofO3andNO2.
5.4.2 Ozone and Other Copollutants
Ozone and Formaldehyde
Early studies with combined exposures to O3 and formaldehyde (HCHO) found evidence
of both synergistic and non-interactive effects. Newer work listed in Table AX5-16 includes
studies of biochemical and histopathological endpoints in rats exposed to 0.4 ppm O3 and
3.6 ppm HCHO, alone and combined, for 8 h/day for 3 days (Cassee and Feron, 1994). They
demonstrated no interactive effects in the nasal respiratory epithelium, despite the high levels of
HCHO when compared to typical ambient levels of 1 to 10 ppb (e.g., Rehle et al., 2001). Mautz
(2003) studied changes in breathing pattern and epithelial cell proliferation using exposures of
0.6 ppm O3 and 10 ppm HCHO alone and in combination for 3 h with exercise at two times
resting ventilation. Even with exercise, HCHO does not substantially penetrate to the lower
respiratory tract to interact with O3 and does not alter breathing patterns to modify local O3 dose.
Parenchymal injury was, therefore, due to O3 alone. In the nasal transitional epithelium and in
the trachea, however, combined exposure produced additive effects due to the increased volume
5-67
-------
of toxicants during exercise. No other combined pollutant studies have been published in the
peer-reviewed literature, although two studies compared the respiratory effects of O3 to HCHO.
Nielsen et al. (1999) compared upper airway sensory irritation caused by HCHO concentrations
up to 4 ppm to the lower airway irritation caused by O3. Using BALB/c mice, they continuously
measured fB, VT, expiratory flow, time of inspiration, time of expiration, and respiratory patterns
during acute, 30-min exposures. They reported a no-effect level of 0.3 ppm for HCHO and
1.0 ppm for O3.
Thus, O3 and HCHO do not appear to have additive effects except during exercise and that
is due to an increased volume of gas reaching the tissue. Any possible synergism occurs in the
nasal epithelium. HCHO exerts its effects primarily in the upper respiratory tract, whereas the
primary site of acute cell injury from O3 occurs in the conducting airways. EPA is currently
completing a toxicological and epidemiological review and risk characterization for HCHO.
Ozone and Tobacco Smoke
Early studies of combined exposures of O3 (1 ppm) and tobacco smoke demonstrated
altered airway responsiveness to inhaled bronchoconstrictor challenge and tracheal vascular
permeability in guinea pigs. Table AX5-17 lists studies completed since the 1996 AQCD which
evaluated tobacco smoke/O3 coexposures.
Wu et al. (1997) reported that inhalation of cigarette smoke evokes a transient
bronchoconstrictive effect in anesthetized guinea pigs. Total pulmonary resistance and Cdyn were
compared before and after acute exposure to 1.5 ppm O3for 1 h. Cigarette smoke alone (7 mL)
at a low concentration (33%) induced a mild and reproducible bronchoconstriction that slowly
developed and reached its peak after a delay of >1 min. After O3 exposure, the same cigarette
smoke inhalation challenge evoked an intense bronchoconstriction that occurred more rapidly,
reaching its peak within 20 s, and being sustained for >2 min. Pretreatment with selective
antagonists of neurokinin type 1 and 2 receptors completely blocked the enhanced airway
responsiveness suggesting that O3 exposure induced AHR to inhaled cigarette smoke, primarily
from the bronchoconstrictive effect of endogenous tachykinins.
The above studies were conducted with undiluted tobacco smoke and high O3
concentrations. To determine the effects of aged and diluted sidestream cigarette smoke (ADSS)
as a surrogate of environmental tobacco smoke (ETS) on O3-induced lung injury, Yu et al.
5-68
-------
(2002) exposed male B6C3F1 mice to (1) FA, (2) ADSS, (3) O3, or (4) ADSS followed by O3
(ADSS/O3). Exposure to 30 mg/m3 ADSS, 6 h/day for 3 days, followed by exposure to
0.5 ppm O3 for 24 h was associated with a significant increase in the number of cells recovered
by BAL compared with exposure to ADSS alone or O3 alone. Neutrophils, lymphocytes, and
total protein levels in BAL were increased following the combined exposure when compared
with all other groups. Within the CAR, the percentage of proliferating cells was unchanged from
control following exposure to ADSS alone but was significantly elevated following exposure
to O3 and further augmented in a statistically significant manner in mice exposed to ADSS/O3.
Following exposure to O3 alone or ADSS/O3, the ability of AMs to release IL-6 under LPS
stimulation was significantly decreased, while exposure to ADSS alone or ADSS/O3 caused a
significantly increased release of TNF-a from AMs under LPS stimulation. These data suggest
that ADSS exposure enhances the sensitivity of animals to O3-induced lung injury.
Acute exposure to ETS also may make a healthy person more susceptible to sequential O3
exposure by affecting lung barrier function or the underlying epithelium. Toxicological studies
with components of ETS (e.g., nicotine receptor agonists, acrolein, and oxidants) have shown
that the vagal bronchopulmonary C-fibers are stimulated by acute exposures that initiate both
central and local responses (Bonham et al., 2001; Mutoh et al., 2000). The central responses
(e.g., tachypnea, cough, bronchoconstriction, increased mucous secretion) are more protective of
the lungs; however, local responses may include increased sensitization of the C-fibers to other
irritants, including O3. Active tobacco smokers should not be similarly affected, because they
already have significant chronic airway inflammation and increased mucus production. In fact,
chronic smokers appear to have diminished lung function responses to O3 (see Chapter 6).
5.4.3 Complex (Multicomponent) Mixtures Containing Ozone
Ambient pollution in most areas is a complex mix of more than two chemicals. A number
of new studies have examined the effects of exposure to multicomponent atmospheres
containing O3. Some of these studies attempted to simulate photochemical reaction products
occurring under actual atmospheric conditions. However, the results of these studies are often
difficult to interpret because of chemical interactions between the components, as well as the
resultant production of variable amounts of numerous secondary reaction products, and a lack of
precise control over the ultimate composition of the exposure environment. In addition, the role
5-69
-------
of O3 in the observed biological responses is often obscure. Prior studies using irradiated
automobile exhaust mixtures containing total oxidant concentrations (expressed as O3) in the
range of 0.2 to 1.0 ppm have demonstrated pulmonary function changes in several species.
A more recent attempt has been made to examine multicomponent mixtures resulting from
the reaction of O3 with unsaturated hydrocarbons [e.g., isoprene (C5H8) and terpene (C10H16)],
which produces HCHO, formic acid, acetone, acrolein, acetic acid, and other oxidation products
(many of which are strong airway irritants). Wilkins et al. (2001) evaluated sensory irritation by
measuring mean fB in the mouse bioassay and found a 50% reduction after 30 min of exposure to
reaction products of O3 and isoprene. The mixture at this time period contained <0.2 ppm O3, so
the authors attributed the observed effects to the oxidation products. Clausen et al. (2001), using
the same mouse model, evaluated the reaction products of O3 and limonene. A 33% reduction in
mean fB was produced after 30 min of exposure to the mixture containing <0.3 ppm O3, again
implicating the effects of strong irritant products. Further work needs to be done with these
complex reaction mixtures, because of their potential impact on the respiratory tract. The results
would be particularly important, however, to the reaction of O3 indoors (see Chapter 3).
Pollutant mixtures containing acid aerosols comprise another type of commonly examined
exposure atmosphere (studies summarized in Table AX5-18). Earlier studies that employed
simultaneous single, repeated, or continuous exposures of various animal species to mixtures of
acid sulfates and O3 found responses for several endpoints, including tracheobronchial
mucociliary clearance, alveolar clearance, pulmonary mechanics, and lung morphology, to be
due solely to O3 Some synergism was noted for bacterial infectivity, response to antigen, and
effects on lung protein content and the rate of collagen synthesis.
More recent studies found some differences in airway responses to inhaled acid particle-O3
mixtures that may have been partly due to airway dosimetry. Various physical and chemical
mechanisms may be responsible (see Schlesinger, 1995). For example, physical adsorption or
absorption of O3 or its reaction products on a particle could result in transport to more sensitive
sites, or to sites where O3, by itself, would not normally be reactive (Madden et al., 2000).
Chemical reactions on the surface of particles can form secondary products that are more
lexicologically active, or chemical characteristics of the particle may change the residence time
or reactivity of oxidation products at the site of deposition. The hypothesis that synergism
between O3 and sulfates is due to decreased pH changing the residence time or reactivity of
5-70
-------
reactants, such as free radicals, was tested by Chen et al. (1995) and El-Fawal et al. (1995).
Male New Zealand White rabbits were exposed for 3 h to 125 |ig/m3 H2SO4, 0.1, 0.3, or
0.6 ppm O3, and to combinations. Chen et al. (1995) demonstrated that decreased pH following
exposure to acid aerosol was correlated with phagocytic activity and capacity of harvested
macrophages and that exposure to O3/H2SO4 removed this relationship. El-Fawal et al. (1995)
showed that responsiveness of rabbit harvested bronchial rings to ACh was increased following a
3 h O3 exposure, but that 0.1 to 0.6 ppm O3/0.5 to 0.125 mg/m3 H2SO4 combinations resulted in
antagonism.
As discussed in Section 5.2.2.1, Churg et al. (1996) demonstrated increased uptake of
asbestos or TiO2 in response to 10 min O3 (up to 1.0 ppm) preexposure, suggesting that low
concentrations of O3 may increase the penetration of some types of PM into epithelial cells.
Using human epithelial cell cultures, Madden et al. (2000) demonstrated a greater potency for
ozonized diesel PM to induce PGE2 production. This suggests that 0.1 ppm O3 for 24 h can
modify the biological activity of PM derived from diesel exhaust.
Effects of combined exposures of O3 and resuspended urban particles on cell proliferation
in epithelial cells of the terminal bronchioles and the alveolar ducts were examined by Vincent
et al. (1997) and Adamson et al. (1999). Rats exposed to 0.8 ppm O3 in combination with 5 or
50 mg/m3 particles for 4 h demonstrated greatly potentiated proliferative effects compared to O3
exposure alone. These findings using resuspended dusts, although at high concentrations, are
consistent with the studies demonstrating interaction between H2SO4 aerosols and O3. Effects of
acute coexposure to 0.6 ppm O3 and fine or ultrafine H2SO4 (0.5 to 0.3 mg/m3) aerosols on lung
morphology were examined by Kimmel et al. (1997). They demonstrated that alveolar septal
volume was increased in animals coexposed to O3 and ultrafine, but not fine, H2SO4.
Interestingly, cell proliferation was increased only in animals coexposed to fine H2SO4 and O3,
as compared to animals exposed to O3 alone. Subchronic exposure to acid aerosols (20 to
150 |ig/m3 H2SO4) had no interactive effect on the biochemical and morphometric changes
produced by either intermittent or continuous exposure to 0.12 to 0.2 ppm O3 for up to 90 days,
which suggests that the interactive effects of O3 and acid aerosol coexposure in the lung
disappeared during the long-term exposure (Last and Pinkerton, 1997). Sindhu et al. (1998)
observed an increase in rat lung putrescine levels after repeated, combined exposures to O3 and a
nitric acid (HNO3) vapor for 40 weeks.
5-71
-------
Other studies have examined interactions between carbon particles and O3. The
interactions of intratracheally instilled carbon particles followed by either a 7-day or 60-day
exposure to 0.5 ppm O3 in rats was evaluated by Creutzenberg et al. (1995). The carbon
particles caused diminished phagocytotic capacity and chemotactic migration capability of AMs
that was stimulated by the subsequent O3 exposure. Inflammatory responses following
exposures to low- and high-concentration mixtures of O3 and acidic aerosols (0.2 ppm O3 +
50 |ig/m3 carbon + 100 |ig/m3 H2SO4; 0.4 ppm O3 + 250 |ig/m3 carbon + 500 |ig/m3 H2SO4,
respectively) for 1 or 5 days was examined by Kleinman et al. (1999). The response with
the O3-particle mixture was greater after 5 days (4 h/day) than after day 1. This contrasted
with O3 exposure alone (0.4 ppm), which caused marked inflammation on acute exposure,
but no inflammation after 5 consecutive days of exposure.
The effects of a mixture of elemental carbon particles, 0.2 ppm O3, and 0.5 mg/m3
ammonium bisulfate on rat lung collagen content and macrophage activity was examined by
Kleinman et al. (2000). Decreases in lung collagen, and increases in macrophage respiratory
burst and phagocytosis were observed relative to other pollutant combinations. Mautz et al.
(2001) used a similar mixture (i.e., elemental carbon particles, 0.16 to 0.59 ppm O3, ammonium
bisulfate 0.5 to 0.22 mg/m3, but with 0.11 to 0.39 ppm NO2 also) and exposure regimen as
Kleinman et al. (2000). Also observed were decreases in pulmonary macrophage Fc-receptor
binding and phagocytosis and increases in acid phosphatase staining. Bronchoalveolar epithelial
permeability and cell proliferation were increased. Altered breathing-patterns also were
observed, with some adaptations occurring.
Bolarin et al. (1997) exposed rats to 50 or 100 |ig/m3 carbon particles in combination with
ammonium bisulfate and 0.2 ppm O3. Despite 4 weeks of exposure, they observed no changes in
protein concentration in BALF or in blood prolyl 4-hydroxylase, an enzyme involved in collagen
metabolism. Slight decreases in plasma fibronectin were present in animals exposed to the
combined pollutants versus O3 alone. Thus, the potential for adverse effects in the lungs of
animals challenged with a combined exposure to particles and gaseous pollutants is dependent
on numerous factors, including the gaseous copollutant, concentration, and time.
In a complex series of studies, Oberdorster and colleagues examined the interaction of
several pulmonary oxidative stress pollutants. Elder et al. (2000a,b) reported the results of
combined exposure to ultrafine carbon particles (100 |ig/m3) and O3 (1 ppm for 6 h) in young
5-72
-------
and old Fischer 344 rats that were pretreated with aerosolized endotoxin. In old rats, exposure to
carbon and O3 produced an interaction that resulted in a greater influx in neutrophils than that
produced by either agent alone. This interaction was not seen in young rats. Oxidant release
from lavage fluid cells also was assessed and the combination of endotoxin, carbon particles,
and O3 produced an increase in oxidant release in old rats. This mixture produced the opposite
response in the cells recovered from the lungs of the young rats, indicating that the lungs of the
aged animals underwent greater oxidative stress in response to a complex pollutant mix of
particles, O3, and a biogenic agent. Johnston et al. (2000a, 2002) reported the results of
combined exposure to O3 (1.0 and 2.5 ppm for 4, 20, or 24 h) and low-dose endotoxin, or to O3
and endotoxin separately, in newborn and adult C57BL/6J mice. In the first study, adult
(8 weeks old) mice showed greater sensitivity to O3 than newborn (36 h old) mice on the basis of
mRNAs encoding for various chemokines and cytokines. In contrast, adult and newborn mice
responded similarly 2 h after endotoxin exposure (10 ng for 10 min), suggesting that age
differences in O3-generated inflammation is secondary to epithelial cell injury. In the second
study, 8-week-old mice exposed to O3 (1 ppm for 24 h) followed by endotoxin (37.5 EU for
10 min) showed increased responsiveness over either exposure alone, on the basis of increased
expression of chemokine and cytokine messages and increased BALF levels of protein and
PMNs.
Fanucchi et al. (1998) and Wagner et al. (2001a,b) examined the synergistic effect of
coexposure to O3 and endotoxin on the nasal transitional epithelium of rats that also was
mediated, in part, by neutrophils. Fisher 344 rats intranasally instilled with endotoxin and
exposed to 0.5 ppm O3, 8 h per day for 3 days developed mucous cell metaplasia in the nasal
transitional epithelium, an area normally devoid of mucous cells; whereas, intratracheal
instillation of endotoxin (20 jig) caused mucous cell metaplasia rapidly in the respiratory
epithelium of the conducting airways. A synergistic increase of intraepithelial mucosubstances
and morphological evidence of mucous cell metaplasia were found in rat maxilloturbinates upon
exposure to both O3 and endotoxin, compared to each pollutant alone. A similar response was
reported in OVA-sensitized Brown Norway rats exposed to 0.5 ppm O3, 8 h/day for 3 days
(Wagner et al., 2002), indicating that coexposure to O3 and inflammatory biogenic substances
like allergens (e.g., OVA) or bacterial endotoxin can augment epithelial and inflammatory
responses in rat nasal passages.
5-73
-------
In follow-up studies, Wagner et al. (2003) reported that coexposure of rats to O3 and
endotoxin also enhanced epithelial and neutrophilic inflammatory responses in the pulmonary
airways. Fisher 344 rats were intranasally instilled with endotoxin and exposed to 1.0 ppm O3
for 8 h, which was repeated 24 h later. Three days after the last exposure, BALF was analyzed
for inflammatory cells and secreted mucosubstances (mucin SAC), and lung tissue was
processed for morphometric analysis. Endotoxin instillation alone caused a dose-dependent
increase in BALF neutrophils that was further increased 2-fold in O3-exposed rats given 20 jig
endotoxin, the highest dose. Mucin glycoprotein SAC also was increased in the BALF at this
dose but not at lower endotoxin doses. Ozone exposure alone did not cause mucus
hypersecretion, but it did potentiate mucus secretion in rats given both 2 and 20 jig endotoxin
and increased intraepithelial mucosub stances 2-fold, which was further substantiated by
significant increases in mucin gene (rMucSAC) mRNA levels in the conducting airways.
The effect of O3 modifying the biological potency of PM (diesel PM and carbon black) was
examined by Madden et al. (2000) in rats. Reaction of National Institute of Standards and
Technology (NIST) Standard Reference Material # 2975 diesel PM with 0.1 ppm O3 for 48 h
increased the potency (compared to unexposed or air-exposed diesel PM) to induce neutrophil
influx, total protein, and LDH in lung lavage fluid in response to intratracheal instillation.
Exposure of the diesel PM to high, nonambient O3 concentration (1.0 ppm) attenuated the
increased potency, suggesting destruction of the bioactive reaction products. Unlike the diesel
particles, carbon black particles exposed to 0.1 ppm O3 did not exhibit an increase in biological
potency, which suggested that the reaction of organic components of the diesel PM with O3 were
responsible for the increased potency.
Ulrich et al. (2002) investigated the effect of ambient PM from Ottawa Canada (EHC-93)
on O3-induced inflammation. Male Wistar rats were exposed to 0.8 ppm O3 for 8 h and allowed
to recover before intratracheal instillation of 0.5, 1.5, and 5 mg of EHC-93 in 0.3 mL of saline.
The high concentrations of PM used were sufficient to induce pulmonary inflammation, which
was not exacerbated by preexposure to O3. Rats from the combined exposure group had higher
and more persistent lung lavage protein and albumin levels, as well as increased plasma
fibrinogen levels, when compared to PM exposure alone.
The interaction of PM and O3 was further examined in a murine model of O VA-induced
asthma. Kobzik et al. (2001) investigated whether coexposure to inhaled, concentrated ambient
5-74
-------
particles (CAPs) from Boston, MA and to O3 could exacerbate asthma-like symptoms. On days
7 and 14 of life, half of the BALB/c mice used in this study were sensitized by i.p. injection of
OVA and then exposed to OVA aerosol on three successive days to create the asthma phenotype.
The other half received the i.p. OVA but were exposed to a phosphate-buffered saline aerosol
(controls). The mice were further subdivided (n > 6I/group) and exposed for 5 h to CAPs,
ranging from 63 to 1,569 |ig/m3, 0.3 ppm O3, CAPs + O3, or to FA. Pulmonary resistance and
airway responsiveness to an aerosolized MCh challenge were measured after exposures.
A small, statistically significant increase in RL and airway responsiveness, respectively, was
found in both normal and asthmatic mice immediately after exposure to CAPs alone and to
CAPs + O3 but not to O3 alone or to FA. By 24 h after exposure, the responses returned to
baseline levels. There were no significant increases in airway inflammation after any of the
pollutant exposures. In this well-designed study of a small-animal model of asthma, O3 and
CAPs did not appear to be synergistic. In further analysis of the data using specific elemental
groupings of the CAPs, the acutely increased RL was found to be associated with the AISi
fraction of PM. Thus, some components of concentrated fine particulate matter (PM2 5) may
affect airway caliber in sensitized animals, but the results are difficult to extrapolate to people
with asthma.
Animal studies have examined the adverse cardiopulmonary effects of complex mixtures in
urban and rural environments of Italy (Gulisano et al., 1997), Spain (Lorz and Lopez, 1997), and
Mexico (Vanda et al., 1998; Moss et al., 2001). Some of these studies have taken advantage of
the differences in pollutant mixtures of urban and rural environments to report primarily
morphological changes in the nasopharynx and lower respiratory tract (Gulisano et al., 1997;
Lorz and Lopez, 1997) of lambs and pigeons, respectively, after natural, continuous exposures to
ambient pollution. Each study has provided evidence that animals living in urban air pollutants
have greater pulmonary changes than those that would occur in a rural, and presumably cleaner,
environment. However, these studies either did not report ambient O3 levels or reported only
annual means.
The study by Moss et al. (2001) examined the nasal and lung tissue of rats exposed
(23 h/day) to Mexico City air for up to 7 weeks and compared them to controls similarly exposed
to FA. No inflammatory or epithelial lesions were found using quantitative morphological
techniques; however, the concentrations of pollutants were low. Extrapolation of these results to
5-75
-------
humans is restricted, however, by uncontrolled exposure conditions, small sample sizes, and
other unknown exposure and nutritional factors in the studies in mammals and birds, and the
negative studies in rodents. They also bring up the issue of which species of "sentinel" animals
is more useful for predicting urban pollutant effects in humans. Thus, in these field studies, it is
difficult to assign a specific role to any specific component of the mixture for the significant
cardiopulmonary effects reported.
Similar morphological changes (Calderon-Garciduefias et al., 2000a, 2001) and chest X-ray
evidence of mild lung hyperinflation (Calderon-Garciduefias et al., 2000b) have been reported in
children residing in urban and rural areas of Mexico City. The ambient air in urban areas,
particularly in southwestern Mexico City, is a complex mixture of particles and gases, including
high concentrations of O3 and aldehydes that previously have been shown to cause airway
inflammation and epithelial lesions in humans (e.g., Calderon-Garciduefias et al., 1992, 1994,
1996) and laboratory animals (Morgan et al., 1986; Heck et al., 1990; Harkema et al., 1994,
1997a,b). The described effects demonstrate a persistent, ongoing upper and lower airway
inflammatory process and chest X-ray abnormalities in children residing predominantly in highly
polluted areas. Again, extrapolation of these results to urban populations of the United States is
difficult, because of the unique complex mixture of urban air in Mexico City, uncontrolled
exposure conditions, and other unknown exposure and nutritional factors.
5.4.4 Summary and Conclusions—Interactions of Ozone with Other
Co-Occurring Pollutants
It is difficult to summarize the role that O3 plays in exposure responses to binary mixtures,
and even harder to determine its role in responses to multicomponent, complex atmospheres.
Though the specific mechanisms of action of the individual pollutants within a mixture may be
known, the exact bases for toxic interactions have not been elucidated clearly. Certain generic
mechanisms that may underlie pollutant interactions: (1) physical, involving adsorption of one
pollutant onto another and subsequent transport to more or less sensitive sites or to sites where
one of the components of the mixture normally would not deposit in concentrated amounts
(probably not involved in O3-related interactions); (2) production of secondary products that may
be more lexicologically active than the primary materials, demonstrated or suggested in a
number of studies as a basis for interaction between O3 and NO2 and between O3 and PM;
5-76
-------
(3) biological or chemical alterations at target sites that affect response to O3 or the copollutant,
H2SO4, which has been suggested to underlie interactions with mixtures of O3 and acid sulfates;
(4) O3- or copollutant-induced physiological change, such as alteration in ventilation pattern,
resulting in changes in the penetration or deposition of one pollutant when another is present.
This has been implicated in enhanced responses to various O3-containing mixtures with exercise.
Evaluation of interactions between O3 and copollutants is a complex task. Responses are
dependent on a number of host and environmental factors, such that different studies using the
same copollutants may show different types or magnitudes of interactions. The occurrence and
nature of any interaction is dependent on the endpoint being examined and is also highly related
to the specific conditions of each study, such as animal species, health status, exposure method,
dose, exposure sequence, and the physicochemical characteristics of the copollutants. Because
of this, it is difficult to compare studies, even those examining similar endpoints, that were
performed under different exposure conditions. Thus, any description of interactions is really
valid only for the specific conditions of the study in question and cannot be generalized to all
conditions of exposure to a particular chemical mixture. Furthermore, it is generally not possible
to extrapolate the effect of pollutant mixtures from studies of the effects of each component
when given separately. In any case, what can be concluded from the database is that interactions
of O3-containing mixtures are generally synergistic (antagonism has been noted in a few studies),
depending on the various factors noted above and that O3 may produce more significant
biological responses as a component of a mixture than when inhaled alone. Furthermore,
although most studies have shown that interaction occurs only at higher than ambient
concentrations with acute exposure, some have demonstrated interaction at more
environmentally relevant levels (e.g., 0.05 to 0.1 ppm O3 with NO2) and with repeated exposures.
5.5 EFFECTS OF OTHER PHOTOCHEMICAL OXIDANTS
Peroxyacetyl nitrate and peroxypropionyl nitrate (PPN) are the most abundant non-O3
oxidants in ambient air of industrialized areas, other than the inorganic nitrogenous oxidants
such as NO2, and possibly F£NO3. Ambient levels of PAN and PPN were reported to be
decreasing over the 1990s, and available air quality data (Grosjean et al., 2001; Grosjean, 2003;
Jakobi and Fabian, 1997) indicate that present peak concentrations of PAN and PPN in ambient
5-77
-------
air from urban areas are in the low parts per billion range (e.g., <1 to 10 ppb). The levels found
in nonurban areas are considerably lower (Gaffney et al., 1993).
Reactions occur in the troposphere between O3 and hydrocarbons (e.g., D-limonene) to
produce epoxides, hydroperoxides, and peroxides. The majority of the measured ambient
hydroperoxides produced is H2O2, although a small amount of organic hydroperoxides (ROOH)
also may be formed. Friedlander and Yeh (1998) have estimated that atmospheric aerosols can
carry as high as 1 mM of H2O2 and ROOH (e.g., hydroxymethylhydroperoxide) in the associated
water. In vitro cell and tissue damage are induced by high concentrations of liquid phase H2O2
(50 |iM to 1 mM). Morio et al. (2001) (see Table AX5-19) demonstrated that a 2 h exposure of
10 and 20 ppb of inhaled H2O2 vapor can penetrate into lower lung regions where it causes
inflammation. It is likely that hygroscopic components of PM transport ambient H2O2 into the
lower lung and induce tissue injury as well. Exposure of rats to a H2O2-fine particle mixture
(215 or 429 |ig/m3 (NH4)SO4) resulted in increased neutrophil influx, and production of
inflammatory mediators by AMs (Morio et al., 2001). Hygroscopic secondary organic aerosols
generated by O3/hydrocarbon reactions and their co-occurrence with H2O2 also provide another
possible mechanism, yet to be validated, whereby H2O2 can be transported into the lower
respiratory tract (e.g., Friedlander and Yeh, 1998). Interaction of inhaled O3 with unsaturated
fatty acids on cell membranes and mucus in the airways generates epoxides, hydroperoxides, and
secondary ozonation products such as HNE (see Section 5.2.1)
Inhalation toxicological information on the effects of the non-O3 oxidants has been limited
to a few studies on PAN, but at concentrations much higher (approximately 100- to 1,000-fold)
than levels typically found in ambient air. Such acute high levels can cause changes in lung
morphology, behavioral modifications, weight loss, and susceptibility to pulmonary infections.
Therefore, acute toxicity of PAN is much lower than O3, and it is unlikely that present ambient
PAN levels would affect pulmonary function responses to O3 (reviewed in Vyskocil et al., 1998).
Cytogenetic studies indicate that PAN is not a potent mutagen, clastogen, or DNA damaging
agent in mammalian cells in vivo or in vitro at concentrations several orders of magnitude higher
than those generally encountered in ambient air in most cities (Vyskocil et al., 1998; Kligerman
et al., 1995; Heddle et al., 1993). Some studies suggest that PAN may be a weak bacterial
mutagen at concentrations much higher than exist in present urban atmospheres (DeMarini et al.,
2000; Kleindienst et al., 1990).
5-78
-------
An additional level of complexity exists due to the possibility that other ambient oxidants
may contribute to effects attributed to O3. As discussed in Chapter 2, both short-lived radicals
and secondary particles containing highly polar compounds are generated in the troposphere by
the same photochemical mechanisms that produce O3. It is plausible that, in addition to the
direct effects of O3, health effects are produced by ambient exposures to these gaseous and
particulate secondary compounds. Little is known regarding the composition of these reaction
products, and little research has been undertaken evaluating their toxicologic effects. Due to the
many oxidizing species present in the atmosphere, interpretation of toxicology data based on O3
exposures alone have the potential for underestimating health effects of ambient oxidant
mixtures.
5.5.1 Summary and Conclusions—Effects of Other Photochemical Oxidants
Concentrations of PAN and PPN (<1 to 10 ppb) in ambient air are unlikely to affect
pulmonary function or cause DNA damage. Levels of 10-20 ppm H2O2 can penetrate to the
lower lung directly or be transported there by PM, where inflammation can result; however,
ambient H2O2 levels are typically < ~5 ppb. As toxicology studies of other photochemical
oxidants are rare, quantitative scientific evaluations of possible health effects of environmental
exposures cannot be completed at this time.
5-79
-------
REFERENCES
Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. I; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir. Crit.
Care Med. 159: 373-382.
Adamson, I. Y. R.; Vincent, R.; Bjarnason, S. G. (1999) Cell injury and interstitial inflammation in rat lung after
inhalation of ozone and urban particulates. Am. J. Respir. Cell Mol. Biol. 20: 1067-1072.
Aizawa, H.; Shigyo, M; Nakano, H.; Matsumoto, K.; Inoue, H.; Hara, N. (1999a) Effect of the Chinese herbal
medicine, Bakumondo-to, on airway hyperresponsiveness induced by ozone exposure in guinea-pigs.
Respirology 4: 349-354.
Aizawa, H.; Shigyo, M.; Matsumoto, K.; Inoue, H.; Koto, H.; Hara, N. (1999b) PACAP reverses airway
hyperresponsiveness induced by ozone exposure in guinea pigs. Respiration 66: 538-542.
Alfaro, M. F.; Putney, L.; Tarkington, B. K.; Hatch, G. E.; Hyde, D. M.; Schelegle, E. S. (2004) Effect of rapid
shallow breathing on the distribution of 18O-labeled ozone reaction product in the respiratory tract of the rat.
Inhalation Toxicol. 16: 77-85.
Anderson, S. D.; Daviskas, E. (2000) The mechanism of exercise-induced asthma is ... J. Allergy Clin. Immunol.
106: 453-459.
Arito, H.; Takahashi, M.; Iwasaki, T.; Uchiyama, I. (1997) Age-related changes in ventilatory and heart rate
responses to acute ozone exposure in the conscious rat. Ind. Health 35: 78-86.
Arsalane, K.; Cosset, P.; Vanhee, D.; Voisin, C.; Hamid, Q.; Tonnel, A.-B.; Wallaert, B. (1995) Ozone stimulates
synthesis of inflammatory cytokines by alveolar macrophages in vitro. Am. J. Respir. Cell Mol. Biol.
13:60-68.
Avila-Costa, M. R.; Colin-Barenque, L.; Fortoul, T. L; Machado-Salas, J. P.; Espinosa-Villanueva, J.;
Rugerio-Vargas, C.; Rivas-Arancibia, S. (1999) Memory deterioration in an oxidative stress model and its
correlation with cytological changes on rat hippocampus CA1. Neurosci. Lett. 270: 107-109.
Avital, A.; Springer, C.; Bar-Yishay, E.; Godfrey, S. (1995a) Adenosine, methacholine, and exercise challenges in
children with asthma or paediatric chronic obstructive pulmonary disease. Thorax 50: 511-516.
Avital, A.; Picard, E.; Uwyyed, K.; Springer, C. (1995b) Comparison of adenosine 5'-monophosphate and
methacoline for the differentiation of asthma from chronic airway diseases with the use of the auscultative
method in very young children. J. Pediatr. 127: 438-440.
Ballinger, C. A.; Cueto, R.; Squadrito, G.; Coffin, J. F.; Velsor, L. W.; Pryor, W. A.; Postlethwait, E. M. (2005)
Antioxidant-mediated augmentation of ozone-induced membrane oxidation. Free Radical Biol. Med.
38: 515-526.
Bassett, D.; Elbon-Copp, C.; Otterbein, S.; Barraclough-Mitchell, H.; DeLorme, M.; Yang, H. (2001) Inflammatory
cell availability affects ozone-induced lung damage. J. Toxicol. Environ. Health A 64: 547-565.
Becker, S.; Quay, J.; Koren, H. S. (1991) Effect of ozone on immunoglobulin production by humanB cells in vitro.
J. Toxicol. Environ. Health 34: 353-366.
Beeson, W. L.; Abbey, D. E.; Knutsen, S. F. (1998) Long-term concentrations of ambient air pollutants and incident
lung cancer in California adults: results from the AHSMOG study. Environ. Health Perspect. 106: 813-823.
Bermudez, E. (2001) Detection of poly(ADP-ribose) synthetase activity in alveolar macrophages of rats exposed to
nitrogen dioxide and ozone. Inhalation Toxicol. 13: 69-84.
Bermudez, E.; Ferng, S. F.; Castro, C. E.; Mustafa, M. G. (1999) DNA strand breaks caused by exposure to ozone
and nitrogen dioxide. Environ. Res. 81: 72-80.
Bhalla, D. K. (1996) Alteration of alveolar macrophage chemotaxis, cell adhesion, and cell adhesion molecules
following ozone exposure of rats. J. Cell. Physiol. 169: 429-438.
Bhalla, D. K.; Gupta, S. K. (2000) Lung injury, inflammation, and inflammatory stimuli in rats exposed to ozone.
J. Toxicol. Environ. Health 59: 211-228.
Bhalla, D. K.; Gupta, S. K.; Reinhart, P. G. (1999) Alteration of epithelial integrity, alkaline phosphatase activity,
and fibronectin expression in lungs of rats exposed to ozone. J. Toxicol. Environ. Health A 56: 329-343.
Bhalla, D. K.; Reinhart, P. G.; Bai, C.; Gupta, S. K. (2002) Amelioration of ozone-induced lung injury by anti-tumor
necrosis factor-a. Toxicol. Sci. 69: 400-408.
Bignami, G.; Musi, B.; DeH'Omo, G.; Laviola, G.; Alleva, E. (1994) Limited effects of ozone exposure during
pregnancy on physical and neurobehavioral development of CD-I mice. Toxicol. Appl. Pharmacol.
129: 264-271.
Bimonte, H. A.; Nelson, M. E.; Granholm, A. C. (2003) Age-related deficits as working memory load increases:
relationships with growth factors. Neurobiol. Aging 24: 37-48.
5-80
-------
Bolarin, D. M; Bhalla, D. K.; Kleinman, M. T. (1997) Effects of repeated exposures of geriatric rats to ozone and
particle-containing atmospheres: an analysis of bronchoalveolar lavage and plasma proteins. Inhalation
Toxicol. 9: 423-434.
Bonham, A. C.; Chen, C. Y.; Mutoh, T.; Joad, J. P. (2001) Lung C-fiber CNS reflex: role in the respiratory
consequences of extended environmental tobacco smoke exposure ini young guinea pigs. Environ. Health
Perspect. 109(suppl. 4): 573-578.
Boorman, G. A.; Hailey, R.; Grumbein, S.; Chou, B. J.; Herbert, R. A.; Goehl, T.; Mellick, P. W.; Roycroft, J. H.;
Haseman, J. K.; Sills, R. (1994) Toxicology and carcinogenesis studies of ozone and ozone
4-(A^-nitrosomethylamino)-l-(3-pyridyl)-l-butanone in Fischer-344/N rats. Toxicol. Pathol. 22: 545-554.
Bornholdt, J.; Dybdahl, M.; Vogel, U.; Hansen, M.; Loft, S.; Wallin, H. (2002) Inhalation of ozone induces DNA
strand breaks and inflammation in mice. Mutat. Res. 520: 63-71.
Brannan, J. D.; Koskela, H.; Anderson, S. D.; Chew, N. (1998) Responsiveness to Mannitol in asthmatic subjects
with exercise- and hyperventiliation-induced asthma. Am. J. Respir. Crit. Care Med. 158: 1120-1126.
Bridges, J. P.; Davis, H. W.; Demodarasamy, M.; Kuroki, Y.; Howies, G.; Hui, D. Y.; McCormack, F. X. (2000)
Pulmonary surfactant proteins A and D are potent endogenous inhibitors of lipid peroxidation and oxidative
cellular injury. J. Biol. Chem. 275: 38848-38855.
Broeckaert, F.; Clippe, A.; Wattiez, R.; Falmagne, P.; Bernard, A. (2003) Lung hyperpermeability, Clara-cell
secretory potein (CC16), and susceptibility to ozone of five inbred strains of mice. Inhalation Toxicol.
15: 1209-1230.
Bush, M. L.; Zhang, W.; Ben-Jebria, A.; Ultman, J. S. (2001) Longitudinal distribution of ozone and chlorine in the
human respiratory tract: simulation of nasal and oral breathing with the single-path diffusion model. Toxicol.
Appl. Pharmacol. 173: 137-145.
Calderon-Garciduenas, L.; Osorno-Velaquez, A.; Bravo-Alvarez, H.; Delgado-Chavez, R.; Barrios-Marquez, R.
(1992) Histopathologic changes of the nasal mucosa in Southwest metropolitan Mexico City inhabitants. Am.
J. Pathol. 140: 225-232.
Calderon-Garciduenas, L.; Rodriguez-Alcaraz, A.; Garcia, R.; Sanchez, G.; Barragan, G.; Camacho, R.; Ramirez, L.
(1994) Human nasal mucosal changes after exposure to urban pollution. Environ. Health Perspect.
102: 1074-1080.
Calderon-Garciduenas, L.; Osnaya-Brizuela, N.; Ramirez-Martinez, L.; Villarreal-Calderon, A. (1996) DNA strand
breaks in human nasal respiratory epithelium are induced upon exposure to urban pollution. Environ. Health
Perspect. 104: 160-168.
Calderon-Garciduenas, L.; Wen-Wang, L.; Zhang, Y.-J.; Rodriguez-Alcaraz, A.; Osnaya, N.; Villarreal-Calderon,
A.; Santella, R. M. (1999) 8-hydroxy-2'-deoxyguanosine, a major mutagenic oxidative DNA lesion, and DNA
strand breaks in nasal respiratory epithelium of children exposed to urban pollution. Environ. Health Perspect.
107: 469-474.
Calderon-Garciduenas, L.; Devlin, R. B.; Miller, F. J. (2000a) Respiratory tract pathology and cytokine imbalance in
clinically healthy children chronically and sequentially exposed to air pollutants. Med. Hypotheses
55: 373-378.
Calderon-Garciduenas, L.; Mora-Tiscareno, A.; Chung, C. J.; Valencia, G.; Fordham, L. A.; Garcia, R.; Osnaya, N.;
Romero, L.; Acuna, H.; Villarreal-Calderon, A. (2000b) Exposure to air pollution is associated with lung
hyperinflation in healthy children and adolescents in southwest Mexico City: a pilot study. Inhalation Toxicol.
12: 537-561.
Calderon-Garciduenas, L.; Rodrigues-Alcaraz, A.; Valencia-Salazar, G.; Mora-Tascareno, A.; Garcia, R.; Osnaya,
N.; Villarreal-Calderon, A.; Devlin, R. B.; Van Dyke, T. L. (2001) Nasal biopsies of children exposed to air
pollutants. Toxicol. Pathol. 29: 558-564.
Cassee, F. R.; Feron, V. J. (1994) Biochemical and histopathological changes in nasal epithelium of rats after 3-day
intermittent exposure to formaldehyde and ozone alone or in combination. Toxicol. Lett. 72: 257-268.
Chang, M. M.-J.; Wu, R.; Plopper, C. G.; Hyde, D. M. (1998) IL-8 is one of the major chemokines produced by
monkey airway epithelium after ozone-induced injury. Am. J. Physiol. 275: L524-L532.
Cheek, J. M.; McDonald, R. J.; Rapalyea, L.; Tarkington, B. K.; Hyde, D. M. (1995) Neutrophils enhance removal
of ozone-injured alveolar epithelial cells in vitro. Am. J. Physiol. 269: L527-L535.
Chen, L. C.; Qu, Q.; Amdur, M. O.; Schlesinger, R. B. (1995) Alteration of pulmonary macrophage intracellular pH
following inhalation exposure to sulfuric acid/ozone mixtures. Exp. Lung Res. 21: 113-128.
Chen, L.; Yang, W.; Jennison, B. L.; Goodrich, A.; Omaye, S. T. (2002) Air pollution and birth weight in northern
Nevada, 1991-1999. Inhalation Toxicol. 14: 141-157.
5-81
-------
Chen, C.-Y.; Bonham, A. C.; Plopper, C. G.; Joad, J. P. (2003) Plasticity in respiratory motor control: selected
contribution: neuroplasticity in nucleus tractus solitarius neurons following episodic ozone exposure in infant
primates. J. Appl. Physiol. 94: 819-827.
Cho, H. Y.; Hotchkiss, J. A.; Harkema, J. R. (1999a) Inflammatory and epithelial responses during the development
of ozone-induced mucous cell metaplasia in the nasal epithelium of rats. Toxicol. Sci. 51: 135-145.
Cho, H. Y.; Hotchkiss, J. A.; Bennett, C. B.; Harkema, J. R. (1999b) Effects of pre-existing rhinitis on
ozone-induced mucous cell metaplasia in rat nasal epithelium. Toxicol. Appl. Pharmacol. 158: 92-102.
Cho, H. Y.; Hotchkiss, J. A.; Bennett, C. B.; Harkema, J. R. (2000) Neutrophil-dependent and
neutrophil-independent alterations in the nasal epithelium of ozone-exposed rats. Am. J. Respir. Crit. Care
Med. 162: 629-636.
Cho, H.-Y.; Zhang, L.-Y.; Kleeberger, S. R. (2001) Ozone-induced lung inflammation and hyperreactivity are
mediated via tumor necrosis factor-a receptors. Am. J. Physiol. 280: L537-L546.
Churg, A.; Brauer, M; Keeling, B. (1996) Ozone enhances the uptake of mineral particles by tracheobronchial
epithelial cells in organ culture. Am. J. Respir. Crit. Care Med. 153: 1230-1233.
Clausen, P. A.; Wilkins, C. K.; Wolkoff, P.; Nielsen, G. D. (2001) Chemical and biological evaluation of a reaction
mixture of R-(+)-limonene/ozone: formation of strong airway irritants. Environ. Int. 26: 511-522.
Cohen, M. D.; Zelikoff, J. T.; Qu, Q.; Schlesinger, R. B. (1996) Effects of ozone upon macrophage-interferon
interactions. Toxicology 114: 243-252.
Cohen, M. D.; Sisco, M.; Li, Y.; Zelikoff, J. T.; Schlesinger, R. B. (2001) Ozone-induced modulation of
cell-mediated immune responses in the lungs. Toxicol. Appl. Pharmacol. 171: 71-84.
Cohen, M. D.; Sisco, M.; Baker, K.; Li, Y.; Lawrence, D.; VanLoveren, H.; Zelikoff, J. T.; Schlesinger, R. B.
(2002) Effects of inhaled ozone on pulmonary immune cells critical to antibacterial responses in situ.
Inhalation Toxicol. 14: 599-619.
Colin-Barenque, L.; Avila-Costa, M. R.; Fortoul, T.; Rugerio-Vargas, C.; Machado-Salas, J. P.;
Espinosa-Villanueva, J.; Rivas-Arancibia, S. (1999) Morphologic alteration of the olfactory bulb after acute
ozone exposure in rats. Neurosci. Lett. 274: 1-4.
Connor, L. M.; Ballinger, C. A.; Albrecht, T. B.; Postlethwait, E. M. (2004) Interfacial phospholipids inhibit ozone
reactive absorption-mediated cytotoxicity in vitro. Am. J. Physiol. 286: L1169-L1178.
Cottet-Emard, J.-M.; Dalmaz, Y.; Pequignot, J.; Peyrin, L.; Pequignot, J.-M. (1997) Long-term exposure to ozone
alters peripheral and central catecholamine activity in rats. Pfluegers Arch. 433: 744-749.
Creutzenberg, O.; Bellmann, B.; Klingebiel, R.; Heinrich, U.; Muhle, H. (1995) Phagocytosis and chemotaxis of rat
alveolar macrophages after a combined or separate exposure to ozone and carbon black. Exp. Toxicol. Pathol.
47: 202-206.
Cross, C. E.; VanDer Vliet, A.; Louie, S.; Thiele, J. J.; Halliwell, B. (1998) Oxidative stress and antioxidants at
biosurfaces: plants, skin, and respiratory tract surfaces. Environ. Health Perspect. 106(suppl. 5): 1241-1251.
Custodio-Ramirez, V.; Paz, C. (1997) Ozone produces functional deficits in the rat visual pathway.
Electroencephalogr. Clin. Neurophysiol. 104: 269-273.
Daly, C.; Fox, K.; Henein, M. (2002) Natriuretic peptides in the diagnosis of heart disease—first amongst equals?
Int. J. Cardiol. 84: 107-113.
Delaunois, A.; Segura, P.; Montano, L. M.; Vargas, M. H.; Ansay, M.; Gustin, P. (1998) Comparison of
ozone-induced effects on lung mechanics and hemodynamics in the rabbit. Toxicol. Appl. Pharmacol.
150: 58-67.
DeH'Omo, G.; Fiore, M.; Petruzzi, S.; Alleva, E.; Bignami, G. (1995a) Neurobehavioral development of CD-I mice
after combined gestational and postnatal exposure to ozone. Arch. Toxicol. 69: 608-616.
DeH'Omo, G.; Wolfer, D.; Alleva, E.; Lipp, H.-P. (1995b) Developmental exposure to ozone induces subtle changes
in swimming navigation of adult mice. Toxicol. Lett. 81: 91-99.
DeMarini, D. M.; Shelton, M. L.; Kohan, M. J.; Hudgens, E. E.; Kleindienst, T. E.; Ball, L. M.; Walsh, D.; de Boer,
J. G.; Lewis-Bevan, L.; Rabinowitz, J. R.; Claxton, L. D.; Lewtas, J. (2000) Mutagenicity in lung of Big
Blue(R) mice and induction of tandem-base substitutions in Salmonella by the air pollutant peroxyacetyl
nitrate (PAN): predicted formation of intrastrand cross-links. Mutat. Res. 457: 41-55.
Depuydt, P.; Joos, G. F.; Pauwels, R. A. (1999) Ambient ozone concentrations induce airway hyperresponsiveness
in some rat strains. Eur. Respir. J. 14: 125-131.
Depuydt, P. O.; Lambrecht, B. N.; Joos, G. F.; Pauwels, R. A. (2002) Effect of ozone exposure on allergic
sensitization and airway inflammation induced by dendritic cells. Clin. Exp. Allergy 32: 391-396.
5-82
-------
Dorado-Martinez, C.; Parades-Carbajal, C.; Mascher, D.; Borgonio-Perez, G.; Rivas-Arancibia, S. (2001) Effects of
different ozone doses on memory, motor activity and lipid peroxidation levels, in rats. Int. J. Neurosci.
108: 149-161.
Dormans, J. A. M. A.; Boere, A. J. F.; vanLoveren, H.; Rombout, P. J. A.; Marra, M.; vanBree, L. (1996)
Age-related toxicity in rat lungs following acute and repeated ozone exposure. Inhalation Toxicol. 8: 903-925.
Dormans, J. A. M. A.; VanBree, L.; Boere, A. J. F.; Marra, M.; Rombout, P. J. A. (1999) Interspecies differences in
time course of pulmonary toxicity following repeated exposure to ozone. Inhalation Toxicol. 11: 309-329.
Driscoll, K. E.; Simpson, L.; Carter, J.; Hassenbein, D.; Leikauf, G. D. (1993) Ozone inhalation stimulates
expression of a neutrophil chemotactic protein, macrophage inflammatory protein 2. Toxicol. Appl.
Pharmacol. 119:306-309.
Dungworth, D. L. (1989) Noncarcinogenic responses of the respiratory tract to inhaled toxicants. In: McClellan,
R. O.; Henderson, R. F., eds. Concepts in inhalation toxicology. New York, NY: Hemisphere Publishing
Corp.; pp. 273-298.
Dye, J. A.; Madden, M. C.; Richards, J. H.; Lehmann, J. R.; Devlin, R. B.; Costa, D. L. (1999) Ozone effects on
airway responsiveness, lung injury, and inflammation. Comparative rat strain and in vivo/in vitro
investigations. Inhalation Toxicol. 11: 1015-1040.
El-Fawal, H. A. N.; McGovern, T.; Schlesinger, R. B. (1995) Nonspecific bronchial responsiveness assessed in vitro
following acute inhalation exposure to ozone and ozone/sulfuric acid mixtures. Exp. Lung Res. 21: 129-139.
Elder, A. C. P.; Gelein, R.; Finkelstein J. N.; Cox, C.; Oberdorster, G. (2000a) Endotoxin priming affects the lung
response to ultrafine particles and ozone in young and old rats. In: Phalen, R. F., ed. Inhalation toxicology:
proceedings of the third colloquium on paniculate air pollution and human health (first special issue); June,
1999; Durham, NC. Inhalation Toxicol. 12(suppl. 1): 85-98.
Elder, A. C. P.; Gelein, R.; Finkelstein, J. N.; Cox, C.; Oberdorster, G. (2000b) Pulmonary inflammatory response to
inhaled ultrafine particles is modified by age, ozone exposure, and bacterial toxin. In: Grant, L. D., ed.
PM2000: paniculate matter and health. Inhalation Toxicol. 12(suppl. 4): 227-246.
Elsayed, N. M. (2001) Diet restriction modulates lung response and survivability of rats exposed to ozone.
Toxicology 159: 171-182.
Evans, M. J.; Fanucchi, M. V.; Baker, G. L.; Van Winkle, L. S.; Pantle, L. M.; Nishio, S. J.; Schelegle, E. S.;
Gershwhin, L. J.; Miller, L. A.; Hyde, D. M.; Sannes, P. L.; Plopper, C. G. (2003) Atypical development of
the trachea! basement membrane zone of infant rhesus monkeys exposed to ozone and allergen. Am. J.
Physiol. 285:L931-L939.
Fakhrzadeh, L.; Laskin, J. D.; Laskin, D. L. (2002) Deficiency in inducible nitric oxide synthase protects mice from
ozone-induced lung inflammation and tissue injury. Am. J. Respir. Cell Mol. Biol. 26: 413-419.
Fanucchi, M. V.; Hotchkiss, J. A.; Harkema, J. R. (1998) Endotoxin potentiates ozone-induced mucous cell
metaplasia in rat nasal epithelium. Toxicol. Appl. Pharmacol. 152: 1-9.
Farman, C. A.; Pinkerton, K. E.; Rajini, P.; Witschi, H.; Last, J. A. (1997) Evolution of lung lesions in rats exposed
to mixtures of ozone and nitrogen dioxide. Inhalation Toxicol. 9: 647-677'.
Farman, C. A.; Watkins, K.; Van Hoozen, B.; Last, J. A.; Witschi, H.; Pinkerton, K. E. (1999) Centriacinar
remodeling and sustained procollagen gene expression after exposure to ozone and nitrogen dioxide. Am. J.
Respir. Cell Mol. Biol. 20: 303-311.
Ferng, S.-F.; Castro, C. E.; Afifi, A. A.; Bermudez, E.; Mustafa, M. G. (1997) Ozone-induced DNA strand breaks in
guinea pig tracheobronchial epithelial cells. J. Toxicol. Environ. Health 51: 353-367.
Folkerts, G.; Busse, W. W.; Nijkamp, F. P.; Sorkness, R.; Gern, J. E. (1998) Virus-induced airway
hyperresponsiveness and asthma. Am. J. Respir. Crit. Care Med. 157: 1708-1720.
Foster, W. M.; Freed, A. N. (1999) Regional clearance of solute from peripheral airway epithelia: recovery after
sublobar exposure to ozone. J. Appl. Physiol. 86: 641-646.
Frampton, M. W.; Pryor, W. A.; Cueto, R.; Cox, C.; Morrow, P. E.; Utell, M. J. (1999) Aldehydes (nonanal and
hexanal) in rat and human bronchoalveolar lavage fluid after ozone exposure. Cambridge, MA: Health Effects
Institute; research report no. 90. Available: www.healtheffects.org/Pubs/Frampton-C.pdf [2000, February 9].
Freed, A. N.; Chou, C. L.; Fuller, S. D.; Croxton, T. L. (1996) Ozone-induced vagal reflex modulates airways
reactivity in rabbits. Respir. Physiol. 105: 95-102.
Freed, A. N.; Cueto, R.; Pryor, W. A. (1999) Antioxidant transport modulates peripheral airway reactivity and
inflammation during ozone exposure. J. Appl. Physiol. 87: 1595-1603.
Friedlander, S. K.; Yeh, E. K. (1998) The submicron atmospheric aerosol as a carrier of reactive chemical species:
case of peroxides. Appl. Occup. Environ. Hyg. 13: 416-420.
5-83
-------
Gaffney, J. S.; Marley, N. A.; Prestbo, E. W. (1993) Measurements of peroxyacetyl nitrate at a remote site in the
southwestern United States: tropospheric implications. Environ. Sci. Technol. 27: 1905-1910.
Garssen, J.; Van Bree, L.; Van Der Vliet, H.; Van Loveren, H. (1997) Ozone-induced impairment of pulmonary type
IV hypersensitivity and airway hyperresponsiveness in mice. Inhalation Toxicol. 9: 581-599.
Gohil, K.; Cross, C. E.; Last, J. A. (2003) Ozone-induced disruptions of lung transcriptomes. Biochem. Biophys.
Res. Commun. 305: 719-728.
Goldsmith, C.-A. W.; Ning, Y.-Y.; Qin, G.; Imrich, A.; Lawrence, J.; Murthy, G. G., K.; Catalano, P. J.; Kobzik, L.
(2002) Combined air pollution particle and ozone exposure increases airway responsiveness in mice.
Inhalation Toxicol. 14: 325-347.
Gonzalez-Pina, R.; Paz, C. (1997) Brain monoamine changes in rats after short periods of ozone exposure.
Neurochem. Res. 22: 63-66.
Grosjean, D. (2003) Ambient PAN and PPN in southern California from 1960 to the SCOS97-NARSTO. Atmos.
Environ. 37(suppl. 2): S221-S238.
Grosjean, E.; Grosjean, D.; Woodhouse, L. F. (2001) Peroxyacetyl nitrate and peroxypropionyl nitrate during SCOS
97-NARSTO. Environ. Sci. Technol. 35: 4007-4014.
Guerrero, A. L.; Dorado-Martinez, C.; Rodriguez, A.; Pedroza-Rios, K.; Borgonio-Perez, G.; Rivas-Arancibia, S.
(1999) Effects of vitamin E on ozone-induced memory deficits and lipid peroxidation in rats. NeuroReport
10: 1689-1692.
Gulisano, M.; Marceddu, S.; Barbara, A.; Pacini, A.; Buiatti, E.; Martini, A.; Pacini, P. (1997) Damage to the
nasopharyngeal mucosa induced by current levels of urban air pollution: a field study in lambs. Eur. Respir. J.
10: 567-572.
Gtinther, T.; Hollriegl, V.; Vormann, J. (1993) Perinatal development of iron and antioxidant defence systems.
J. Trace Elem. Electrolytes Health Dis. 7: 47-52.
Gupta, S. K.; Reinhart, P. G.; Bhalla, D. K. (1998) Enhancement of fibronectin expression in rat lung by ozone and
an inflammatory stimulus. Am. J. Physiol. 275: L330-L335.
Haddad, E.-B.; Liu, S. F.; Salmon, M.; Robichaud, A.; Barnes, P. J.; Chung, K. F. (1995) Expression of inducible
nitric oxide synthase mRNA in Brown Norway rats exposed to ozone: effect of dexamethasone. Eur. J.
Pharmacol. Environ. Toxicol. Pharmacol. Sect. 293: 287-290.
Hamilton, R. F.; Li, L.; Eschenbacher, W. L.; Szweda, L.; Holian, A. (1998) Potential involvement of
4-hydroxynonenal in the response of human lung cells to ozone. Am. J. Physiol. 274: L8-L16.
Harkema, J. R.; Plopper, C. G.; Hyde, D. M.; St. George, J. A.; Wilson, D. W.; Dungworth, D. L. (1987) Response
of the macaque nasal epithelium to ambient levels of ozone: a morphologic and morphometric study of the
transitional and respiratory epithelium. Am. J. Pathol. 128: 29-44.
Harkema, J. R.; Morgan, K. T.; Gross, E. A.; Catalano, P. J.; Griffith, W. C. (1994) Consequences of prolonged
inhalation of ozone on F344/N rats: collaborative studies. Part VII: effects on the nasal mucociliary apparatus.
Cambridge, MA: Health Effects Institute; research report no. 65.
Harkema, J. R.; Catalano, P. J.; Hotchkiss, J. A. (1997a) Consequences of prolonged inhalation of ozone onF344/N
rats: collaborative studies. Part XII. Atrophy of bone in nasal turbinates. Cambridge, MA: Health Effects
Institute; research report no. 65.
Harkema, J. R.; Hotchkiss, J. A.; Griffith, W. C. (1997b) Mucous cell metaplasia in rat nasal epithelium after a
20-month exposure to ozone: a morphometric study of epithelial differentiation. Am. J. Respir. Cell Mol.
Biol. 16: 521-530.
Harkema, J. R.; Hotchkiss, J. A.; Barr, E. B.; Bennett, C. B.; Gallup, M.; Lee, J. K.; Basbaum, C. (1999)
Long-lasting effects of chronic ozone exposure on rat nasal epithelium. Am. J. Respir. Cell Mol. Biol.
20: 517-529.
Kara, R.; Paz, C. (1993) Effects of ozone exposure during pregnancy on ontogeny of sleep in rats. Neurosci. Lett.
164: 67-70.
Hawgood, S.; Poulain, F. R. (2001) The pulmonary collectins and surfactant metabolism. Annu. Rev. Physiol.
63:495-519.
Hawgood, S.; Ochs, M.; Jung, A.; Akiyama, J.; Allen, L.; Brown, C.; Edmondson, J.; Levitt, S.; Carlson, E.;
Gillespie, A. M.; Villar, A.; Epstein, C. J.; Poulain, F. R. (2002) Sequential targeted deficiency of SP-A and
-D leads to progressive alveolar lipoproteinosis and emphysema. Am. J. Physiol. 283: L1002-L1010.
Heck, H. d'A.; Casanova, M.; Starr, T. B. (1990) Formaldehyde toxicity-new understanding. Crit. Rev. Toxicol.
20: 397-426.
Heddle, J. A.; Shepson, P. B.; Gingerich, J. D.; So, K. W. (1993) Mutagenicity of peroxyacetyl nitrate (PAN) in
vivo: tests for somatic mutations and chromosomal aberrations. Environ. Mol. Mutagen. 21: 58-66.
5-84
-------
Herbert, R. A.; Hailey, J. R.; Grumbein, S.; Chou, B. I; Sills, R. C.; Haseman, J. K.; Goehl, T.; Miller, R. A.;
Roycroft, J. H.; Boorman, G. A. (1996) Two-year and lifetime toxicity and carcinogenicity studies of ozone in
B6C3F1 mice. Toxicol. Pathol. 24: 539-548.
Highfill, J. W.; Watkinson, W. P. (1996) Ozone toxicity in the rat. II. Modeling changes due to ambient temperatures
and duration. J. Appl. Physiol. 80: 1811-1818.
Hoffer, E.; Baum, Y.; Tabak, A.; Frevert, C. (1999) Adhesion molecules of blood polymorphonuclear leukocytes
and alveolar macrophages in rats: modulation by exposure to ozone. Hum. Exp. Toxicol. 18: 547-551.
Hotchkiss, J. A.; Harkema, J. R.; Johnson, N. F. (1997) Kinetics of nasal epithelial cell loss and proliferation in F344
rats following a single exposure to 0.5 ppm ozone. Toxicol. Appl. Pharmacol. 143: 75-82.
Hotchkiss, J. A.; Hilaski, R.; Cho, H.; Regan, K.; Spencer, P.; Slack, K.; Harkema, J. R. (1998) Fluticasone
propionate attenuates ozone-induced rhinitis and mucous cell metaplasia in rat nasal airway epithelium.
Am. J. Respir. CellMol. Biol. 18: 91-99.
Huffman, L. J.; Judy, D. J.; Brumbaugh, K.; Frazer, D. G.; Reynolds, J. S.; McKinney, W. G.; Goldsmith, W. T.
(2001) Hyperthyroidism increases the risk of ozone-induced lung toxicity in rats. Toxicol. Appl. Pharmacol.
173: 18-26.
Huitron-Resendiz, S.; Custodio-Ramirez, V.; Escalante-Membrillo, C.; Gonzalez-Pina, R.; Paz, C. (1994) Sleep
alterations and brain regional changes of serotonin and its metabolite in rats exposed to ozone. Neurosci. Lett.
177: 119-122.
Hyde, D. M.; Miller, L. A.; McDonald, R. J.; Stovall, M. Y.; Wong, V.; Pinkerton, K. E.; Wegner, C. D.; Rothlein,
R.; Plopper, C. G. (1999) Neutrophils enhance clearance of necrotic epithelial cells in ozone-induced lung
injury in rhesus monkeys. Am. J. Physiol. 277: L1190-L1198.
Igarashi, A.; lijima, H.; Tamura, G.; Shirato, K. (1998) Tazanolast inhibits ozone-induced airway
hyperresponsiveness in guinea pigs. Am. J. Respir. Crit. Care Med. 157: 1531-1535.
lijima, M. K.; Kobayashi, T.; Kamada, H.; Shimojo, N. (2001) Exposure to ozone aggravates nasal allergy-like
symptoms in guinea pigs. Toxicol. Lett. 123: 77-85.
Ishii, Y.; Yang, H.; Sakamoto, T.; Nomura, A.; Hasegawa, S.; Hirata, F.; Bassett, D. J. P. (1997) Rat alveolar
macrophage cytokine production and regulation of neutrophil recruitment following acute ozone exposure.
Toxicol. Appl. Pharmacol. 147: 214-223.
Ishii, Y.; Hashimoto, K.; Hirano, K.; Morishima, Y.; Mochizuki, M.; Masuyama, K.; Nomura, A.; Sakamoto, T.;
Uchida, Y.; Sagai, M.; Sekizawa, K. (2000) Ebselen decreases ozone-induced pulmonary inflammation in
rats. Lung 178: 225-234.
Iwasaki, T.; Takahashi, M.; Saito, H.; Arito, H. (1998) Adaptation of extrapulmonary responses to ozone exposure in
conscious rats. Ind. Health 36: 57-60.
Jakobi, G.; Fabian, P. (1997) Indoor/outdoor concentrations of ozone and peroxyacetyl nitrate (PAN). Int. J.
Biometeorol. 40: 162-165.
Jang, A.-S.; Choi, I.-S.; Koh, Y.-I.; Park, C.-S.; Lee, J.-S. (2002) The relationship between alveolar epithelial
proliferation and airway obstruction after ozone exposure. Allergy 57: 737-740.
Jimba, M.; Skornik, W. A.; Killingsworth, C. R.; Long, N. C.; Brain, J. D.; Shore, S. A. (1995) Role of C fibers in
physiological responses to ozone in rats. J. Appl. Physiol. 78: 1757-1763.
Joad, J. P.; Kott, K. S.; Bonham, A. C. (1998) Exposing guinea pigs to ozone for 1 wk enhances responsiveness of
rapidly adapting receptors. J. Appl. Physiol. 84: 1190-1197.
Joad, J. P.; Brie, J. M.; Weir, A. J.; Putney, L.; Hyde, D. M.; Postlewait, E. M.; Plopper, C. G. (2000) Effect of
respiratory pattern on ozone injury to the airways of isolated rat lungs. Toxicol. Appl. Pharmacol. 169: 26-32.
Johnston, C. J.; Stripp, B. R.; Piedbeouf, B.; Wright, T. W.; Mango, G. W.; Reed, C. K.; Finkelstein, J. N. (1998)
Inflammatory and epithelial responses in mouse strains that differ in sensitivity to hyperoxic injury. Exp.
Lung Res. 24: 189-202.
Johnston, C. J.; Stripp, B. R.; Reynolds, S. D.; Avissar, N. E.; Reed, C. K.; Finkelstein, J. N. (1999a) Inflammatory
and antioxidant gene expression in C57BL/6J mice after lethal and sublethal ozone exposures. Exp. Lung
Res. 25: 81-97.
Johnston, C. J.; Finkelstein, J. N.; Oberdorster, G.; Reynolds, S. D.; Stripp, B. R. (1999b) Clara cell secretory
protein-deficient mice differ from wild-type mice in inflammatory chemokine expression to oxygen and
ozone, but not to endotoxin. Exp. Lung Res. 25: 7-21.
Johnston, C. J.; Oberdorster, G.; Gelein, R.; Finkelstein, J. N. (2000) Newborn mice differ from adult mice in
chemokine and cytokine expression to ozone, but not to endotoxin. Inhalation Toxicol. 12: 205-224.
Johnston, C. J.; Oberdorster, G.; Gelein, R.; Finkelstein, J. N. (2002) Endotoxin potentiates ozone-induced
pulmonary chemokine and inflammatory responses. Exp. Lung Res. 28: 419-433.
5-85
-------
Kafoury, R. M; Pryor, W. A.; Squadrito, G. L.; Salgo, M. G.; Zou, X.; Friedman, M. (1999) Induction of
inflammatory mediators in human airway epithelial cells by lipid ozonation products. Am. J. Respir. Crit.
Care Med. 160: 1934-1942.
Kenyon, N. J.; VanDer Vliet, A.; Schock, B. C.; Okamoto, T.; McGrew, G. M.; Last, J. A. (2002) Susceptibility to
ozone-induced acute lung injury in iNOS-deficient mice. Am. J. Physiol. 282: L540-L545.
Kim, M. Y.; Son, J. W.; Cho, M. H.; Choi, C. S.; Chae, C. H.; Lee, M. H. (2001) Oviductural carcinoma inB6C3Fl
female mice exposed to 0.5 ppm ozone. Vet. Hum. Toxicol. 43: 370-372.
Kimmel, T. A.; Chen, L. C.; Bosland, M. C.; Nadziejko, C. (1997) Influence of acid aerosol droplet size on
structural changes in the rat lung caused by acute exposure to sulfuric acid and ozone. Toxicol. Appl.
Pharmacol. 144: 348-355.
Kleeberger, S. R.; Levitt, R. C.; Zhang, L.-Y.; Longphre, M.; Harkema, J.; Jedlicka, A.; Eleff, S. M.; DiSilvestre, D.;
Holroyd, K. J. (1997) Linkage analysis of susceptibility to ozone-induced lung inflammation in inbred mice.
Nat. Genet. 17: 475-478.
Kleeberger, S. R.; Reddy, S.; Zhang, L.-Y.; Jedlicka, A. E. (2000) Genetic susceptibility to ozone-induced lung
hyperpermeability: role of toll-like receptor 4. Am. J. Respir. CellMol. Biol. 22: 620-627.
Kleeberger, S. R.; Reddy, S. P.; Zhang, L.-Y.; Cho, H.-Y.; Jedlicka, A. E. (2001a) Toll-like receptor 4 mediates
ozone-induced murine lung hyperpermeability via inducible nitric oxide synthase. Am. J. Physiol.
280: L326-L333.
Kleeberger, S. R.; Ohtsuka, Y.; Ahang, L.-Y.; Longphre, M. (200Ib) Airway responses to chronic ozone exposure
are partially mediated through mast cells. J. Appl. Physiol. 90: 713-723.
Kleindienst, T. E.; Shepson, P. B.; Smith, D. F.; Hudgens, E. E.; Nero, C. M.; Cupitt, L. T.; Bufalini, J. J.; Claxton,
L. D. (1990) Comparison of mutagenic activities of several peroxyacyl nitrates. Environ. Mol. Mutagen.
16: 70-80.
Kleinman, M. T.; Mautz, W. J.; Bjarnason, S. (1999) Adaptive and non-adaptive responses in rats exposed to ozone,
alone and in mixtures, with acidic aerosols. Inhalation Toxicol. 11: 249-264.
Kleinman, M. T.; Bufalino, C.; Rasmussen, R.; Hyde, D.; Bhalla, D. K.; Mautz, W. J. (2000) Toxicity of chemical
components of ambient fine paniculate matter (PM25) inhaled by aged rats. J. Appl. Toxicol. 20: 357-364.
Kligerman, A. D.; Mottus, K.; Erexson, G. L. (1995) Cytogenetic analyses of the in vitro and in vivo responses of
murine cells to peroxyacetyl nitrate (PAN). Mutat. Res. 341: 199-206.
Kobzik, L.; Goldsmith, C.-A. W.; Ning, Y. Y.; Qin, G.; Morgan, B.; Imrich, A.; Lawrence J.; Murthy, G. G. K.;
Catalano, P. J. (2001) Effects of combined ozone and air pollution particle exposure in mice. Boston, MA:
Health Effects Institute; research report no. 106. Available: http://www.healtheffects.org/Pubs/Kobzik.pdf
[27 January 2003].
Kodavanti, U. P.; Hatch, G. E.; Starcher, B.; Giri, S. N.; Winsett, D.; Costa, D. L. (1995) Ozone-induced pulmonary
functional, pathological, and biochemical changes in normal and vitamin C-deficient guinea pigs. Fundam.
Appl. Toxicol. 24: 154-164.
Koto, H.; Salmon, M.; Haddad el-B.; Huang, T.-J.; Zagorski, J.; Chung, K. F. (1997) Role of cytokine-induced
neutrophil chemoattractant (CINC) in ozone-induced airway inflammation and hyperresponsiveness. Am. J.
Respir. Crit. Care Med. 156: 234-239.
Koyama, Y.; Hayaishi, O. (1994) Modulation by prostaglandins of activity of sleep-related neurons in the
preoptic/anterior hypothalamic areas in rats. Brain Res. Bull. 33: 367-372.
Larson, S. D.; Schelegle, E. S.; Walby, W. F.; Gershwin, L. J.; Fanuccihi, M. V.; Evans, M. J.; Joad, J. P.;
Tarkington, B. K.; Hyde, D. M.; Plopper, C. G. (2004) Postnatal remodeling of the neural components of the
epithelial-mesenchymal trophic unit in the proximal airways of infant rhesus monkeys exposed to ozone and
allergen. Toxicol. Appl. Pharmacol. 194: 211-220.
Laskin, D. L.; Laskin, J. D. (2001) Role of macrophages and inflammatory mediators in chemically induced toxicity.
Toxicology (Ireland) 160: 111-118.
Laskin, D. L.; Pendino, K. J.; Punjabi, C. J.; del Valle, M. R.; Laskin, J. D. (1994) Pulmonary and hepatic effects of
inhaled ozone in rats. Environ. Health Perspect. 102(suppl. 10): 61-64.
Laskin, J. D.; Heck, D. E.; Laskin, D. L. (1996) Nitric oxide production in the lung and liver following inhalation of
the pulmonary irritant ozone. In: Snyder, R.; Kocsis, J. J.; Sipes, I. G.; Kalf, G. F.; Jollow, D. J.; Greim, H.;
Monks, T. J.; Witmer, C. M., eds. Biological Reactive Intermediates V: Basic Mechanistic Research in
Toxicology and Human Risk Assessment: proceedings of the Fifth International Symposium; January 1995;
Munich, Germany. Adv. Exp. Med. Biol. 387: 141-146.
Laskin, D. L.; Sunil, V.; Guo, Y.; Heck, D. E.; Laskin, J. D. (1998) Increased nitric oxide synthase in the lung after
ozone inhalation is associated with activation of NF-KB. Environ. Health Perspect. 106(suppl. 5): 1175-1178.
5-86
-------
Laskin, D. L.; Fakhrzadeh, L.; Heck, D. E.; Gerecke, D.; Laskin, J. D. (2002) Upregulation of phosphoinositide
3-kinase and protein kinase B in alveolar macrophages following ozone inhalation. Role of NF-KB and
STAT-1 in ozone-induced nitric oxide production and toxicity. Mol. Cell. Biochem. 234-235: 91-98.
Last, J. A.; Pinkerton, K. E. (1997) Chronic exposure of rats to ozone and sulfuric acid aerosol: biochemical and
structural responses. Toxicology 116: 133-146.
Lavnikova, N.; Prokhorova, S.; Lakhotia, A. V.; Gordon, R.; Laskin, D. L. (1998) Distinct inflammatory responses
of adherent vascular lung neutrophils to pulmonary irritants. J. Inflammation 48: 56-66
Lee, C.; Watt, K. C.; Chang, A. M; Plopper, C. G.; Buckpitt, A. R.; Pinkerton, K. E. (1998) Site-selective
differences in cytochrome P450 isoform activities: comparison of expression in rat and rhesus monkey lung
and induction in rats. Drug Metab. Dispos. 26: 396-400.
Lemos, M.; Lichtenfels, A. J. F. C.; Amaro, E., Jr.; Macchione, M.; Martins, M. A.; King, M.; Bohm, G. M.;
Saldiva, P. H. N. (1994) Quantitative pathology of nasal passages in rats exposed to urban levels of air
pollution. Environ. Res. 66: 87-95.
Longphre, M.; Zhang, L.-Y.; Harkema, J. R.; Kleeberger, S. R. (1999) Ozone-induced pulmonary inflammation and
epithelial proliferation are partially mediated by PAF. J. Appl. Physiol. 86: 341-349.
Lorz, C.; Lopez, J. (1997) Incidence of air pollution in the pulmonary surfactant system of the pigeon (Columbia
livid). Anat. Rec. 249: 206-212.
Madden, M. C.; Richards, J. H.; Dailey, L. A.; Hatch, G. E.; Ohio, A. J. (2000) Effect of ozone on diesel exhaust
particle toxicity in rat lung. Toxicol. Appl. Pharmacol. 168: 140-148.
Mango, G. W.; Johnston, C. J.; Reynolds, S. D.; Finkelstein, J. N.; Plopper, C. G.; Stripp, B. R. (1998) Clara cell
secretory protein deficiency increases oxidant stress response in conducting airways. Am. J. Physiol.
275: L348-L356.
Matsubara, S.; Kikkawa, H.; Kaminuma, O.; Ikezawa, K. (1997a) Angiotensin-converting enzyme inhibitors can
potentiate ozone-induced airway hyperresponsiveness. Eur. J. Pharmacol. 337: 259-265.
Matsubara, S.; Fushimi, K.; Kaminuma, O.; Kikkawa, H.; Ikezawa, K.; Naito, K. (1997b) Prevention of
ozone-induced airway hyperresponsiveness and epithelial injury by phosphodiesterase inhibitors in guinea
pigs. Environ. Toxicol. Pharmacol. 3: 201-209.
Mautz, W. J. (2003) Exercising animal models in inhalation toxicology: interactions with ozone and formaldehyde.
Environ. Res. 92: 14-26.
Mautz, W. J.; Kleinman, M. T.; Bhalla, D. K.; Phalen, R. F. (2001) Respiratory tract responses to repeated inhalation
of an oxidant and acid gas-particle air pollutant mixture. Toxicol. Sci. 61: 331-341.
McGraw, D. W.; Forbes, S. L.; Mak, J. C. W.; Witte, D. P.; Carrigan, P. E.; Leikauf, G. D.; Liggett, S. B. (2000)
Transgenic overexpression of p2-adrenergic receptors in airway epithelial cells decreases bronchoconstriction.
Am. J. Physiol. 279: L379-L389.
McKinney, W. J.; Jaskot, R. H.; Richards, J. H.; Costa, D. L.; Dreher, K. L. (1998) Cytokine mediation of
ozone-induced pulmonary adaptation. Am. J. Respir. Cell. Mol. Biol. 18: 696-705.
Miller, F. J. (1995) Uptake and fate of ozone in the respiratory tract. Toxicol. Lett. 82/83: 277-285.
Miller, L. A.; Barnett, N. L.; Sheppard, D.; Hyde, D. M. (2001) Expression of the P6 integrin subunit is associated
with sites of neutrophil influx in lung epithelium. J. Histochem. Cytochem. 49: 41-48.
Morgan, M. S.; Meyer, P.; Holub, R.; Frank, R. (1986) Overall and regional lung function in dogs exposed acutely to
ozone. Environ. Res. 41: 546-557.
Morio, L. A.; Hooper, K. A.; Brittingham, J.; Li, T.-H.; Gordon, R. E.; Turpin, B. J.; Laskin, D. L. (2001) Tissue
injury following inhalation of fine paniculate matter and hydrogen perioxide is associated with altered
production of inflammatory mediators and antioxidants by alveolar macrophages. Toxicol. Appl. Pharmacol.
177: 188-199.
Moss, O. R.; Gross, E. A.; James, R. A.; Janszen, D. B.; Ross, P. W.; Roberts, K. C.; Howard, A. M.; Harkema,
J. R.; Calderon-Garciduenas, L.; Morgan, K. T. (2001) Respiratory tract toxicity in rats exposed to Mexico
City air. Cambridge, MA: Health Effects Institute; research report no. 100. Available:
http://www.healtheffects.org/pubs-research.htm [15 May, 2003].
Mticke, W. (1996) The environment and the eye. Topics of ophthalmic toxicology. Leban. Med. J. 44: 146-150.
Mudway, I. S.; Kelly, F. J. (1998) Modeling the interactions of ozone with pulmonary epithelial lining fluid
antioxidants. Toxicol. Appl. Pharmacol. 148: 91-100.
Mudway, I. S.; Kelly, F. J. (2000) Ozone and the lung: a sensitive issue. Mol. Aspects. Med. 21: 1-48.
Murphy, D. J. (2002) Assessment of respiratory function in safety pharmacology. Fundam. Clin. Pharmacol.
16: 183-196.
5-87
-------
Mutoh, T.; Joad, J. P.; Bonham, A. C. (2000) Chronic passive cigarette smoke exposure augments
bronchopulmonary C-fibre inputs to nucleus tractus solitarii neurones and reflex output in young guinea-pigs.
J. Physiol. (London) 523: 223-233.
Nakano, H.; Aizawa, H.; Matsumoto, K.; Fukuyama, S.; Inoue, H.; Hara, N. (2000) Cyclooxygenase-2 participates
in the late phase of airway hyperresponsiveness after ozone exposure in guinea pigs. Eur. J. Pharmacol.
403: 267-275.
National Toxicology Program. (1994) NTP technical report on the toxicology and carcinogenesis studies of ozone
(CAS no. 10028-15-6) and ozone/NNK (CAS no. 10028-15-6/64091-91-4) in F344/N rats and B6C3FJ mice
(inhalation studies). Research Triangle Park, NC: U.S. Department of Health and Human Services, National
Institutes of Health; publication no. 95-3371. (National Toxicology Program technical report series: no. 440).
Neuhaus-Steinmetz, U.; Uffhausen, F.; Herz, U.; Renz, H. (2000) Priming of allergic immune responses by repeated
ozone exposure in mice. Am. J. Respir. Cell Mol. Biol. 23: 228-233.
Nichols, B. G.; Woods, J. S.; Luchtel, D. L.; Corral, J.; Koenig, J. Q. (2001) Effects of ozone exposure on nuclear
factor-KB activation and tumor necrosis factor-a expression in human nasal epithelial cells. Toxicol. Sci.
60: 356-362.
Nielsen, G. D.; Hougaard, K. S.; Larsen, S. T.; Hammer, M; Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Alarie, Y.
(1999) Acute airway effects of formaldehyde and ozone inBALB/c mice. Hum. Exp. Toxicol. 18: 400-409.
Nino-Cabrera, H. G.; Colin-Barenque, L.; Avila-Costa, M. R.; Espinosa-Villanueva, J.; Fortoul, T. L;
Rivas-Arancibia, S. (2002) Differences between hippocampus and cerebral cortex in aged rats in an oxidative
stress model. Int. J. Neurosci. 112: 373-381.
Noviski, N.; Brewer, J. P.; Skornik, W. A.; Galli, S. J.; Drazen, J. M.; Martin, T. R. (1999) Mast cell activation is
not required for induction of airway hyperresponsiveness by ozone in mice. J. Appl. Physiol. 86: 202-210.
Paige, R. C.; Royce, F. H.; Plopper, C. G.; Buckpitt, A. R. (2000a) Long-term exposure to ozone increases acute
pulmonary centriacinar injury by 1-nitronaphthalene: I. Region-specific enzyme activity. J. Pharmacol. Exp.
Ther. 295: 934-941.
Paige, R. C.; Wong, V.; Plopper, C. G. (2000b) Long-term exposure to ozone increases acute pulmonary centriacinar
injury by 1-nitronaphthalene: II. Quantitative histopathology. J. Pharmacol. Exp. Ther. 295: 942-950.
Paquette, N. C.; Tankersley, C. G.; Zhang, L.-Y.; Kleeberger, S. R. (1994) Repeated subacute ozone exposure of
inbred mice: airway inflammation and ventilation. Exp. Lung Res. 20: 579-594.
Paz, C. (1997) Some consequences of ozone exposure on health. Arch. Med. Res. 28: 163-170.
Paz, C.; Bazan-Perkins, B. (1992) Sleep-wake disorganization in cats exposed to ozone. Neurosci. Lett.
140: 270-272.
Paz, C.; Huitron-Resendiz, S. (1996) The effects of ozone exposure on the sleep-wake cycle and serotonin contents
in the pons of the rat. Neurosci. Lett. 204: 49-52.
Pearson, A. C.; Bhalla, D. K. (1997) Effects of ozone onmacrophage adhesion in vitro and epithelial and
inflammatory responses in vivo: the role of cytokines. J. Toxicol. Environ. Health 50: 143-157.
Peden, D. B.; Dailey, L. (1995) Modulation of mast cell functions by in vitro ozone exposure. Am. J. Physiol.
268: L902-L910.
Pendino, K. J.; Shuler, R. L.; Laskin, J. D.; Laskin, D. L. (1994) Enhanced production of interleukin-1, tumor
necrosis factor-a, and fibronectin by rat lung phagocytes following inhalation of a pulmonary irritant. Am. J.
Respir. Cell Mol. Biol. 11: 279-286.
Petruzzi, S.; Fiore, M.; DeH'Omo, G.; Bignami, G.; Alleva, E. (1995) Medium and long-term behavioral effects in
mice of extended gestational exposure to ozone. Neurotoxicol. Teratol. 17: 463-470.
Petruzzi, S.; De Acetis, L.; Chiarotti, F.; Sorace, A.; Alleva, E. (1999) Limited changes in handedness and morphine
reactivity in CD-I mice after pre- and postnatal ozone exposure. Acta Neurobiol. Exp. 59: 115-122.
Pinkerton, K. E.; Weller, B. L.; Menache, M. G.; Plopper, C. G. (1998) Consequences of prolonged inhalation of
ozone on F344/N rats: collaborative studies. Part XIII. A comparison of changes in the tracheobronchial
epithelium and pulmonary acinus in male rats at 3 and 20 months. Cambridge, MA: Health Effects Institute;
research report no. 65.
Plopper, C. G.; Fanucchi, M. V. (2000) Do urban environmental pollutants exacerbate childhood lung diseases?
Environ. Health Perspect. 108: A252-A253.
Plopper, C. G.; Hatch, G. E.; Wong, V.; Duan, X.; Weir, A. J.; Tarkington, B. K.; Devlin, R. B.; Becker, S.;
Buckpitt, A. R. (1998) Relationship of inhaled ozone concentration to acute tracheobronchial epithelial injury,
site-specific ozone dose and glutathione depletion in rhesus monkeys. Am. J. Respir. Cell Mol. Biol.
19: 387-399.
-------
Polosa, R.; Holgate, S. T. (1997) Adenosine bronchoprovocation: a promising marker of allergic inflammation in
asthma? Thorax 52: 919-923.
Pope, C. A., Ill; Burnett, R. T.; Thun, M. I; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA J. Am. Med.
Assoc. 287: 1132-1141.
Postlethwait, E. M.; Cueto, R.; Velsor, L. W.; Pryor, W. A. (1998) O3-induced formation of bioactive lipids:
estimated surface concentrations and lining layer effects. Am. J. Physiol. 274: L1006-L1016.
Postlethwait, E. M.; Joad, J. P.; Hyde, D. M.; Schelegle, E. S.; Brie, J. M.; Weir, A. J.; Putney, L. F.; Wong, V. J.;
Velsor, L. W.; Plopper, C. G. (2000) Three-dimensional mapping of ozone-induced acute cytotoxicity in
tracheobronchial airways of isolated perfused rat lung. Am. J. Respir. Cell Mol. Biol. 22: 191-199.
Pryor, W. A.; Squadrito, G. L.; Friedman, M. (1995) A new mechanism for the toxicity of ozone. Toxicol. Lett.
82/83:287-293.
Pryor, W. A.; Bermudez, E.; Cueto, R.; Squadrito, G. L. (1996) Detection of aldehydes in bronchoalveolar lavage of
rats exposed to ozone. Fundam. Appl. Toxicol. 34: 148-156.
Pulfer, M. K.; Murphy, R. C. (2004) Formation of biologically active oxysterols during ozonolysis of cholesterol
present in lung surfactant. J. Biol. Chem. 279: 26331-26338.
Pulfer, M. K.; Taube, C.; Gelfand, E.; Murphy, R. C. (2005) Ozone exposure in vivo and formation of biologically
active oxysterols in the lung. J. Pharmacol. Exp. Ther. 312: 256-264.
Quinlan, T.; Spivack, S.; Mossman, B. T. (1994) Regulation of antioxidant enzymes in lung after oxidant injury.
Environ. Health Perspect. 102(suppl. 2): 79-87.
Rehle, D.; Leleux, D.; Erdelyi, M.; Tittel, F.; Fraser, M.; Friedfeld, S.; et al. (2001) Ambient formaldehyde detection
with a laser spectrometer based on difference-frequency generation in PPLN. Appl. Phys. B: Lasers Opt.
72: 947-952.
Reinhart, P. G.; Gupta, S. K.; Bhalla, D. K. (1999) Attenuation of ozone-induced lung injury by interleukin-10.
Toxicol. Lett. 110:35-42.
Renner, R. (2002) Bad air and birth defects. Environ. Health Perspect. 110: A291.
Ritz, B.; Yu, F. (1999) The effect of ambient carbon monoxide on low birth weight among children born in southern
California between 1989 and 1993. Environ. Health Perspect. 107: 17-25.
Rivas-Arancibia, S.; Vazquez-Sandoval, R.; Gonzalez-Kladiano, D.; Schneider-Rivas, S.; Lechuga-Guerrero, A.
(1998) Effects of ozone exposure in rats on memory and levels of brain and pulmonary superoxide dismutase.
Environ. Res. 76: 33-39.
Rivas-Arancibia, S.; Dorado-Martinez, C.; Borgonio-Perez, G.; Hiriart-Urdanivia, M.; Verdugo-Diaz, L.;
Duran-Vazquez, A.; Colin-Baranque, L.; Avila-Costa, M. R. (2000) Effects of taurine on ozone-induced
memory deficits and lipid peroxidation levels in brains of young, mature, and old rats. Environ. Res. 82: 7-17.
Rivas-Arancibia, S.; Dorado-Martinez, C.; Colin-Barenque, L.; Kendrick, K. M.; De laRiva, C.; Guevara-Guzman,
R. (2003) Effect of acute ozone exposure on locomotor behavior and striatal function. Pharmacol. Biochem.
Behav. 74: 891-900.
Rivas-Manzano, P.; Paz, C. (1999) Cerebellar morphological alterations in rats induced by prenatal ozone exposure.
Neurosci. Lett. 276: 37-40.
Rose, R. C.; Richer, S. P.; Bode, A. M. (1998) Ocular oxidants and antioxidant protection. Proc. Soc. Exp. Biol.
Med. 217: 397-407.
Savov, J. D.; Whitehead, G. S.; Wang, J.; Liao, G.; Usuka, J.; Peltz, G.; Foster, W. M.; Schwartz, D. A. (2004)
Ozone-induced acute pulmonary injury in inbred mouse strains. Am. J. Respir. Cell Mol. Biol. 31: 69-11.
Schelegle, E. S.; Alfaro, M. F.; Putney, L.; Stovall, M.; Tyler, N.; Hyde, D. M. (2001) Effect of C-fiber-mediated,
ozone-induced rapid shallow breathing on airway epithelial injury in rats. J. Appl. Physiol. 91: 1611-1618.
Schelegle, E. S.; Miller, L. A.; Gershwin, L. J.; Fanucchi, M. V; Van Winkle, L. S.; Gerriets, J. E.; Walby, W. F.;
Mitchell, V; Tarkington, B. K.; Wong, V. J.; Baker, G. L.; Pantle, L. M.; Joad, J. P.; Pinkerton, K. E.;
Wu, R.; Evans, M. J.; Hyde, D. M.; Plopper, C. G. (2003a) Repeated episodes of ozone inhalation amplifies
the effects of allergen sensitization and inhalation on airway immune and structural development in Rhesus
monkeys. Toxicol. Appl. Pharmacol. 191: 74-85.
Schelegle, E. S.; Walby, W. F.; Alfaro, M. F.; Wong, V. J.; Putney, L.; Stovall, M. Y.; Sterner-Kock, A.; Hyde,
D. M.; Plopper, C. G. (2003b) Repeated episodes of ozone inhalation attenuates airway injury/repair and
release of substance P, but not adaptation. Toxicol. Appl. Pharmacol. 186: 127-142.
Schlesinger, R. B. (1995) Interaction of gaseous and paniculate pollutants in the respiratory tract: mechanisms and
modulators. Toxicology 105: 315-325.
5-89
-------
Schlesinger, R. B.; Cohen, M. D.; Gordon, T.; Nadziejko, C.; Zelikoff, J. T.; Sisco, M.; Regal, J. F.; Menache, M. G.
(2002a) Ozone differentially modulates airway responsiveness in atopic versus nonatopic guinea pigs.
Inhalation Toxicol. 14: 431-457.
Schlesinger, R. B.; Cohen, M.; Gordon, T.; Nadziejko, C.; Zelikoff, J. T.; Sisco, M.; Regal, J. F.; Menache, M. G.
(2002b) Ozone-induced modulation of airway hyperresponsiveness in guinea pigs. Boston, MA: Health
Effects Institute; research report no. 109.
Schwartz, D. A. (2002) TLR4 and LPS hyporesponsiveness in humans. Int. J. Hyg. Environ. Health 205: 221-227.
Segura, P.; Montano, L. M.; Bazan-Perkins, B.; Gustin, P.; Vargas, M. H. (1997) Ozone at high-pollution urban
levels causes airway hyperresponsiveness to substance P but not to other agonists. Environ. Toxicol.
Pharmacol. 3: 91-95.
Sen, S.; Dulchavsky, S. A.; Dutta, S. (1993) Effects of triiodothyronine (T3) supplementation upon ozone-induced
lung injury. Free Radic. Res. Commun. 18: 299-308.
Shore, S. A.; Abraham, J. H.; Schwartzman, I. N.; Murthy, G. G.; Laporte, J. D. (2000) Ventilatory responses to
ozone are reduced in immature rats. J. Appl. Physiol. 88: 2023-2030.
Shore, S. A.; Johnston, R. A.; Schwartzman, I. N.; Chism, D.; Krishna Murthy, G. G. (2002) Ozone-induced airway
hyperresponsiveness is reduced in immature mice. J. Appl. Physiol. 92: 1019-1028.
Shore, S. A.; Rivera-Sanchez, Y. M.; Schwartzman, I. N.; Johnston, R. A. (2003) Responses to ozone are increased
in obese mice. J. Appl. Physiol. 95: 938-945.
Sindhu, R. K.; Mautz, W. J.; Kikkawa, Y. (1998) Chronic exposure to ozone and nitric acid vapor results in
increased levels of rat pulmonary putrescine. Arch. Toxicol. 72: 445-449.
Slade, R.; Watkinson, W. P.; Hatch, G. E. (1997) Mouse strain differences in ozone dosimetry and body temperature
changes. Am. J. Physiol. 272: L73-L77.
Sorace, A.; De Acetis, L.; Alleva, E.; Santucci, D. (2001) Prolonged exposure to low doses of ozone: short- and
long-term changes to behavioral performance in mice. Environ. Res. 85: 122-134.
Spannhake, E. W. (1996) Down-regulation of canine airway mast cell function following exposure to ozone in vivo.
Exp. Lung Res. 22: 163-178.
Sterner-Kock, A.; Kock, M.; Braun, R.; Hyde, D. M. (2000) Ozone-induced epithelial injury in the ferret is similar
to nonhuman primates. Am. J. Respir. Crit. Care Med. 162: 1152-1156.
Sun, J.; Chung, K. F. (1997) Interaction of ozone exposure with airway hyperresponsiveness and inflammation
induced by trimellitic anhydride in sensitized guinea pigs. J. Toxicol. Environ. Health 51: 77-87.
Sun, J.; Koto, H.; Chung, K. F. (1997) Interaction of ozone and allergen challenges on bronchial responsiveness and
inflammation in sensitised guinea pigs. Int. Arch. Allergy Immunol. 112: 191-195.
Szarek, J. L.; Stewart, N. L.; Zhang, J. Z.; Webb, J. A.; Valentovic, M. A.; Catalano, P. (1995) Contractile responses
and structure of small bronchi isolated from rats after 20 months' exposure to ozone. Fundam. Appl. Toxicol.
28: 199-208.
Tager, I. B. (1999) Air pollution and lung function growth. Is it ozone? Am. J. Respir. Crit. Care Med. 160: 387-389.
Takahashi, T.; Miura, M.; Katsumata, U.; Ichinose, M.; Kimura, K.; Inoue, H.; Takishima, T.; Shirato, K. (1993)
Involvement of superoxide in ozone-induced airway hyperresponsiveness in anesthetized cats. Am. Rev.
Respir. Dis. 148: 103-106.
Takahashi, N.; Yu, X.-Y.; Schofield, B. H.; Kleeberger, S. R.; Scott, A. L.; Hasegawa, S.; Spannhake, E. W. (1995a)
Expression of ICAM-1 in airway epithelium after acute ozone exposure in the mouse. J. Appl. Physiol.
79: 1753-1761.
Takahashi, M.; Kleeberger, S. R.; Croxton, T. L. (1995b) Genetic control of susceptibility to ozone-induced changes
in mouse trachea! electrophysiology. Am. J. Physiol. 269: L6-L10.
Takebayashi, T.; Abraham, J.; Murthy, G. G. K.; Lilly, C.; Rodger, I.; Shore, S. A. (1998) Role of tachykinins in
airway responses to ozone in rats. J. Appl. Physiol. 85: 442-450.
Tankersley, C. G.; Kleeberger, S. R. (1994) Ozone-induced inflammation and altered ventilation in genetically
susceptible mice: a comparison of acute and subacute exposures. Toxicol Lett. 72: 279-289.
Tankersley, C. G.; Fitzgerald, R. S.; Mitzner, W. A.; Kleeberger, S. R. (1993) Hypercapnic ventilatory responses in
mice differentially susceptible to acute ozone exposure. J. Appl. Physiol. 75: 2613-2619.
Tesfaigzi, J.; Hotchkiss, J. A.; Harkema, J. R. (1998) Expression of the Bcl-2 protein in nasal epithelia of F344/N
rats during mucous cell metaplasia and remodeling. Am. J. Respir. Cell Mol. Biol. 18: 794-799.
Thiele, J. J. (2001) Oxidative targets in the stratum corneum. A new basis for antioxidative strategies. Skin
Pharmacol. Appl. Skin Physiol. 14(suppl. 1): 87-91.
Thiele, J. J.; Traber, M. G.; Packer, L. (1998) Depletion of human stratum corneum vitamin E: an early and sensitive
in vivo marker of UV induced photo-oxidation. J. Invest. Dermatol. 110: 756-761.
5-90
-------
Thiele, J. I; Hsieh, S. N.; Briviba, K.; Sies, H. (1999) Protein oxidation in human stratum corneum: susceptibility of
keratins to oxidation in vitro and presence of a keratin oxidation gradient in vivo. J. Invest. Dermatol.
113:335-339.
Tsai, J.-J.; Lin, Y.-C.; Kwan, Z.-H.; Kao, H.-L. (1998) Effects of ozone on ovalbumin sensitization in guinea pigs.
J. Microbiol. Immunol. Infect. 31: 225-232.
U.S. Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park,
NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
nos. EPA/600/8-9!/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525, and
PB95-124517.
U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EP A/600/ AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available online at:
www.epa.gov/ncea/ozone.htm.
Uhlson, C.; Harrison, K.; Allen, C. B.; Ahmad, S.; White, C. W.; Murphy, R. C. (2002) Oxidized phospholipids
derived from ozone-treated lung surfactant extract reduce macrophage and epithelial cell viability. Chem.
Res. Toxicol. 15: 896-906.
Ulrich, M. M. W.; Alink, G. M; Kumarathasan, P.; Vincent, R.; Boere, A. J.; Cassee, F. R. (2002) Health effects
and time course of paniculate matter on the cardiopulmonary system in rats with lung inflammation.
J. Toxicol. Environ. Health Part A 65: 1571-1595.
Valacchi, G.; Weber, S. U.; Luu, C.; Cross, C. E.; Packer, L. (2000) Ozone potentiates vitamin E depletion by
ultraviolet radiation in the murine stratum corneum. FEES Lett. 466: 165-168.
Valacchi, G.; Pagnin, E.; Okamoto, T.; Corbacho, A. M.; Olano, E.; Davis, P. A.; Van der Vliet, A.; Packer, L.;
Cross, C. E. (2003) Induction of stress proteins and MMP-9 by 0.8 ppm of ozone in murine skin. Biochem.
Biophys. Res. Commun. 305: 741-746.
Valverde, M.; del Carmen Lopez, M.; Lopez, I.; Sanchez, I.; Fortoul, T. I.; Ostrosky-Wegman, P.; Rojas, E. (1997)
DNA damage in leukocytes and buccal and nasal epithelial cells of individuals exposed to air pollution in
Mexico City. Environ. Mol. Mutagen. 30: 147-152.
Van Bree, L.; Dormans, J. A. M. A.; Boere, A. J. F.; Rombout, P. J. A. (2001) Time study on development and
repair of lung injury following ozone exposure in rats. Inhalation Toxicol. 13: 703-717.
Van Bree, L.; Dormans, J. A. M. A.; Koren, H. S.; Devlin, R. B.; Rombout, P. J. A. (2002) Attenuation and recovery
of pulmonary injury in rats following short-term, repeated daily exposure to ozone. Inhalation Toxicol.
14: 883-900.
Van Hoof, I. H. J. M.; Van Bree, L.; Bast, A. (1996) Changes in receptor function by oxidative stress in guinea pig
trachea! smooth muscle. Cent. Eur. J. Public Health 4(suppl.): 3-5.
Van Hoof, H. J. M.; Van Acker, F. A. A.; Voss, H.-P.; Van Bree, L.; Bast, A. (1997a) Acute exposure to ozone does
not influence neuroreceptor density and sensitivity in guinea pig lung. Toxicol. Lett. 90: 53-60.
Van Hoof, H. J. M.; Voss, H.-P.; Kramer, K.; Boere, A. J. F.; Dormans, J. A. M. A.; Van Bree, L.; Bast, A. (1997b)
Changes in neuroreceptor function of tracheal smooth muscle following acute ozone exposure of guinea pigs.
Toxicology 120: 159-169.
Vanda, B.; de Buen, N.; Jasso, R.; Valero, G.; Vargas, M. H.; Olmos, R.; Arreola, J. L.; Santillan, P.; Alonso, P.
(1998) Inflammatory cells and ferruginous bodies inbronchoalveolar lavage in urban dogs. Acta Cytol.
42: 939-944.
Vargas, M. H.; Romero, L.; Sommer, B.; Zamudio, P.; Gustin, P.; Montano, L. M. (1998) Chronic exposure to ozone
causes tolerance to airway hyperresponsiveness in guiea pigs: lack of SOD role. J. Appl. Physiol.
84: 1749-1755.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994a) Ozone increases amino- and
carboxy-terminal atrial natriuretic factor prohormone peptides in lung, heart, and circulation. J. Biochem.
Toxicol. 9: 107-112.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994b) Increase in atrial natriuretic factor in
the lungs, heart, and circulatory system owing to ozone. Chest 105: 1551-1554.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994c) Ozone increases atrial natriuretic
peptides in heart, lung and circulation of aged vs. adult animals. Gerontology (Basel) 40: 227-236.
Vesely, K. R.; Schelegle, E. S.; Stovall, M. Y.; Harkema, J. R.; Green, J. F.; Hyde, D. M. (1999a) Breathing pattern
response and epithelial labeling in ozone-induced airway injury in neutrophil-depleted rats. Am. J. Respir.
Cell Mol. Biol. 20:699-709.
5-91
-------
Vesely, K. R.; Hyde, D. M; Stovall, M. Y.; Harkema, J. R.; Green, J. F.; Schelegle, E. S. (1999b)
Capsaicin-sensitive C-fiber-mediated protective responses in ozone inhalation in rats. J. Appl. Physiol.
86: 951-962.
Vincent, R.; Janzen, E. G.; Chen, G.; Kumarathasan, P.; Haire, D. L.; Guenette, J.; Chen, J. Z.; Bray, T. M. (1996)
Spin trapping study in the lungs and liver of F344 rats after exposure to ozone. Free Radical Res. 25: 475-488.
Vincent, R.; Bjarnason, S. G.; Adamson, I. Y. R.; Hedgecock, C.; Kumarathasan, P.; Guenette, J.; Potvin, M.;
Goegan, P.; Bouthillier, L. (1997) Acute pulmonary toxicity of urban paniculate matter and ozone. Am. J.
Pathol. 151: 1563-1570.
Vyskocil, A.; Viau, C.; Lamy, S. (1998) Peroxyacetyl nitrate: review of toxicity. Hum. Exp. Toxicol. 17: 212-220.
Wagner, J. G.; Hotchkiss, J. A.; Harkema, J. R. (200la) Effects of ozone and endotoxin coexposure on rat airway
epithelium: potentiation of toxicant-induced alterations. Environ. Health Perspect. 109(suppl. 4): 591-598.
Wagner, J. G.; VanDyken, S. J.; Hotchkiss, J. A.; Harkema, J. R. (200Ib) Endotoxin enhancement of ozone-induced
mucous cell metaplasia is neutrophil-dependent in rat nasal epithelium. Toxicol. Sci. 60: 338-347.
Wagner, J. G.; Hotchkiss, J. A.; Harkema, J. R. (2002) Enhancement of nasal inflammatory and epithelial responses
after ozone and allergen coexposure in brown Norway rats. Toxicol. Sci. 67: 284-294.
Wagner, J. G.; VanDyken, S. J.; Wierenga, J. R.; Hotchkiss, J. A.; Harkema, J. R. (2003) Ozone exposure enhances
endotoxin-induced mucous cell metaplasia in rat pulmonary airways. Toxicol. Sci. 74: 437-446.
Wang, G.; Umstead, T. M.; Phelps, D. S.; Al-Mondhiry, H.; Floras, J. (2002) The effect of ozone exposure on the
ability of human surfactant protein A variants to stimulate cytokine production. Environ. Health Perspect.
110:79-84.
Watkinson, W. P.; Wiester, M. J.; Highfill, J. W. (1995) Ozone toxicity in the rat. I. Effect of changes in ambient
temperature on extrapulmonary physiological parameters. J. Appl. Physiol. 78: 1108-1120.
Watkinson, W. P.; Campen, M. J.; Nolan, J. P.; Costa, D. L. (2001) Cardiovascular and systemic responses to
inhaled pollutants in rodents: effects of ozone and paniculate matter. Environ. Health Perspect.
109(suppl. 4): 539-546.
Watkinson, W. P.; Campen, M. J.; Wichers, L. B.; Nolan, J. P.; Costa, D. L. (2003) Cardiac and thermoregulatory
responses to inhaled pollutants in healthy and compromised rodents: modulation via interaction with
environmental factors. Environ. Res. 92: 35-47.
Watt, K. C.; Plopper, C. G.; Weir, A. J.; Tarkington, B.; Buckpitt, A. R. (1998) Cytochrome P450 2E1 in rat
tracheobronchial airways: response to ozone exposure. Toxicol. Appl. Pharmacol. 149: 195-202.
Wattiez, R.; Noel-Georis, L; Cruyt, C.; Broeckaert, F.; Bernard, A.; Falmagne, P. (2003) Susceptibility to oxidative
stress: proteomic analysis of bronchoalveolar lavage from ozone-sensitive and ozone-resistant strains of mice.
Proteomics 3: 658-665.
Weber, S. U.; Thiele, J. J.; Cross, C. E.; Packer, L. (1999) Vitamin C, uric acid, and glutathione gradients in murine
stratum corneum and their susceptibility to ozone exposure. J. Invest. Dermatol. 113:1128-1132.
Weber, S. U.; Jothi, S.; Thiele, J. J. (2000) High-pressure liquid chromatography analysis of ozone-induced
depletion of hydrophilic and lipophilic antioxidants in murine skin. Methods Enzymol. 319: 536-546.
Weber, S. U.; Han, N.; Packer, L. (2001) Ozone: an emerging oxidative stressorto skin. Curr. Probl. in Dermatol.
29: 52-61.
Weller, B. L.; Crapo, J. D.; Slot, J.; Posthuma, G.; Plopper, C. G.; Pinkerton, K. E. (1997) Site- and cell-specific
alteration of lung copper/zinc and manganese superoxide dismutases by chronic ozone exposure. Am. J.
Respir. CellMol. Biol. 17: 552-560.
Weller, B. L.; Witschi, H.; Pinkerton, K. E. (2000) Quantitation and localization of pulmonary manganese
superoxide dismutase and tumor necrosis factor a following exposure to ozone and nitrogen dioxide. Toxicol.
Sci. 54: 452-461.
Wells, C. A.; Ravasi, T.; Faulkner, G. J.; Carninci, P.; Okazaki, Y.; Hayashizaki, Y.; Sweet, M.; Wainwright, B. J.;
Hume, D. A. (2003) Genetic control of the innate immune response. BMC Immunol. 4: 5. Available:
http://www.biomedcentral.eom/1471-2172/4/5 [18 February, 2003]
Wiester, M. J.; Watkinson, W. P.; Costa, D. L.; Crissman, K. M.; Richards, J. H.; Winsett, D. W.; Highfill, J. W.
(1996) Ozone toxicity in the rat. III. Effect of changes in ambient temperature on pulmonary parameters.
J. Appl. Physiol. 81: 1691-1700.
Wilkins, C. K.; Clausen, P. A.; Wolkoff, P.; Larsen, S. T.; Hammer, M.; Larsen, K.; Hansen, V; Nielsen, G. D.
(2001) Formation of strong irritants in mixtures of isoprene/ozone and isoprene/ozone/nitrogen dioxide.
Environ. Health Perspect. 109: 937-941.
Witschi, H.; Espiritu, L; Pinkerton, K. E.; Murphy, K.; Maronpot, R. R. (1999) Ozone carcinogenesis revisited.
Toxicol. Sci. 52: 162-167.
5-92
-------
Wu, Z.-X.; Morton, R. F.; Lee, L.-Y. (1997) Role of tachykinins in ozone-induced airway hyperresponsiveness to
cigarette smoke in guinea pigs. J. Appl. Physiol. 83: 958-965.
Yamauchi, T.; Shima, M.; Kuwaki, T.; Ando, M; Ohmichi, M.; Fukuda, Y.; Adachi, M. (2002) Acute effects of
ozone exposure on lung function in mice sensitized to ovalbumin. Toxicology (Ireland) 172: 69-78.
Yu, M.; Pinkerton, K. E.; Witschi, H. (2002) Short-term exposure to aged and diluted sidestream cigarette smoke
enhances ozone-induced lung injury inB6C3Fl mice. Toxicol. Sci. 65: 99-106.
Zhang, L.-Y.; Levitt, R. C.; Kleeberger, S. R. (1995) Differential susceptibility to ozone-induced airways
hyperreactivity in inbred strains of mice. Exp. Lung Res. 21: 503-518.
Zhao, Q.; Simpson, L. G.; Driscoll, K. E.; Leikauf, G. D. (1998) Chemokine regulation of ozone-induced neutrophil
and monocyte inflammation. Am. J. Physiol. 274: L39-L46.
5-93
-------
6. CONTROLLED HUMAN EXPOSURE STUDIES
OF OZONE AND RELATED
PHOTOCHEMICAL OXIDANTS
6.1 INTRODUCTION
In the previous chapter, results of ozone (O3) studies in laboratory animals and in vitro test
systems were presented. The extrapolation of results from animal studies is one mechanism by
which information on potential adverse human health effects from exposure to O3 is obtained.
More direct evidence of human health effects due to O3 exposure can be obtained through
controlled human exposure studies of volunteers or through field and epidemiologic studies of
populations exposed to ambient O3 (see Chapter 7). Controlled human exposure studies
typically use fixed concentrations of O3 under carefully regulated environmental conditions and
subject activity levels. This chapter discusses studies in which volunteers were exposed for up
to 8 h to O3 concentrations ranging from 0.04 to 0.75 ppm O3 while at rest or during varying
intensities of exercise.
The majority of controlled human studies have investigated the effects of exposure to O3 in
young nonsmoking healthy adults (18 to 35 years of age) performing continuous exercise (CE)
or intermittent exercise (IE). Various combinations of O3 concentration, exercise routine, and
exposure duration have been used in these studies. The responses to ambient O3 concentrations
include decreased inspiratory capacity; mild bronchoconstriction; rapid, shallow breathing
patterns during exercise; and symptoms of cough and pain on deep inspiration. Reflex inhibition
of inspiration results in a decrease in forced vital capacity (F VC) and total lung capacity (TLC)
and, in combination with mild bronchoconstriction, contributes to a decrease in the forced
expiratory volume in 1 s (FEVj). In addition to physiological pulmonary responses and
respiratory symptoms, O3 has been shown to result in airway hyperresponsiveness, epithelial
permeability, and inflammation.
The most salient observations from studies reviewed in the 1996 EPA Ozone Air Quality
Criteria Document or O3 AQCD (U.S. Environmental Protection Agency, 1996) were that:
(1) young healthy adults exposed to O3 concentrations >0.08 ppm develop significant reversible,
6-1
-------
transient decrements in pulmonary function if minute ventilation (VE ) or duration of exposure is
increased sufficiently, (2) children experience similar spirometric responses but lesser symptoms
from O3 exposure relative to young adults, (3) O3-induced spirometric responses are decreased in
the elderly relative to young adults, (4) there is a large degree of intersubject variability in
physiologic and symptomatic responses to O3 but responses tend to be reproducible within a
given individual over a period of several months, (5) subjects exposed repeatedly to O3 for
several days develop a tolerance to successive exposures, as demonstrated by an attenuation of
responses that is lost after about a week without exposure, and (6) acute O3 exposure initiates an
inflammatory response which may persist for at least 18 to 24 h postexposure.
There are several important limitations associated with these clinical studies: (1) the
ability to study only short-term, acute effects; (2) difficulties in trying to link short-term effects
with long-term consequences; (3) the use of a small number of volunteers that may not be
representative of the general population; and (4) statistical limitations associated with the small
sample size. Sample size affects the power of a study, and having a small number of samples
causes a risk of Type II error, i.e., the incorrect conclusion that no difference exists between
treatments or groups when comparisons are not significantly different. This affects the
confidence in estimates of a minimum O3 concentration at which some degree of pulmonary
impairment will occur in both the general population and susceptible subpopulations. As a
result, the conclusions drawn from many of the studies cited in this chapter may underestimate
the presence of responses at low O3 concentrations and low activity levels.
Most of the scientific information summarized in this chapter comes from the literature
published since the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996). In addition
to further study of physiological pulmonary responses and symptoms of breathing discomfort,
much of this literature has focused on mechanisms of inflammation and cellular responses to
injury induced by O3 inhalation. A more thorough discussion and review of this literature
appears in Annex AX6 of this document. In summarizing the literature, effects are described if
they are statistically significant at a probability (p-value) of less than 0.05; otherwise, trends are
noted as such.
As spirometry typically improves in healthy young adults with exercise exposures to
filtered air (FA), the term "O3-induced" is used herein and in the annex to designate effects that
have been corrected for responses during FA exposures. For healthy adults, an O3-induced
6-2
-------
change in lung function is the difference between the decrement experienced with O3 exposure
and the improvement observed with FA exposure. However, the distinction between an O3-
induced change and a post- versus preexposure change is particularly important in individuals
with respiratory disease who may experience exercise-induced decrements in pulmonary
function during both FA and O3 exposures. Hence, in subjects with respiratory disease, exercise-
induced responses could be mistaken for O3-induced responses in the absence of a correction for
FA responses.
6.2 PULMONARY FUNCTION EFFECTS OF OZONE EXPOSURE
IN HEALTHY SUBJECTS
6.2.1 Introduction
As reviewed in the 1986 and 1996 O3 AQCD's (U.S. Environmental Protection Agency,
1986, 1996), 0.5 ppm is the lowest O3 concentration at which statistically significant reductions
in FVC and FEVj have been reported in sedentary subjects. On average, young adults (n = 23;
mean age, 22 yrs) exposed at rest for 2 h to 0.5 ppm O3 had O3-induced decrements of-4% in
FVC and -7% in FEVj (Folinsbee et al., 1978; Horvath et al., 1979). During exercise,
spirometric and symptoms responses are observed at lower O3 concentrations. For acute
exposures of 1 to 2 h to >0.12 ppm O3, if VE is sufficiently increased by exercise, healthy human
subjects generally experience (a) decreases in TLC, inspiratory capacity (1C), FVC, FEVl3 mean
forced expiratory flow from 25% to 75% of FVC (FEF25.75), and tidal volume (VT) and
(b) increases in specific airways resistance (sRaw), breathing frequency (fB), and airway
responsiveness. These exposures also cause symptoms of cough; pain on deep inspiration; rapid,
shallow breathing patterns during exercise; throat irritation; and wheezing. With exposures of 4-
to 8-h in duration, statistically significant pulmonary function and symptoms responses are
observed at lower O3 concentrations and lower VE than in shorter duration studies.
A large body of data regarding the interdependent effects of O3 concentration (C), minute
ventilation (VE), and duration of exposure (time, T) on pulmonary responses was assessed in the
1986 and 1996 O3 AQCD's (U.S. Environmental Protection Agency, 1986, 1996). In an attempt
to describe O3 dose-response characteristics, acute responses were modeled as a function of total
-------
inhaled O3 dose (C * T x VE), which was found to be a better predictor of response than O3
concentration, VE, or T of exposure, alone, or as a combination of any two of these factors. For
example, in an analysis of 6 studies with 1- to 2-h exposures to between 0.12 and 0.18 ppm O3
with exercise, Folinsbee et al. (1988) found a good correlation (r = 0.81) between total inhaled
O3 dose and FEVj decrements. Other studies, however, reported that O3 concentration appears to
be more important than either VE or T in determining pulmonary responses during acute
exposures at constant concentration (Adams et al., 1981, 1987; Folinsbee et al, 1978; Hazucha,
1987; Larsen et al., 1991). When considering variable O3 concentrations, where response is not
a linear function of dose, the use of O3 dose as a predictor of response becomes more complex,
if appropriate at all (e.g., see Section 6.2.4).
6.2.2 Acute Exposure for Up to 2 h
With heavy CE (VE = 89 L/min), an O3-induced decrement of 9.7% in FEVj has
been reported for healthy young adults (n = 17; age, 24 ± 3 yrs) exposed for only 1 h to
0.12 ppm O3 (Gong et al., 1986). With moderate-to-heavy IE (15 min intervals of rest and
exercise [VE = 68 L/min]), McDonnell et al. (1983) reported a physiologically small, but
significant, O3-induced decrement of 3.4% in FEVj for young healthy adults (n = 22, age,
22 ± 3 yrs) exposed for 2 h to 0.12 ppm O3. Using the same 2 h IE exposure protocol, Linn et al.
(1986) found no statistically significant spirometic responses at O3 concentrations of 0.16 ppm
and lower. However, the subjects in the Linn et al. (1986) study were potentially exposed
concurrently in Los Angeles to ambient O3 levels of between 0.12 and 0.16 ppm and were on
average 3 yrs older than the subjects in the McDonnell et al. (1983) study. (The attenuating
effects of increasing age and repeated O3 exposures are discussed in Sections 6.5.1 and 6.6,
respectively.) The disparities between the Linn et al. (1986) and McDonnell et al. (1983) studies
demonstrate the difficulty in determining a no-effect-level for O3 based on relatively small study
populations.
Studies analyzing large data sets (>300 subjects) provide better predictive ability of acute
changes in FEVj at low levels of O3 and VEthan is possible via comparisons between smaller
studies. Such an analysis was performed by McDonnell et al. (1997), who examined FEVj
responses in 485 healthy white males (18 to 36 years of age; subjects recruited from the area
6-4
-------
around Chapel Hill, NC) exposed once for 2 h to O3 concentrations of up to 0.40 ppm at rest or
with IE. Decrements in FEVj were modeled by sigmoid-shaped curve as a function of subject
age, O3 concentration, VE, and duration of exposure. Regarding applicability to the general
population, the McDonnell et al. (1997) model has an apparent limitation of considering only
data for young adult white males. However, two other analyses (93 black females; 94 white
females; 92 black males; 93 white males, a subset of subjects included in the McDonnell et al.
study) found no significant gender or race effects on spirometric responses to O3 exposure (Seal
et al., 1993, 1996). It should be emphasized that the McDonnell et al. (1997) model only
predicts average responses. For exposure conditions causing a predicted average FEVj
decrement of 10%, individual decrements range from approximately 0 to 40% (see Figure 1 of
McDonnell et al., 1997). (The reader is referredto Sections 6.5.1 and 6.5.4 for further
discussion of this model and its prediction of age and physical activity effects on FEV,
responses.)
In a more recent study, McDonnell et al. (1999) also reported a model predicting average
symptom responses from O3 exposure. Unfortunately, neither of these papers (McDonnell et al.,
1997, 1999) provide predictions of intersubject variability in response. Ultman et al. (2004)
recently reported pulmonary responses in 60 young heathy nonsmoking adults (32 M, 28 F)
exposed to 0.25 ppm O3 for 1 h with CE at a target VE of 30 L/min. Consistent with findings
reported in the 1996 O3 criteria document, considerable intersubject variability in FEVj
decrements was reported by Ultman et al. (2004) with responses ranging from a 4%
improvement to a 56% decrement. One-third of the subjects had FEVj decrements of >15% and
7% of the subjects had decrements of >40%. (Section 6.4 of this Chapter discusses intersubject
variability in response to O3 exposure.}
In addition to overt effects of O3 exposure on the large airways indicated by spirometric
responses, O3 exposure also affects the function of the small airways and parenchymal lung.
Foster et al. (1993, 1997) examined the effect of O3 on ventilation distribution in healthy adult
males. In healthy nonsmoking males (26.7 ± 7 years old) exposed to FA or 0.33 ppm O3 for 2 h
with IE, there was a significant reduction in ventilation to the lower lung (31% of lung volume)
and significant increases in ventilation to the upper- and middle-lung regions relative to the FA
values in 7 of the 9 subjects (Foster et al., 1993). In another study, 15 healthy nonsmoking
6-5
-------
males (25.4 ± 2 years old) were exposed to FA or 0.35 ppm O3 for 2.2 h with IE (Foster et al.,
1997). Following O3 exposure, an inert gas washout was delayed and resembled a two-
compartment washout, whereas pre-O3 exposure a log-linear gas clearance as a function of
expired volume resembled a single-compartment washout. The pronounced slow phase of gas
washout occurring post-O3, represented a 24% decrease in the washout rate relative to pre-O3.
At 24-h post-O3, 6 of the 12 subjects still had (or developed) a delayed washout relative to the
pre-O3 maneuver. This suggests a prolonged O3 effect on the small airways and ventilation
distribution in some individuals.
6.2.3 Prolonged Ozone Exposures
In the exposure range of 0.08 to 0.16 ppm O3, a number of studies using moderate
quasi continuous exercise (QCE; 50 min exercise and 10 min rest per h) for 4 to 8 h have
shown significant responses under the following conditions: 0.16 ppm for 4 h with QCE
at VE « 40 L/min (Folinsbee et al., 1994), 0.08 to 0.12 ppm for 6.6 h with QCE at VE « 35 to
40 L/min (Adams, 2002; Adams, 2003a; Folinsbee et al., 1988; Horstman et al., 1990), and
0.12 ppm for 8 h of IE (30 min per h) at VE « 40 L/min (Hazucha et al., 1992). Symptoms and
spirometric responses increased with duration of exposure, O3 concentration, and total VE.
Airway resistance is only modestly affected with moderate or even heavy exercise combined
with O3 exposure (Folinsbee et al., 1978; McDonnell et al., 1983; Seal et al., 1993).
6.2.3.1 Effect of Exercise Ventilation Rate on FEVt Response to 6.6 h Ozone Exposure
It is well established that response to O3 exposure is a function of VE in studies of 2 h or
less in duration (See Section AX6.2.2). It is reasonable to expect that response to a prolonged
6.6-h O3 exposure is also a function of VE, although quantitative analyses are lacking. Data
from five similar prolonged exposure studies are available for evaluation of FEVj responses as a
function of exercise VE (Adams, 2000; Adams and Ollison, 1997; Folinsbee et al., 1988, 1994;
Horstman et al., 1990). Each of these studies exposed similarly aged subjects (mean ages 22 to
25 yrs) to 0.12 ppm O3 for 6.6 h. In total, ten sets of mean FEVj decrements were available for
exercise VE ranging from 20 to 43 L/min, although no data were available for VE between
20 and 30 L/min (data illustrated in Figure AX6-2). As in 2-h exposure studies, FEVj
6-6
-------
decrements are a function of VE in prolonged 6.6-h exposure studies as indicated by a significant
correlation between these variables (Pearson, r = 0.95, p < 0.001; Spearman, r = 0.84, p < 0.01).
6.2.3.2 Exercise Ventilation Rate as a Function of Body/Lung Size on FEVt Response
to 6.6 h Ozone Exposure
Based on the assumption that the total inhaled O3 dose (product of O3 concentration,
exposure duration, and VE) is proportional to the lung size, exercise VE is typically selected to be
a multiple of body surface area (BSA) or FVC. Data from several recent studies do not support
the contention that VE should be normalized. In an analysis of data from 485 young adults,
McDonnell et al. (1997) found that any effect of BSA, height, or baseline FVC on percent
decrement in FEVj was small to nonexistent. This is consistent with Messineo and Adams
(1990), who compared pulmonary function responses in young adult women having small
(n= 14) or large (n = 14) lung sizes (mean FVC of 3.74 and 5.11 L, respectively) and found no
significant group difference in FEVj decrements. For 30 subjects (15M, 15F) exposed to
0.12 ppm O3 for 6.6 h, Adams (2000) also reported that FEVj responses were more closely
related to VE than to VE normalized to BSA. The O3 dosimetry study of Bush et al. (1996)
suggested that normalization of the O3 dose might more appropriately be a function of anatomic
dead space. Ozone penetrates deeper into the lungs of individuals with larger conducting airway
volumes, however, differences in FEVj responses between subjects exposed for 2 h to 0.25 ppm
O3 (VE =30 L/min) do not appear to be explained by intersubject differences in the fraction of
inhaled O3 retained in the lung (Ultman et al., 2004).
6.2.3.3 Comparison of 2 h IE to 6.6 h O3 Exposure Effects on Pulmonary Function
Adams (2003b) examined whether prolonged 6.6-h QCE exposure to a relatively low O3
concentration (0.08 ppm) and a 2-h IE exposure at a relatively high O3 concentration (0.30 ppm)
elicited consistent individual subject FEVj responses. Individual subject O3 exposure
reproducibility was first examined via a regression plot of the postexposure FEVj response to the
6.6-h chamber exposure as a function of postexposure FEVj response to the 2-h IE chamber
exposure. The resulting R2 of 0.40, although statistically significant, was substantially less than
that observed in a comparison of individual FEVj response to the two 2-h IE exposures by
6-7
-------
chamber and face mask, respectively (R2 = 0.83). The Spearman rank order correlation for the
chamber 6.6-h and chamber 2-h exposure comparison was also substantially less (0.49) than that
obtained for the two 2-h IE exposures (0.85). The primary reason for the greater variability in
the chamber 6.6-h exposure FEVj response as a function of that observed for the two 2-h IE
exposures is very likely related to the increased variability in response upon repeated exposure
to O3 concentrations lower than 0.18 ppm (R = 0.57, compared to a mean R of 0.82 at higher
concentrations) reported by McDonnell et al. (1985). This rationale is supported by the lower r2
(0.40) observed by Adams (2003b) for the FEVj responses found in 6.6 h chamber and face
mask exposures to 0.08 ppm O3, compared to an r2 of 0.83 observed for responses found at 0.30
ppm O3.
6.2.4 Triangular Ozone Exposures
To further explore the factors that determine responsiveness to O3, Hazucha et al. (1992)
designed a protocol to examine the effect of varying, rather than constant, O3 concentrations.
Subjects were exposed to an O3 level that increased linearly from 0 to 0.24 ppm for the first 4 h
and then decreased linearly from 0.24 to 0 ppm over the second 4 h of the 8 h exposure
(triangular concentration profile) and to a constant level exposure of 0.12 ppm O3 for 8 h. While
total inhaled O3 doses for the constant and the triangular concentration profile were almost
identical, the FEVj response was dissimilar. For the constant 0.12 ppm O3 exposure, FEVj
declined -5% by the fifth hour and then remained at that level. With the triangular O3
concentration profile, there was minimal FEVj response over the first 3 h followed by a rapid
decrease in FEVj (-10.3%) over the next 3 h. During the seventh and eighth hours, mean FEVj
decrements improved to -6.3% as the O3 concentration decreased from 0.12 to 0.00 ppm
(mean = 0.06 ppm). (The reader is referred to Figure AX6-3for illustration ofFEV, responses
during square-wave and triangular O3 exposures).
Adams (2003a) used a less abrupt triangular O3 exposure profile (ranging from 0.03 to
0.15 ppm) having an average exposure concentration of only 0.08 ppm, assumed to be typical of
outdoor ambient conditions. Postexposure values for FEVj and symptoms were not significantly
different between the 6.6 h triangular and a square-wave 0.08 ppm O3 exposure. During the
triangular exposure, however, FEVj responses became statistically significant after 4.6 h,
6-8
-------
whereas they were not significant until 6.6 h during the square wave exposure (Adams, 2003a).
Perhaps due to the lower O3 concentrations, evidence of FEVj response recovery with the
triangular exposure was less pronounced than as observed by Hazucha et al. (1992). Figure 6-1
illustrates the average O3-induced FEVj responses and the O3 exposure schemes for the Adams
(2003a) and Hazucha et al. (1992) studies.
0.00
Time
Figure 6-1. Ozone-induced changes in FEVt (top panel) and O3 concentration profiles
(bottom panel) as a function of exposure duration. Open (o) and closed (•)
circles illustrate average data from Hazucha et al. (1992) and Adams (2003a),
respectively. For clarification, the "O3-induced changes in FEVt" are the FEVt
responses following O3 exposure minus the FEVt responses following FA.
6-9
-------
In a very recent chamber study, Adams (2006) employed the earlier design of square-wave
and triangular exposure profiles. The O3 concentrations for the 6.6-h exposures were 0.00, 0.06,
and 0.08 ppm for square-wave profiles and averaged 0.04, 0.06, and 0.08 ppm for triangular
profiles. The 6.6-h postexposure responses from the 0.08 ppm O3 (average) triangular exposure
did not differ significantly from those observed in the 0.08 ppm O3 square-wave exposure.
However, FEVj and symptoms were significantly different from preexposure at 4.6 h (when
the O3 concentration was 0.15 ppm) in the triangular exposure, but not until 6.6 h in the square-
wave exposure. These observations have confirmed and expanded findings of the earlier study
(Adams, 2003a) showing a clear divergence in the peak and hourly responses in FEVj between
square-wave and triangular exposure profiles at the 0.08 ppm level. However, at the lower O3
concentration of 0.06 ppm, no temporal pattern differences in FEVj responses between square-
wave and triangular exposure profiles could be discerned. The author concluded that the results
support previous evidence that O3 concentration has a greater singular effect in the total inhaled
O3 dose than do VE and exposure duration. The author observed no significant differences in
FEVj and symptom responses compared to FA for the 0.04 and 0.06 ppm exposures.
With square-wave O3 exposures between 0.08 to 0.12 ppm, FEVj decrements may increase
with time of exposure (and O3 dose) or reach a plateau (Horstman et al., 1990; McDonnell et al.,
1991). For the triangular exposures used by Hazucha et al. (1992) and Adams (2003a, 2006),
maximal FEVj responses occurred 1 h to 2 h after peak O3 concentration and 1 h to 2 h before
the maximal O3 dose occurred (at the end of the O3 exposure). These three studies suggest that a
triangular exposure profile can potentially lead to higher FEVj responses than square-wave
profiles at overall equivalent O3 doses.
6.2.5 Mechanisms of Pulmonary Function Responses
Inhalation of O3 for several hours while physically active elicits both subjective respiratory
tract symptoms and acute pathophysiologic changes. The typical symptomatic responses
consistently reported in studies are that of airway irritation, cough, and pain on deep inspiration.
Depending on the individual's responsiveness to O3, this can be accompanied by several
pathophysiologic changes, e.g., decrements in lung capacities and volumes, bronchoconstriction,
airway hyperresponsiveness, airway inflammation, immune system activation, and epithelial
injury. The severity of symptoms and the magnitude of response depend on inhaled dose, O3
6-10
-------
sensitivity of an individual, and the extent of tolerance resulting from the individual's previous
exposures. The development of effects is time-dependent during both exposure and recovery
periods, with considerable overlap of evolving and receding effects. The time sequence,
magnitude and the type of responses of this complex series of events, both in terms of
development and recovery, indicate that several mechanisms, activated at different times of
exposure, must contribute to the overall lung function response (U.S. Environmental Protection
Agency, 1996).
Available information on recovery from O3 exposure indicates that an initial phase of
recovery proceeds relatively rapidly, and some 40 to 65% of the acute spirometric and symptom
response appears to occur within about 2 h (Folinsbee and Hazucha, 1989). Following a 2-h
exposure to 0.4 ppm O3 with IE, Nightingale et al. (2000) observed a 13.5% decrement in FEVj.
By 3 h postexposure; however, only a 2.7% FEVj decrement persisted, as illustrated in
Figure 6-2. A similar postexposure recovery in FVC was also observed. Gerrity et al. (1993)
suggested that, for healthy young adults, transient increases in mucus clearance (mediated by
cholinergic receptors) due to O3 exposure may be coincident to pulmonary function responses,
i.e., the transient increases in clearance and decrements in lung function return to baseline values
within 2 to 3 h postexposure. However, there is some indication that the spirometric responses,
especially at higher O3 concentrations, are not fully recovered within 24 h (Folinsbee and
Horvath, 1986; Folinsbee et al., 1998). In hyperresponsive individuals, the recovery takes
longer, as much as 48 h, to return to baseline values. Collectively, these observations suggest
that there is a rapid recovery of O3-induced spirometric responses and symptoms, which may
occur during resting exposure to O3 (Folinsbee et al., 1977) or as O3 concentration is reduced
during exposure (Hazucha et al., 1992), and a slower phase, which may take at least 24 h to
complete (Folinsbee and Hazucha, 2000). Repeated exposure studies at higher concentrations
typically show that FEVj response to O3 is enhanced on the second of several days of exposure
(Table AX6-8). This enhanced response suggests a residual effect of the previous exposure,
about 22 h earlier, even though the preexposure spirometry may be the same as on the previous
day. The absence of the enhanced response with repeated exposure at lower O3 concentrations
may be the result of a more complete recovery or less damage to pulmonary tissues (Folinsbee
etal., 1994).
6-11
-------
4.0n
£• 3.0-
LU
2.0-
Pre 0 2 4
Time from Exposure (hours)
24
Figure 6-2. Recovery of FEVt responses following a 2 h exposure to 0.4 ppm O3 with IE.
Note that the 2 h exposure period is indicated by the Pre to 0 h interval.
Immediately postexposure, FEVt was significantly (**p < 0.001) decreased.
At 3 h postexposure, FEVt was at 97% of the preexposure value.
Adapted from Nightingale et al. (2000).
6.2.5.1 Pathophysiologic Mechanisms
Breathing pattern changes
Human studies consistently report that inhalation of O3 alters the breathing pattern without
significantly affecting minute ventilation. A progressive decrease in tidal volume and a
"compensatory" increase in frequency of breathing to maintain steady minute ventilation during
exposure suggests a direct modulation of ventilatory control. These changes parallel a response
of many animal species exposed to O3 and other lower airway irritants (Tepper et al., 1990).
Bronchial C-fibers and rapidly adapting receptors appear to be the primary vagal afferents
responsible for O3-induced changes in ventilatory rate and depth in both humans (Folinsbee and
Hazucha, 2000) and animals (Coleridge et al., 1993; Hazucha and Sant'Ambrogio, 1993;
Schelegle et al., 1993).
6-12
-------
The potential modulation of breathing pattern by activation of sensory afferents located in
extrathoracic airways by O3 has not yet been studied in humans. Nasal-only O3 exposure of rats
produces changes in breathing pattern that are similar to changes observed in humans (Kleinman
etal., 1999).
Symptoms and lung function changes
As discussed, in addition to changes in ventilatory control, O3 inhalation by humans can
also induce a variety of symptoms, reduce vital capacity (VC) and related functional measures,
and increase airway resistance.
Schelegle et al. (2001) demonstrated that the reduction in VC due to O3 exposure is a reflex
action and not a voluntary termination of inspiration as result of discomfort. They reported
that O3-induced symptom responses (mediated in part by bronchial C-fibers) are substantially
reduced by inhaled topical anesthetic. However, the anesthetic had a minor and irregular effect
on pulmonary function decrements and tachypnea. Since respiratory symptom responses were
largely abolished, these findings support reflex inhibition of VC due to stimulation of both
bronchial and pulmonary C-fibers.
The involvement of nociceptive bronchial C-fibers modulated by opioid receptors
in limiting maximal inspiration and eliciting subjective symptoms in humans was studied
by Passannante et al. (1998). Sufentanil (an opioid agonist and analgesic) rapidly
reversed O3-induced symptom responses and reduced spirometric decrements in "strong"
responders. The incomplete recovery in FEVj following sufentanil administration, however,
suggests involvement of non-opioid receptor modulated mechanisms as well. Interestingly,
naloxone (an opioid receptor antagonist) had no significant effect on FEVj decrements in "weak"
responders. Plasma levels of p-endorphin (a potent pain suppressor) were not related to
O3 responses.
Airway hyperreactivity
In addition to limitation of maximal inspiration and its effects on other spirometric
endpoints, activation of airway sensory afferents also plays a role in receptor-mediated
bronchoconstriction and an increase in airway resistance. Despite this common mechanism,
post-O3 pulmonary function changes and either early or late airway hyperresponsiveness (AHR)
6-13
-------
to inhaled aerosolized methacholine or histamine are poorly correlated either in time or
magnitude. Fentanyl and indomethacin, the drugs that have been shown to attenuate O3-induced
lung function decrements in humans, did not prevent induction of AHR when administered to
guinea pigs prior to O3 exposure (Yeadon et al., 1992). Neither does post-O3 AHR seem to be
related to baseline airway responsiveness. These findings imply that the mechanisms are either
not related or are activated independently in time. Animal studies (with limited support from
human studies) have suggested that an early post-O3 AHR is, at least in part, vagally mediated
(Freed, 1996) and that stimulation of C-fibers can lead to increased responsiveness of bronchial
smooth muscle independently of systemic and inflammatory changes which may be even absent
(load et al., 1996). An in vitro study of isolated human bronchi has shown that O3-induced
airway sensitization involves changes in smooth muscle excitation-contraction coupling
(Marthan, 1996). Characteristic O3-induced inflammatory airway neutrophilia, which at one
time was considered a leading AHR mechanism, has been found in a murine model to be only
coincidentally associated with AHR, i.e., there was no cause and effect relationship (Zhang et al.,
1995). However, this observation does not rule out possible involvement of other cells (such as
eosinophils or T-helper cells) in AHR modulation. There is some evidence that release of
inflammatory mediators by these cells can sustain AHR and bronchoconstriction. In vitro and
animal studies have also suggested that airway neutral endopeptidase activity can be a strong
modulator of AHR (Marthan et al., 1996; Yeadon et al., 1992). Late AHR observed in some
studies is plausibly due to sustained damage of the airway epithelium and continual release of
inflammatory mediators (Foster et al., 2000). Thus, O3-induced AHR appears to be a product of
multiple mechanisms acting at different time periods and levels of the bronchial smooth muscle
signaling pathways (effects ofO3 on AHR are described in Section 6.8).
6.2.5.2 Mechanisms at a Cellular and Molecular Level
Stimulation of vagal afferents by O3 and reactive products, the primary mechanism of lung
function impairment, is enhanced and sustained by what can be considered in this context to be
secondary mechanisms activated at a cellular and molecular level. The complexity of these
mechanisms is beyond the scope of this section, and the reader is directed to Section 6.9 of this
chapter for more detail. A comprehensive review by Mudway and Kelly (2000) discusses the
cellular and molecular mechanisms of O3-induced pulmonary response in great detail.
6-14
-------
Stimulation of bronchial C-fibers by O3 not only inhibits maximal inspiration but, through
local axon reflexes, induces neurogenic inflammation. This pathophysiologic process is
characterized by release of tachykinins and other proinflammatory neuropeptides. Ozone
exposure has been shown to elevate the C-fiber-associated tachykinin, substance P, in human
bronchial lavage fluid (Hazbun et al. 1993) and to deplete neuropeptides synthesized and
released from C-fibers in human airway epithelium rich in substance P-immunoreactive axons.
Substance P and other transmitters are known to induce granulocyte adhesion and subsequent
transposition into the airways, increase vascular permeability and plasma protein extravasation,
cause bronchoconstriction, and promote mucus secretion (Solway and Left, 1991). Although the
initial pathways of neurogenic, antigen-induced, and innate immune-mediated inflammation are
not the same, they eventually converge, leading to further amplification of airway inflammatory
processes by subsequent release of cytokines, eicosanoids, and other mediators. Significantly
negative correlations between O3-induced leukotriene (LTC4/D4/E4) production and spirometric
decrements (Hazucha et al., 1996) and an increased level of postexposure PGE2, a mediator
known to stimulate bronchial C-fibers, show that these mediators play an important role in lung
function impairment due to O3 exposure (Mohammed et al., 1993; Hazucha et al., 1996). The
reported post O3 exposure dysfunction of small airways assessed by decrement in FEF25_75
(Weinman et al., 1995; Frank et al., 2001) is likely induced by both neurogenic and
inflammatory mediators, since the density of bronchial C-fibers is much lower in the small than
large airways. Also, because of the relative slowness of inflammatory responses as compared to
reflex effects, O3-triggered inflammatory mechanisms are unlikely to initially contribute to
progressive lung function reduction. It is plausible, however, that when fully activated, they
sustain and possibly further aggravate already impaired lung function. Indeed, a prolonged
recovery of residual spirometric decrements following the initial rapid improvement after
exposure termination could be due to slowly resolving airway inflammation. Bronchial
biopsies performed 6 h postexposure have shown that O3 caused a significant decrease in
immunoreactivity to substance P in the submucosa (Krishna et al., 1997). A strong negative
correlation with FEVj also suggests that the release of substance P may be an important
contributing mechanism to persistent post-O3 bronchoconstriction (Krishna et al., 1997).
Persistent spirometric changes observed for up to 48 h postexposure could plausibly be
6-15
-------
sustained by the inflammatory mediators, many of which have bronchoconstrictive properties
(Blomberg et al., 1999).
6.3 SUBJECTS WITH PREEXISTING DISEASE
Individuals with respiratory disease are of primary concern in evaluating the health effects
of O3 because even a small change in function is likely to have more impact on a person with
reduced reserve, i.e., O3-induced effects are superimposed on preexisting pulmonary impairment.
6.3.1 Subjects with Chronic Obstructive Pulmonary Disease
For patients with COPD performing light to moderate IE, no decrements in pulmonary
function were observed after 1- and 2-h exposures to <0.30 ppm O3 (Kehrl et al., 1985; Linn
et al., 1982a, 1983a; Solic et al., 1982), and only small decreases in forced expiratory volume
were observed for 3-h exposures of chronic bronchitics to 0.41 ppm O3 (Kulle et al., 1984).
More recently, Gong et al. (1997a) found no significant difference in response between age-
matched controls and COPD patients to a 4 h exposure to 0.24 ppm O3 with IE. Although the
clinical significance is uncertain, small transient decreases in arterial blood oxygen saturation
have also been observed in some of these studies.
6.3.2 Subjects with Asthma
Based on studies reviewed in the 1996 O3 AQCD (U.S. Environmental Protection Agency,
1996), asthmatic subjects appear to be at least as sensitive to acute effects of O3 as healthy
nonasthmatic subjects.
Several recent studies support a tendency for slightly increased spirometric responses in
mild asthmatic versus healthy subjects. Alexis et al. (2000) reported reductions in FVC (12%,
10%) and FEVj (13%, 11%) for 13 mild asthmatic and 9 healthy subjects, respectively, exposed
to 0.4 ppm O3 for 2 h with IE ( VE = 30 L/min). The FVC and FEVj responses were attenuated
by indomethacin in the healthy subjects, but not in the asthmatic subjects. As assessed by the
magnitude of reductions in mid-flows (viz. FEF25, FEF50, FEF60p, FEF75) following O3 exposure,
the small airways tended to be more affected in the asthmatic than in the healthy subjects. In a
6-16
-------
larger study, Torres et al. (1996) exposed 24 asthmatics, 12 allergic rhinitics, and 10 healthy
subjects to 0.25 ppm O3 for 3 h with IE ( VE = 30 L/min). The O3-induced FEVj decrements
tended to be greater in the diseased populations (allergic rhinitis, 14.1%; asthmatic, 12.5%;
healthy controls, 10.2%). Also, Scannell et al. (1996) exposed 18 asthmatics to 0.2 ppm O3 for
4 h with IE ( VE « 25 L/min/m2 BSA). An O3-induced increase in sRaw tended to be greater in
the asthmatics compared to 81 healthy subjects who underwent similar experimental protocols
(Aris et al., 1995; Balmes et al., 1996).
Similar O3-induced spirometric responses of asthmatic and healthy subjects are suggested
by some studies. The Scannell et al. (1996) study of 18 asthmatics reported FEVj and FVC
decrements that were similar to 81 healthy subjects (Aris et al., 1995; Balmes et al., 1996).
Similar group decrements in FEVj and FVC were also reported by Hiltermann et al. (1995) for
6 asthmatic and 6 healthy subjects exposed to 0.4 ppm O3 for 2 h with light IE. Basha et al.
(1994) also reported similar spirometric responses between 5 asthmatic and 5 healthy subjects
exposed to 0.2 ppm O3 for 6 h with IE. The lack of significant differences in the Hiltermann
et al. (1995) and Basha et al. (1994) studies is not very compelling, however, given the
extremely small sample sizes and corresponding lack of statistical power. The Basha et al.
(1994) study was also confounded by the asthmatics having an average preexposure FEVj
that was about 430 mL lower (a 12% difference) on the O3-exposure day relative to the
FA-exposure day.
One other study has reported results suggesting that some asthmatics may have smaller
O3-induced FEVj decrements relative to healthy subjects (3% versus 8%, respectively) when
exposed to 0.2 ppm O3 for 2 h with IE (Mudway et al., 2001). However, the asthmatics in the
Mudway et al. (2001) study also tended to be older than the healthy subjects, which could
partially explain their lesser response.
In a longer exposure duration (7.6 h) study, Horstman et al. (1995) exposed 17 mild-to-
moderate asthmatics and 13 healthy controls to 0.16 ppm O3 or FA with quasi continuous
exercise (VE «30 L/min). The FEVj decrement observed in the asthmatics was significantly
larger than in the healthy subjects (19% versus 10%, respectively). There was also a notable
tendency for a greater O3-induced decrease in FEF25.75 in asthmatics relative to the healthy
subjects (24% versus 15%, respectively). A significant positive correlation in asthmatics was
6-17
-------
also reported between O3-induced spirometric responses and baseline lung function, i.e.,
responses increased with severity of disease.
With repeated O3 exposures, asthmatic subjects, like healthy subjects (see Section 6.6),
develop tolerance. Gong et al. (1997b) exposed 10 asthmatics to 0.4 ppm O3, 3 h per day with
IE (VE -32 L/min), for 5 consecutive days. Symptom and spirometric responses were greatest
on the first (-35 % FEVj) and second (-34 % FEVj) exposure days, and progressively
diminished toward baseline levels (-6 % FEVj) by the fifth exposure day. Similar to healthy
subjects, asthmatics lose their tolerance to O3 after 4 to 7 days without O3 exposure.
Some, but not all, studies have reported that asthmatics have a somewhat exaggerated
inflammatory response to acute O3 exposure relative to healthy control subjects (e.g., McBride
et al., 1994; Basha et al., 1994; Peden et al., 1995, 1997; Peden, 2001a; Scannell et al., 1996;
Hiltermann et al., 1997, 1999; Michelson et al., 1999; Vagaggini et al., 1999; Newson et al.,
2000; Holz et al., 2002) see also Section 6.9 and Tables AX6-3 andAX6-12). For example, at
18-h post-O3 exposure (0.2 ppm, 4-h with IE, VE = 25 L/min/m2 BSA) and corrected for FA
responses, Scannell et al. (1996) found significantly increased neutrophils in 18 asthmatics
(12%) compared to 20 healthy subjects (4.5%). This inflammatory response difference was
observed despite no group differences in spirometric responses to O3. Inflammatory responses
do not appear to be correlated with lung function responses in either asthmatic or healthy
subjects (Balmes et al., 1996, 1997; Holz et al., 1999). The lack of correlation between
inflammatory and spirometric responses may be due to differences in the time kinetics of these
different types of responses (Stenfors et al., 2002). In addition, airway responsiveness to
inhaled allergens is increased by O3 exposure in subjects with allergic asthma for up to 24 h
(see Section 6.8}.
6.3.3 Subjects with Allergic Rhinitis
Allergic rhinitis is a condition defined by inflammation of the nasal membranes. Nayak
(2003) recently reviewed the commonalities between asthma and allergic rhinitis. Clinically,
greater than 60% of asthmatics have allergic rhinitis and slightly less than 40% of allergic
rhinitics have asthma. Leukotrienes and histamine are well-recognized mediators of responses
(viz., inflammation, hyperresponsiveness, and bronchoconstriction) in both asthma and allergic
6-18
-------
rhinitis. Although, rhinitis and asthma are distinguished as affecting the upper and lower
airways, respectively, it has been suggested that these diseases are manifestations of a similar
disease process.
Given the prevalence of concomitant asthma and rhinitis and their common response
mediators, it should be expected that allergic rhinitics might respond more similarly to
asthmatics than healthy individuals. Regarding spirometric responses, Torres et al. (1996)
provide the only data demonstrating a trend in support of this supposition.
Studies demonstrating the interaction between air pollutants and allergic processes in the
human nasal airways and rhinoconjunctival tissue have been reviewed by Peden (200Ib) and
Riediker et al. (2001), respectively. Ozone exposure of subjects with allergic rhinitis has been
shown to induce nasal inflammation and increase airway responsiveness to nonspecific
bronchoconstrictors.
Peden et al. (1995), who studied allergic asthmatics exposed to O3 found that O3 causes an
increased response to nasal allergen challenge in addition to nasal inflammatory responses.
Their data suggested that allergic subjects have an increased immediate response to allergen
after O3 exposure. In a follow-up study, Michelson et al. (1999) reported that 0.4 ppm O3 did not
promote early-phase-response mediator release or enhance the response to allergen challenge in
the nasal airways of mild, asymptomatic dust mite-sensitive asthmatic subjects. Ozone did,
however, promote an inflammatory cell influx, which helps induce a more significant late-phase
response in this population.
Torres et al. (1996) found that O3 causes an increased response to bronchial allergen
challenge in subjects with allergic rhinitis. This study also measured responses in healthy
subjects and mildly allergic asthmatics {see Sections AX6.3.2 andAX6.8). All subjects were
exposed to 0.25 ppm O3 for 3 h with IE (VE = 30 L/min). Statistically significant O3-induced
decrements in FEVj occurred in rhinitics (14.1%), asthmatics (12.5%), and the healthy controls
(10.2%), but these responses did not differ statistically between groups. Methacholine
responsiveness was significantly increased in asthmatics, but not in subjects with allergic
rhinitis. Airway responsiveness to an individual's historical allergen (grass and birch pollen,
house dust mite, or animal dander) was significantly increased after O3 exposure when compared
6-19
-------
to FA exposure. The authors concluded that subjects with allergic rhinitis, but without asthma,
could be at risk if a high O3 exposure is followed by a high dose of allergen.
Holz et al. (2002) extended the results of Torres et al. (1996) by demonstrating that
repeated daily exposure to lower concentrations of O3 (0.125 ppm for 4 days) causes an
increased response to bronchial allergen challenge in subjects with preexisting allergic airway
disease, with or without asthma. These investigators observed no major difference in the pattern
of bronchial allergen response between asthmatics or rhinitics, except that the dose of allergen
required to elicit a similar response (>20% decrease in FEVj) was significantly greater for
rhinitic than asthmatic subjects (16 vs. 0.6 mg/mL). Early phase responses were more consistent
in subjects with rhinitis, and late-phase responses were more pronounced in subjects with
asthma. There also was a tendency towards a greater effect of O3 in subjects with greater
baseline response to specific allergens (chosen on the basis of skin prick test and history, viz.,
grass, rye, birch, or alder pollen, house dust mite, or animal dander). These data suggest that the
presence of allergic bronchial sensitization, but not a history of asthma, may be a key
determinant of increased airway allergen responsiveness following exposure to O3 (for a more
complete discussion of airway responsiveness, see Section AX6.8).
6.3.4 Subjects with Cardiovascular Disease
Possibly due to the age of subjects studied, O3 exposure does not appear to result in
significant pulmonary function impairment or evidence of cardiovascular strain in patients with
cardiovascular disease relative to healthy controls. Gong et al. (1998) exposed 10 hypertensive
and 6 healthy adult males, 41 to 78 years of age, to 0.3 ppm O3 for 3 h with IE at 30 L/min. For
all subjects combined (no significant group differences), there was an O3-induced decrement of
7% in FEVj and an 70% increase in the alveolar-arterial oxygen tension gradient. The overall
results did not indicate any major acute cardiovascular effects of O3 in either the hypertensive or
normal subjects. Gong et al. (1998) suggested that by impairing alveolar-arterial oxygen
transfer, the O3 exposure could potentially lead to adverse cardiac events by decreasing oxygen
supply to the myocardium. However, the subjects in their study apparently had sufficient
functional reserve so as to not experience significant ECG changes or myocardial ischemia
and/or injury (see Section 6.10 for additional discussion).
6-20
-------
6.4 INTERSUBJECT VARIABILITY AND REPRODUCIBILITY
OF RESPONSE
Analysis of factors that contribute to intersubject variability is important for the
understanding of individual responses, mechanisms of response, and health risks associated with
acute O3 exposures. Large intersubject variability in response to O3 has been reported by
numerous investigators (Adams et al., 1981; Aris et al., 1995; Folinsbee et al., 1978; Kulle et al.,
1985; McDonnell et al., 1983). The magnitude of individual variability in FEVj response in 2 h
IE exposures increases at higher O3 concentrations (Kulle et al., 1985; McDonnell et al., 1983).
McDonnell (1996) examined the FEVj response data from three 6.6-h exposure studies
conducted at the EPA Health Effects Research Laboratory and showed that the FEVj responses
in FA were small, with most tightly grouped around zero. With increasing O3 concentrations
between 0.08 and 0.12 ppm, the mean response became asymmetrical, with a few individuals
experiencing quite large decrements in FEVj (intersubject variability observed in O3 dosimetry
studies is discussed in Chapter 4, Section 4.2).
As an example of the variation in spirometric responses to O3 exposure, Hazucha et al.
(2003) analyzed the distribution of O3 responsiveness in 240 subjects (18 to 60 years of age)
exposed to 0.42 ppm O3 (on 3 occasions) for 1.5 h with IE at VE = 20 L/min/m2 BSA. Across
all ages, 18% of subjects were weak responders (<5% FEVj decrement), 39% were moderate
responders, and 43% were strong responders (> 15% FEVj decrement). Younger subjects
(<35 years of age) were predominately strong responders, whereas older subjects (>35 years of
age) were mainly weak responders. The influence of age on intersubject variability was also
noted by Passannante et al. (1998), who found that subjects under 35 years of age were more like
to be strong responders than older individuals.
For repeated exposures, Hazucha et al. (2003) reported that the reproducibility of FEVj
responses was related to the length of time between exposures. A Spearman correlation
coefficient of 0.54 was found between responses for exposures separated by 105 days (median),
whereas a correlation coefficient of 0.85 was found between responses for exposures separated
by only 7 days (median). The more reproducible the subject's response, the more precisely it
indicates his/her intrinsic responsiveness. In 2 h IE O3 exposures, McDonnell et al. (1985) found
a relatively poor FEVj reproducibility (R = 0.58) at the lowest concentration, 0.12 ppm, due in
6-21
-------
part to a lack of specific O3 response or a uniformly small response in the majority of subjects.
It was concluded that, for 2 h IE O3 exposures equal to or greater than 0.18 ppm, the intersubject
differences in magnitude of change in FVC and FEVj are quite reproducible over time (21 to
385 days; mean = 33 days) and are due primarily to differences in intrinsic responsiveness of
individual subjects to O3 exposure.
Intersubject variability, mechanisms of response, and health risks associated with acute O3
exposures are complicated by weak associations between various O3-induced responses. In a
retrospective analysis of data for 485 male subjects (aged 18 to 36 yrs) exposed to one of six O3
concentrations at one of three activity levels for 2 h, McDonnell et al. (1999) found significant,
but low, Spearman rank order correlations between FEVj response and symptoms of cough
(R = 0.39), shortness of breath (R = 0.41), and pain on deep inspiration (R = 0.30). These
authors concluded that these different responses are mechanistically related to some degree, but
indicated that there does not appear to be a single factor which is responsible for the observed
individual differences in O3 responsiveness across the spectrum of symptom and lung function
responses.
The effect of large intersubject variability on the ability to predict individual
responsiveness to O3 was demonstrated by McDonnell et al. (1993). These investigators
analyzed data for 290 male subjects (18 to 32 yrs of age) who underwent repeat 2 h IE exposures
to one or more O3 concentrations ranging from 0.12 to 0.40 ppm. They attempted to identify
personal characteristics (i.e., age, height, baseline pulmonary function, presence of allergies, and
past smoking history) that might predict individual differences in FEVj response. Only age
contributed significantly to intersubject responsiveness (younger subjects were more
responsive), accounting for just 4% of the observed variance. Interestingly, O3 concentration
accounted for only 31% of the variance, strongly suggesting the importance of as yet undefined
individual characteristics that determine FEVj responsiveness to O3. The authors concluded that
much individual variability in FEVj response to O3 remains unexplained.
6-22
-------
6.5 FACTORS MODIFYING RESPONSIVENESS TO OZONE
6.5.1 Influence of Age
Children, adolescents, and young adults (<18 yrs of age) appear, on average, to have nearly
equivalent spirometric responses to O3, but have greater responses than middle-aged and older
adults when exposed to comparable O3 doses (U.S. Environmental Protection Agency, 1996).
Symptomatic responses to O3 exposure, however, appear to increase with age until early
adulthood and then gradually decrease with increasing age (U.S. Environmental Protection
Agency, 1996). In contrast to young adults, the diminished symptomatic responses in children
and the elderly may put the latter groups at increased risk for continued O3 exposure. Although
no new laboratory studies investigating O3 responses in children have been published since the
1996 O3 AQCD, the epidemiological studies published during the last decade (see Section
7.2.3.1 for details) are generally in agreement with the earlier laboratory studies.
The ensuing discussion in this section provides information on average FEVj responses to
O3 exposure as a function of age in healthy adults ranging from 18 to 50 years of age. Beyond
approximately 18 years of age, spirometric and symptom responses to O3 exposure begin to
decline with increasing age. In healthy individuals, the rate of decline in O3 responsiveness
appears to be greater in younger (18 to 35 yrs) versus middle aged (35 to 55 yrs) individuals
(Passannante et al., 1998; Hazucha et al., 2003). Beyond this age (>55 yrs), acute O3 exposure
elicits minimal spirometric changes. An average FEVj decrement of-3% has been reported by
Gong et al. (1997a) for this older population under a "worst case" exposure scenario (0.24 ppm
O3 with 4 h IE). Although Gong et al. (1997a) and others have examined responses to O3
exposure in subjects of various ages, the exposure conditions differ between most studies in such
a manner so that age effects remain uncertain.
Three recent studies, which analyzed large data sets (>240 subjects) of similarly exposed
subjects, show clearly discernable changes in FEVj responses to O3 as a function of age. In one,
Seal et al. (1996) analyzed O3-induced spirometric responses in 371 young nonsmokers (aged
18 to 35 yrs) exposed for 2.3 h during IE at a VE of 25 L/min/m2 BSA. On average, for the same
O3 concentration (C), the response of 25-, 30-, and 35-yr old individuals are predicted to be 83,
65, and 48%, respectively, of the response in a 20-yr old. For example, a 5.4% decrement in
FEVj is predicted for a 20-yr old exposed to 0.12 ppm O3 for 2.3 h IE (VE = 25 L/min/m2 BSA),
whereas, a similarly exposed 35-yr old is predicted to have only a 2.6% decrement.
6-23
-------
McDonnell et al. (1997) examined FEVj responses in 485 healthy white males (aged 18 to
36 yrs) exposed once for 2 h to an O3 concentration of 0.0, 0.12, 0.18, 0.24, 0.30, or 0.40 ppm at
rest or one of two levels of IE (VE of 25 and 35 L/min/m2 BSA). For the same exposure
conditions (C, VE, and duration), the average responses of 25-, 30-, and 35-yr old individuals are
predicted to be 69, 48, and 33%, respectively, of the response in 20-yr olds. Hazucha et al.
(2003) analyzed the distribution of O3 responsiveness in 240 subjects (18 to 60 years of age)
exposed to 0.42 ppm O3 for 1.5 h with IE at VE = 20 L/min/m2 BSA. In males, the FEVj
responses of 25-, 35-, and 50-yr olds are predicted to be 94, 83, and 50%, respectively, of the
average response in 20-yr old males. In females, the FEVj responses of 25-, 35-, and 50-yr olds
are predicted to be 82, 46, and 18%, respectively, of the average response in 20-yr old females.
For subjects aged 18 to 36 yrs, McDonnell et al. (1999) recently reported that symptom
responses from O3 exposure also decrease with increasing age. Whether the same age-dependent
pattern of O3 sensitivity decline also holds for airway reactivity or inflammatory endpoints has
not been determined.
6.5.2 Gender and Hormonal Influences
Several studies have suggested that physiological differences between the genders may
predispose females to a greater susceptibility to O3. Housley et al. (1996) reported that females
have lower concentration of uric acid (the most prevalent antioxidant) in nasal lavage fluid than
males. Since the source of uric acid (the plasma) is also known to be lower in females, the
authors suggested that this could be a contributing factor leading to gender differences in
response to oxidants. Plausibly, reduced absorption of O3 in the upper airways may promote its
deeper penetration into the lower respiratory system regions. Dosimetric measurements have
shown that the absorption distribution of O3 is independent of gender when absorption is
normalized to anatomical dead space (Bush et al., 1996). More recently, Ultman et al. (2004)
reported that the whole lung uptake fraction of O3 was significantly greater in males (91.4%)
than females (87.1%). But, this increase in O3 uptake in the males was consistent with their
larger tidal volume and slower breathing frequency relative to the females. Furthermore, O3
uptake was not correlated with spirometric responses. Thus, a differential removal of O3 by uric
6-24
-------
acid seems to have minimal effect. In general, the spirometric responses of healthy young
females to O3 exposure appears comparable to the responses of young males (Hazucha et al.,
2003). Although, during the follicular phase of the menstrual cycle, lung function response to O3
is enhanced (Fox et al., 1993).
6.5.3 Racial, Ethnic, and Socioeconomic Status Factors
In the only laboratory study designed to compare spirometric responses of whites and
blacks exposed to a range of O3 concentrations (0 to 0.4 ppm), Seal et al. (1993) reported
inconsistent and statistically insignificant FEVj differences between white and black males and
females within various exposure levels. Perhaps with larger cohorts, the tendency for greater
responses of black than white males may become significant. Thus, based on this study it is still
unclear if race is a modifier of O3 sensitivity, although the findings of epidemiologic studies
reported in the previous, 1996 O3 AQCD "can be considered suggestive of an ethnic difference"
(U.S. Environmental Protection Agency, 1996).
Gwynn and Thurston (2001) pointed out, it appears that it is more the socioeconomic status
(SES) and overall quality of healthcare that drives PM10- and O3-related hospital admissions than
an innate or acquired sensitivity to pollutants. This assertion appears to be supported somewhat
by the study of Seal et al. (1996), who employed a family history questionnaire to examine the
influence of SES on the O3 responsiveness. Seal et al. found that, of the three SES categories,
individuals in the middle SES category showed greater concentration-dependent decline in %
predicted FEVj (4 to 5% at 0.4 ppm O3) than low and high SES groups. The authors did not
have an "immediately clear" explanation for this finding. With such a paucity of data, it is not
possible to discern the influence of racial or economic-related factors on O3 sensitivity.
6.5.4 Influence of Physical Activity
Any physical activity will increase minute ventilation and therefore the dose of inhaled O3.
Consequently, the intensity of physiological response following an acute exposure will be
strongly associated with minute ventilation (see Figures 6-3 andAX6-2).
6-25
-------
Time (h)
Figure 6-3. Predicted O3-induced decrements in FEVt as a function of exposure duration
and level of IE (line labels are VE levels) in healthy young adults (20 yrs of
age) exposed to 0.3 ppm O3. The illustrated activity levels range from rest
• •
(VE = 10 L/min) to moderate exercise (VE = 40 L/min). Predictions are for
Model 1 coefficients in Table 3 of McDonnell et al. (1997).
Source: Based on McDonnell et al. (1997).
6.5.5 Environmental Factors
Since the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996) few human
laboratory studies have examined the potential influence of environmental factors (e.g., rural
versus urban environment, passive cigarette smoke exposure, and bioactive agents such as
endotoxin) on O3-induced pulmonary function changes in healthy individuals.
New controlled human exposure studies have confirmed that smokers are less responsive
to O3 than nonsmokers. Spirometric and plethysmographic pulmonary function decline,
nonspecific airway hyperreactivity, and inflammatory response of smokers to O3 were all weaker
than those reported for nonsmokers. Although all these responses are intrinsically related, the
functional association between them, as in nonsmokers, has been weak. Similarly, the time
course of development and recovery of these effects, as well their reproducibility, was not
different from nonsmokers. Chronic airway inflammation with desensitization of bronchial
6-26
-------
nerve endings and an increased production of mucus may plausibly explain the smaller responses
to O3 observed in smokers (Frampton et al., 1997; Torres et al., 1997).
The effect of environmental tobacco smoke (ETS) on O3 responses has received very little
attention. In one study (Yu et al., 2002), preexposure of mice to sidestream cigarette smoke
(ETS surrogate) elicited no immediate effects, but potentiated subsequent O3-induced
inflammatory responses (see Chapter 5.4.2 for additional ETS details). Endotoxin is a
biologically active component of both mainstream and sidestream tobacco smoke (Hasday et al.,
1999), which might contribute to the potentiation of O3 effects.
The influence of ambient temperature on pulmonary effects induced by O3 exposure in
humans has been studied infrequently under controlled laboratory conditions. Several
experimental human studies have reported additive effects of heat and O3 exposure (see U.S.
Environmental Protection Agency, 1986, 1996). Foster et al. (2000) exposed 9 young healthy
subjects for 130 min (IE 10 min at 36 to 39 1/min) to filtered air and to ramp profile O3 at 22 °C
and 30 °C, 45-55% RH. The O3 exposure started at 0.12 ppm, reached a peak of 0.24 ppm
midway through, and subsequently declined to 0.12 ppm at the end of exposure. At the end of
exposure, FEVj had decreased significantly (p < 0.5) by -8% at 22 °C and -6.5% at 30 °C.
One day (19 h) later, the decline of 2.3% from baseline was still significant (p < 0.05) at both
temperatures. FVC decrements were smaller and significant only for the 22 °C condition
immediately postexposure. There was a decline in specific airway conductance (sGaw; p < 0.05)
at 30 °C, but not at 22 °C. Nonspecific bronchial responsiveness to methacloline assessed
as PC50 sGaw was significantly (p < 0.05) higher one day following O3 exposure at both
temperatures, but more so at 30 °C. Thus, these findings suggest that elevated temperature may
partially attenuate spirometric responses but enhance airway reactivity.
6.5.6 Oxidant-Antioxidant Balance
The first line of defense against oxidative stress involves antioxidants present in epithelial
lining fluid (ELF), which scavenge free radicals and limit lipid peroxidation. Exposure to O3
depletes the antioxidant level in nasal ELF, probably due to scrubbing of O3 (Mudway et al.,
1999); however, the concentration and the activity of antioxidant enzymes either in ELF or
plasma do not appear to be related to O3 responsiveness (Avissar et al., 2000; Blomberg et al.,
1999; Samet et al., 2001). Carefully controlled studies of dietary antioxidant supplementation
6-27
-------
have shown some protective effects of a-tocopherol and ascorbate on spirometric lung function
from O3 but not on the intensity of subjective symptoms and inflammatory response, including
cell recruitment, activation and release of mediators (Samet et al., 2001; Trenga et al., 2001).
Dietary antioxidants have also afforded partial protection to asthmatics by attenuating
postexposure bronchial hyperresponsiveness (Trenga et al., 2001). Field studies performed in
Mexico City (described in Section 7.2.3.1) and animal studies (described in Section 5.2.1.3) have
also demonstrated protective effects of ELF antioxidants during O3 exposures.
6.5.7 Genetic Factors
Several recent studies (Bergamaschi et al., 2001; Corradi et al., 2002; David et al., 2003;
Romieu et al., 2004; Yang et al., 2005) have reported that genetic polymorphism of antioxidant
enzymes and inflammatory genes may modulate pulmonary function and inflammatory response
to O3 challenge. It has been suggested that healthy carriers of NAD(P)H:quinone oxidoreductase
wild type (NQOlwt) in combination with glutathione S-transferase |i-l genetic deficiency
(GSTMlnull) are more responsive to O3 (Bergamaschi et al., 2001). The authors reported that
subjects having NQOlwt and GSTMlnull genotypes had increased O3 responsiveness (FEVj
decrements and epithelial permeability), whereas subjects with other combinations of these
genotypes were less affected. A subsequent study from the same laboratory reported a positive
association between O3 responsiveness, as characterized by the level of oxidative stress and
inflammatory mediators (8-isoprostane, LTB4 and TEARS) in exhaled breath condensate and the
NQOlwt and GSTMlnull genotypes (Corradi et al., 2002). However, none of the spirometric
lung function endpoints were affected by O3 exposure. It is of interest to note that human nasal
mucosa biopsies of GSTMlnull subjects showed higher antioxidant enzymes activity than
biopsies of GSTM1 positive individuals when exposed to ozone (Otto-Knapp et al., 2003).
Asthmatic children with a genetic deficiency of GSTM1 were reported to be more
responsive to ambient O3 exposure than asthmatic children without this deficiency, as assessed
by decrements in FEF25.75, in this field study (Romieu et al., 2004). Antioxidant supplementation
(vit. C and E) attenuated post-O3 exposure lung function response in these children. More
specific genotyping has shown that O3 responsiveness of asthmatic children may be related to the
presence of variant Ser allele for NQO1. The presence of at least one NQO1 Ser allele in
6-28
-------
combination with GSTM1 null genotype lowered the risk of asthma in O3-exposed asthmatic
children relative to Pro/Pro genotype (David et al., 2003).
The influence of functional polymorphism in TNF-a, lymphotoxin- a (LTA), TLR4, SOD2
and GPX1 genes on ozone-induced lung function changes in healthy individuals, mild asthmatics
and subjects with rhinitis was varied. Of the inflammatory genes studied, only TNF-a has
appeared to show some promise as one of the genetic factors of susceptibility. However, as the
authors stated, "the functional significance of individual TNF-a polymorphisms remains
controversial" (Yang et al., 2005).
These recent studies have shown that an individual's innate susceptibility to O3 may be
linked to the genetic background of an individual. Although a number of potential O3
susceptibility genes have been identified, additional better designed and controlled studies are
needed to ascertain the link between susceptibility and specific polymorphism.
6.6 REPEATED O3 EXPOSURE EFFECTS
Based on studies reviewed here and in the previous O3 criteria documents (U.S.
Environmental Protection Agency, 1986, 1996), several conclusions can be drawn about
repeated 1- to 2-h O3 exposures. Repeated exposures to O3 can cause an enhanced (i.e., greater)
pulmonary function response on the second day of exposure (see Tables AX6-8 andAX6-9for
added detail). This enhancement appears to be dependent on the interval between the exposures
(24 h is associated with the greatest increase) and is absent with intervals greater than 3 days
(Bedi et al., 1985; Folinsbee and Horvath, 1986; Schonfeld et al., 1989). An enhanced response
also appears to depend to some extent on the magnitude of the initial response (Horvath et al.,
1981). Small responses to the first O3 exposure are less likely to result in an enhanced response
on the second day of O3 exposure (Folinsbee et al., 1994). With continued daily exposures (i.e.,
beyond the second day) there is an attenuation of pulmonary function responses, typically after
3 to 5 days of repeated O3 exposure. This attenuation of responses is lost in 1 week (Kulle et al.,
1982; Linn et al., 1982b) or perhaps 2 weeks (Horvath et al., 1981) without O3 exposure.
In temporal conjunction with pulmonary function changes, symptoms induced by O3 (e.g.,
cough, pain on deep inspiration, and chest discomfort), are increased on the second exposure day
and attenuated with repeated O3 exposure thereafter (Folinsbee et al., 1980, 1998; Foxcroft and
6-29
-------
Adams, 1986; Linn et al., 1982b). O3-induced changes in airway responsiveness persist longer
and attenuate more slowly than pulmonary function and symptoms responses (Dimeo et al.,
1981; Kulle et al., 1982), although this has been studied only on a limited basis (Folinsbee et al.,
1994). In longer-duration (4 h to 6.6 h), lower-concentration studies that do not cause an
enhanced second-day response, the attenuation of response to O3 appears to proceed more
rapidly (Folinsbee et al., 1994) {Effects of repeated exposures on inflammatory responses are
discussed in Section 6.9.4).
6.7 EFFECTS ON EXERCISE PERFORMANCE
The effects of acute O3 inhalation on endurance exercise performance have been examined
in numerous controlled laboratory studies. These studies were discussed in the 1996 O3 AQCD
(U.S. Environmental Protection Agency, 1996) and can be divided into two categories: (1) those
that examined the effects of acute O3 inhalation on maximal oxygen uptake (VO2max) and
(2) those that examined the effects of acute O3 inhalation on the ability to complete strenuous
continuous exercise protocols of up to 1 h in duration.
In brief, endurance exercise performance and VO2max may be limited by acute exposure
to O3 (Adams and Schelegle, 1983; Schelegle and Adams, 1986; Gong et al., 1986; Foxcroft and
Adams, 1986; Folinsbee et al., 1977; Linder et al., 1988). Gong et al. (1986) and Schelegle and
Adams (1986) found that significant reductions in maximal endurance exercise performance may
occur in well-conditioned athletes while they perform CE (VE > 80 L/min) for 1 h at O3
concentrations >0.18 ppm. Reports from studies of exposure to O3 during high-intensity
exercise indicate that breathing discomfort associated with maximal ventilation may be an
important factor in limiting exercise performance in some, but not all, subjects.
6.8 EFFECTS ON AIRWAY RESPONSIVENESS
Airway hyperresponsiveness refers to a condition in which the propensity for the airways
to bronchoconstrict due to a variety of stimuli becomes augmented. Airway responsiveness is
typically quantified by measuring the decrement in pulmonary function (i.e., spirometry or
6-30
-------
plethysmography) following inhalation of small amounts of an aerosolized specific (antigen,
allergen) or nonspecific (methacholine, histamine) bronchoconstrictor agent or some other
measured stimulus (e.g., exercise, cold air).
Ozone exposure causes an increase in nonspecific airway responsiveness as indicated by a
reduction in the concentration of methacholine or histamine required to produce a given
reduction in FEVj or increase in SRaw. Increased airway responsiveness is an important
consequence of exposure to O3, because its presence means that the airways are predisposed to
narrowing upon inhalation of a variety of stimuli (e.g., specific allergens, SO2, cold air).
Ozone exposure of asthmatic subjects, who characteristically have increased airway
responsiveness at baseline, can cause further increases in responsiveness (Kreit et al., 1989).
Similar relative changes in airway responsiveness are seen in asthmatics exposed to O3 despite
their markedly different baseline airway responsiveness. Several studies (Torres et al., 1996;
Kehrl et al., 1999; Molfino et al., 1991) have been published suggesting an increase in specific
(i.e., allergen-induced) airway reactivity in response to O3 exposure. An important aspect of
increased airway responsiveness after O3 exposure is that this represents a plausible link between
ambient O3 exposure and increased hospital admissions for asthma.
Changes in airway responsiveness after O3 exposure appear to be resolved more slowly
than changes in FEVj or respiratory symptoms (Folinsbee and Hazucha, 2000). Furthermore, in
studies of repeated exposure to O3, changes in airway responsiveness tend to be somewhat less
susceptible to attenuation with consecutive exposures than changes in FEVj (Dimeo et al., 1981;
Folinsbee et al., 1994; Gong et al., 1997b; Kulle et al., 1982). Ozone-induced increases in
airway responsiveness do not appear to be strongly associated with decrements in lung function
or increases in symptoms.
Mechanisms underlying O3-induced increases in airway responsiveness are only partially
understood, but such increases appear to be associated with a number of cellular and biochemical
changes in airway tissue. Although inflammation could play a role in the increase in airway
responsiveness, cyclooxygenase inhibitors have not been effective at blocking the O3-induced
influx of PMNs into bronchoalveolar lavage (BAL) fluid (Hazucha et al., 1996; Ying et al.,
1990). Therefore, O3-induced airway responsiveness may not be due to the presence of PMNs
in the airway or to the release of arachidonic acid metabolites. Rather, it seems likely that the
mechanism for this response is multifactorial, possibly involving the presence of cytokines,
6-31
-------
prostanoids, or neuropeptides; activation of macrophages, eosinophils, or mast cells; and
epithelial damage that increases direct access of mediators to the smooth muscle or receptors in
the airways that are responsible for reflex bronchoconstriction.
6.9 EFFECTS ON INFLAMMATION AND HOST DEFENSE
6.9.1 Introduction
Short-term exposure of humans to O3 can cause acute inflammation, and long-term
exposure of laboratory animals results in a chronic inflammatory state (see Chapter 5). The
relationship between repetitive bouts of acute inflammation in humans caused by O3 and the
development of chronic respiratory disease is unknown.
The presence of neutrophils (PMNs) in the lung has long been accepted as a hallmark of
inflammation and is an important indicator that O3 causes inflammation in the lungs. It is
apparent, however, that inflammation within airway tissues may persist beyond the point that
inflammatory cells are found in BAL fluid (BALF). Soluble mediators of inflammation such as
the cytokines (IL-6, IL-8) and arachidonic acid metabolites (e.g., PGE2, PGF2a, thromboxane,
and leukotrienes [LTs] such as LTB4) have been measured in the BAL fluid of humans exposed
to O3. In addition to their role in inflammation, many of these compounds have
bronchoconstrictive properties and may be involved in increased airway responsiveness
following O3 exposure.
Some recent evidence suggests that changes in small airways function may provide a
sensitive indicator of O3 exposure and effect, despite the fact that inherent variability in their
measurement by standard spirometric approaches makes their assessment difficult (Frank et al.,
2001). Observations of increased functional responsiveness of these areas relative to the more
central airways, and of persistent effects following repeated exposure, may indicate that further
investigation of inflammatory processes in these regions is warranted.
6.9.2 Inflammatory Responses in the Upper Respiratory Tract
The nasal passages constitute the primary portal for inspired air at rest and, therefore,
the first region of the respiratory tract to come in contact with airborne pollutants. Nikasinovic
et al. (2003) recently reviewed the literature of laboratory-based nasal inflammatory studies
6-32
-------
published since 1985. Nasal lavage (NL) has provided a useful tool for assessing O3-induced
inflammation in the nasopharynx. Increased levels of PMNs in the NL fluid of humans exposed
to 0.5 ppm O3 at rest for 4 h has been reported (Graham et al., 1988; Bascom et al., 1990).
Graham and Koren (1990) compared inflammatory mediators present in both the NL and
BAL fluids of humans exposed to 0.4 ppm O3 for 2 h. Similar increases in PMN were observed
in NL and BAL, suggesting a qualitative correlation between inflammatory changes in the lower
airways (BAL) and the upper respiratory tract (NL). Torres et al. (1997) compared NL and BAL
in smokers and nonsmokers exposed to 0.22 ppm O3 for 4 h. In contrast to Graham and Koren
(1990), they did not find a relationship between numbers or percentages of PMNs in the nose
and the lung, perhaps in part due to the variability observed in their NL recoveries. Albumin,
a marker of epithelial cell permeability, was increased 18 h later, but not immediately after
exposure, as seen by Bascom et al. (1990).
McBride et al. (1994) reported that asthmatic subjects were more sensitive than
nonasthmatics to upper airway inflammation at an O3 concentration (0.24 ppm for 1.5 h with
light IE) that did not affect pulmonary function. In the asthmatics, there was a significant
increase in the number of PMNs in NL fluid both immediately and 24 h after exposure. Peden
et al. (1995) also found that exposure to 0.4 ppm O3 had a direct nasal inflammatory effect and a
priming effect on response to nasal allergen challenge. A subsequent study in dust
mite-sensitive asthmatic subjects indicated that O3 at this concentration enhanced eosinophil
influx in response to allergen but did not promote the early-phase nasal response to allergen
(Michelson et al., 1999). Similar to observations made in the lower airways, the presence of O3
molecular "targets" in nasal lining fluid is likely to provide some level of local protection against
exposure. In a study of healthy subjects exposed to 0.2 ppm O3 for 2 h, Mudway and colleagues
(1999) observed a significant depletion of uric acid in NL fluid at 1.5 h following exposure.
6.9.3 Inflammatory Response in the Lower Respiratory Tract
As reviewed in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996), acute
exposure to O3 results in an inflammatory reaction, increased epithelial cell permeability, and
may stimulate fibrogenic processes. Inflammatory markers are observed in BALF of healthy
subjects by 1 h post-O3 exposure and may persist for at least 18 to 24 h. Not all inflammatory
6-33
-------
markers, however, follow the same time course. Studies published since the 1996 O3 AQCD
support these earlier findings.
Inflammatory effects have been assessed in vivo by lavage (proximal airway and
bronchoalveolar), bronchial biopsy, and more recently, induced sputum. In vitro studies of
human alveolar macrophages (AM) and airway epithelial cells exposed to O3 suggest that these
cell types release mediators found in the BALF of O3-exposed humans (U.S. Environmental
Protection Agency, 1996). Recent evidence suggests that the release of mediators from AMs
may be modulated by the products of O3-induced oxidation of airway lining fluid components,
such as human surfactant protein A (Wang et al., 2002).
Spirometric responses to O3 are independent of inflammatory responses and markers of
epithelial injury (Balmes et al., 1996; Blomberg et al., 1999; Hazucha et al., 1996; Torres et al.,
1997). Significant inflammatory responses to O3 exposures that did not elicit significant
spirometric responses have been reported (Holz et al., 2005; McBride et al., 1994).
A meta-analysis of 21 studies (Mudway and Kelly, 2004) showed that PMN influx in healthy
subjects is associated with total O3 dose (product of O3 concentration, exposure duration,
andVE).
The time course of the inflammatory response to O3 in humans has not been fully
characterized. From a review of the literature by Mudway and Kelly (2000), Figure 6-4
illustrates a plausible time course of acute O3 responses. As the figure shows, different markers
exhibit peak responses at different times. Studies in which lavages were performed 1 h after O3
exposure (1 h at 0.4 ppm or 4 h at 0.2 ppm) have demonstrated that the inflammatory responses
are quickly initiated (Devlin et al., 1996; Schelegle et al., 1991; Torres et al., 1997).
Inflammatory mediators and cytokines such as IL-8, IL-6, and PGE2 are greater at 1 h than at
18 h post-O3 exposure (Devlin et al., 1996; Torres et al., 1997). However, IL-8 still remain
elevated at 18 h post-O3 (4 h at 0.2 ppm O3 versus FA) in healthy subjects (Balmes et al., 1996).
Schelegle et al. (1991) found increased PMNs in the "proximal airway" lavage at 1, 6, and 24 h
after O3 exposure (4 h at 0.2 ppm O3), with a peak response at 6 h. Although, at 18 to 24 h after
O3 exposure, PMNs remain elevated relative to 1 h postexposure (Schelegle et al., 1991; Torres
et al., 1997). In addition to the influx of PMNs, lymphocyte numbers in BALF are elevated
significantly at 6 h following exposure (2 h at 0.2 ppm O3) of healthy subjects (Blomberg et al.,
1997). Flow cytometry also indicated the increased presence of CD3+, CD4+ and CD8+ T cell
6-34
-------
B
C
D
Initiation Progression
Resolution
-»• FEVi
Total celts
Macrophages
^ Neutrophils
>• Lymphocytes
Shed epithelial cells
Albumin/total protein
LDII
LOP
Exposure
Post-Exposure
.,y
,v
Figure 6-4. Time course of acute responses seen in humans exposed to O3. Responses are
divided into three phases: initiation, during O3 exposure; progression, where
responses develop post-O3 exposure; and resolution, during which responses
return to baseline levels. *The IL-8 response is shown, with a progressive
increase peaking at 18 h postexposure (18h-PE). The IL-8 literature is
inconclusive and neutrophils have been demonstrated in the absence of an
IL-8 increase. **Few studies have measured MPO; however, a trend for an
increase at 6h-PE has been reported.
Abbreviations: AA (ascorbic acid), FN (fibronectin), GSH (reduced
glutathione), IL (interleukin), LDH (lactate dehydrogenase), LOP (lipid
ozonation products), LTs4 (leukotriene B4), LTc4 (leukotriene C4), MPO
(myeloperoxidase), PGE2 (prostaglandin E2), UA (uric acid).
Source: Reprinted from Mudway and Kelly (2000) with permission from Elsevier.
6-35
-------
subsets. This same laboratory later demonstrated that within 1.5 h following exposure of healthy
subjects to the same O3 regimen, expression of human leukocyte antigen (HLA)-DR on lavaged
macrophages underwent a significant, 2.5-fold increase (Blomberg et al., 1999).
The inflammatory responses to O3 exposure have also been studied in asthmatic subjects
(Basha et al., 1994; Scannell et al., 1996; Peden et al., 1997). In these studies, asthmatics
showed significantly more neutrophils in the BALF (18 h postexposure) than similarly exposed
healthy individuals. In one of these studies (Peden et al., 1997), which included only allergic
asthmatics who tested positive for Dematophagoides farinae antigen, there was an eosinophilic
inflammation (2-fold increase), as well as neutrophilic inflammation (3-fold increase). In a
study of subjects with intermittent asthma exposed to 0.4 ppm O3 for 2 h, increases in eosinophil
cationic protein, neutrophil elastase and IL-8 were found to be significantly increased 16 h
post-exposure and comparable in induced sputum and BALF (Hiltermann et al, 1999). Scannell
et al. (1996) also reported that IL-8 tends to be higher in the BALF of asthmatics compared to
nonasthmatics following O3 exposure, suggesting a possible mediator for the significantly
increased neutrophilic inflammation in those subjects. Bosson et al. (2003) found significantly
greater epithelial expression of IL-5, IL-8, granulocyte-macrophage colony-stimulating factor
(GM-CSF) and epithelial cell-derived neutrophil-activating peptide 78 (ENA-78) in asthmatics
compared to healthy subjects following exposure to 0.2 ppm O3 for 2 h. In contrast, Stenfors
et al. (2002) did not detect a difference in the O3-induced increases in neutrophil numbers
between 15 mild asthmatic and 15 healthy subjects by bronchial wash at the 6 h postexposure
time point. However, the asthmatics were on average 5 years older than the healthy subjects in
this study, and it is not yet known how age affects inflammatory responses. It is also possible
that the time course of neutrophil influx differs between healthy and asthmatic individuals.
Vagaggini et al. (2002) investigated the effect of prior allergen challenge on responses in
mild asthmatics exposed for 2 h to 0.27 ppm O3 or filtered air. At 6 h postexposure, eosinophil
numbers in induced sputum were found to be significantly greater after O3 than after air
exposures. Studies such as these suggest that the time course of eosinophil and neutrophil influx
following O3 exposure can occur at levels detectable within the airway lumen by as early as 6 h.
They also suggest that the previous or concurrent activation of proinflammatory pathways within
the airway epithelium may enhance the inflammatory effects of O3. For example, in an in vitro
6-36
-------
study of epithelial cells from the upper and lower respiratory tract, cytokine production induced
by rhinovirus infection was enhanced synergistically by concurrent exposure to O3 at 0.2 ppm for
3 h (Spannhake et al, 2002).
Although the release of mediators has been demonstrated to occur at exposure times and
concentrations that are minimally cytotoxic to airway cells, potentially detrimental latent effects
have been demonstrated in the absence of cytotoxicity. These include the generation of DNA
single strand breaks (Kozumbo et al., 1996), the loss of cellular replicative activity (Gabrielson
et al., 1994) in bronchial epithelial cells exposed in vitro, and the formation of protein and
DNA adducts. A highly toxic aldehyde formed during O3-induced lipid peroxidation is 4-
hydroxynonenal (HNE). Healthy human subjects exposed to 0.4 ppm O3 for 1 h underwent BAL
6 h later. Analysis of lavaged AMs by Western Blot indicated increased levels of a 32-kDa
HNE-protein adduct, as well as 72-kDa heat shock protein and ferritin in O3-versus air-exposed
subjects (Hamilton et al., 1998). In a recent study of healthy subjects exposed to 0.1 ppm O3 for
2 h (Corradi et al., 2002), formation of 8-hydroxy-2'-deoxyguanosine (8-OHdG), a biomarker
of reactive oxidant species (ROS)-DNA interaction, was measured in peripheral blood
lymphocytes. At 18 h post exposure, 8-OHdG was significantly increased in cells compared to
preexposure levels, presumably linked to concurrent increases in chemical markers of ROS. Of
interest, the increase in 8-OHdG was only significant in a subgroup of subjects with the wild
genotype for NQO1 and the null genotype for GSTM1, suggesting that polymorphisms in
antioxidant enzymes may confer "susceptibility' to O3 in some individuals.
The generation of ROS following exposure to O3 has been shown to be associated with a
wide range of responses. In a recent study, ROS production by alveolar macrophages lavaged
from subjects exposed to 0.22 ppm for 4 h was assessed by flow cytometry (Voter et al., 2001).
Levels were found to be significantly elevated 18 h postexposure and to be associated with
several markers of increased permeability. An in vitro study of human tracheal epithelial cells
exposed to O3 indicated that generation of ROS resulted in decreased synthesis of the
bronchodilatory prostaglandin, PGE2, as a result of inactivation of prostaglandin endoperoxide
G/H synthase 2 (Alpert et al., 1997).
6-37
-------
6.9.4 Adaptation of Inflammatory Responses
Physiologic and symptomatic responses in humans following repeated exposure to O3 were
discussed in Section 6.6. Inflammatory responses upon repeated O3 exposures are discussed in
this section. Animal studies suggest that while inflammation may be diminished with repeated
exposure, underlying damage to lung epithelial cells continues (Tepper et al., 1989). Markers
from HALF following both 2-h (Devlin et al., 1997) and 4-h (Christian et al., 1998; Torres et al.,
2000) repeated O3 exposures (up to 5 days) indicate that there is ongoing cellular damage
irrespective of the attenuation of some cellular inflammatory responses of the airways,
pulmonary function, and symptom responses.
Devlin et al. (1997) examined the inflammatory responces of humans repeatedly exposed
to 0.4 ppm O3 for 5 consecutive days. Several indicators of inflammation (e.g., PMN influx,
IL-6, PGE2, fibronectin) were attenuated after 5 days of exposure (i.e., values were not different
from FA). Several markers (LDH, IL-8, total protein, epithelial cells) did not show attenuation,
indicating that tissue damage probably continues to occur during repeated exposure. The
recovery of the inflammatory response occurred for some markers after 10 days, but some
responses were not normalized even after 20 days. The continued presence of cellular injury
markers indicates a persistent effect that may not necessarily be recognized due to the
attenuation of spirometric and symptom responses.
Christian et al. (1998) randomly subjected heathy subjects to a single exposure and to
4 consecutive days of exposure to 0.2 ppm O3 for 4 h. Both "bronchial" and "alveolar" fractions
of the BAL showed decreased numbers of PMNs and fibronectin concentration at day 4 versus
the single exposure, and a decrease in IL-6 levels in the alveolar fraction. Following a similar
study design and exposure parameters, Torres et al. (2000) found both functional and BALF
cellular responses to O3 were abolished at 24 h postexposure following the fourth exposure day.
However, levels of total protein, IL-6, IL-8, reduced glutathione and ortho-tyrosine were still
increased significantly. In addition, visual scores for bronchitis, erythema and the numbers of
neutrophils in the mucosal biopsies were increased. Their results indicate that, despite reduction
of some markers of inflammation in BALF and measures of large airway function, inflammation
within the airways persists following repeated exposure to O3.
6-38
-------
Holz et al. (2002) made a comparison of early and late responses to allergen challenge
following O3 in subjects with allergic rhinitis or allergic asthma. With some variation, both early
and late FEVj and cellular responses in the two subject groups were significantly enhanced by
4 consecutive days of exposure to 0.125 ppm O3 for 3 h.
In another study, Frank and colleagues (2001) exposed healthy subjects to FA and to O3
(0.25 ppm, 2 h) on 4 consecutive days each, with pulmonary function measurements being made
prior to and following each exposure. BAL was performed on day 5, 24 h following the last
exposure. On day 5, PMN numbers remained significantly higher following O3 compared to FA.
Of particular note in this study was the observation that small airway function, assessed by
grouping values for isovolumetric FEF25.75, VmaxSO and Vmax75 into a single value, showed
persistent reduction from day 2 through day 5. These data suggest that techniques monitoring
the function in the small peripheral airway regions (the primary sites of O3 uptake in the lung),
may provide important information regarding both acute and cumulative effects of O3 exposure.
6.9.5 Effect of Anti-Inflammatory and Other Mitigating Agents
Pretreatment of healthy subjects with non-steroidal anti-inflammatory drugs (ibuprofen,
etc.) has been found to partially suppress development of airway inflammation and pulmonary
function changes (U.S. Environmental Protection Agency, 1996). Although atropine blocked the
increase in Raw in response to O3 exposure, it did not alter the spirometric or symptom
responses (Beckett et al., 1985). Similarly, albuterol and salbutamol, which had no effect
on O3-induced changes in spirometry, also had no effect of symptom responses (McKenzie et al.,
1987; Gong et al., 1988). The anti-inflammatory medications indomethacin and ibuprofen,
which partially inhibit the spirometric responses to O3 exposure, also cause a reduction in
respiratory symptoms (Schelegle et al., 1987; Hazucha et al., 1994). Indomethacin attenuates
decrements in FEVj and FVC in healthy subjects, but not asthmatics (Alexis et al., 2000).
In contrast, inhalation of the corticosteroid budesonide does not prevent or even
attenuate O3-induced responses in healthy subjects as assessed by measurements of lung
function, bronchial reactivity and airway inflammation (Nightingale et al., 2000). In asthmatic
subjects, budesonide decreases airway neutrophil influx following O3 exposure (Vagaggini
etal., 2001).
6-39
-------
Holz et al. (2005) studied inflammatory responses in healthy ozone-responders (>10%
increase in sputum neutrophils from O3) pretreated with single doses (the highest shown to be
safe and well tolerated) of inhaled fluticasone and oral prednisolone. The O3 exposure caused
small changes in FEVj (-3.6% ± 6.8%) that were not significantly different from baseline or
between treatment groups (i.e., prescreening, placebo, fluticasone, and prednisolone). Relative
to placebo, the inhaled or oral corticosteroids significantly reduced O3-induced neutrophil levels.
These authors noted that their study design was intended to test the anti-inflammatory effects of
the steroids and that such high-dose regimens should not be considered for potential long-term
patient treatment.
6.9.6 Changes in Host Defense Capability Following Ozone Exposures
A number of studies clearly show that a single acute exposure (1 to 4 h) of humans to
moderate concentrations of O3 (0.2 to 0.6 ppm) while exercising at moderate to heavy levels
results in a number of cellular and biochemical changes in the lung, including an inflammatory
response characterized by increased numbers of PMNs, increased permeability of the epithelial
cells lining the respiratory tract, cell damage, and production of proinflammatory cytokines and
prostaglandins. These responses can be detected as early as 1 h after exposure (Koren et al.,
1991; Schelegle et al., 1991) and can persist for at least 18 h (Aris et al., 1993; Koren et al.,
1989). The response profile of these mediators is not defined adequately, although it is clear that
the time course of response varies for different mediators and cells (Devlin et al., 1997;
Schelegle et al., 1991). Following a 2 h exposure to 0.2 ppm O3, Blomberg et al. (2003)
observed increased Clara cell protein (biomarker of epithelial permeability) in blood at 2 h
postexposure, which remained high at 6 h and had returned to baseline by 18 h. These changes
also occur in humans exposed to 0.08 and 0.10 ppm O3 for 6 to 8 h (Devlin et al., 1991; Peden
et al., 1997). Ozone also causes inflammatory changes in the nose, as indicated by increased
levels of PMNs and albumin, a marker for increased epithelial cell permeability. Nasal lavage
analyses, however, are not necessarily parallel to BALF analyses.
There appears to be no strong correlation between any of the measured cellular and
biochemical changes and changes in lung function measurements, suggesting that different
mechanisms may be responsible for these processes (Balmes et al., 1996; Devlin et al., 1991).
The idea of different mechanisms is supported by a study in which ibuprofen, a cyclooxygenase
6-40
-------
inhibitor, blunted the O3-induced decrements in lung function without altering the O3-induced
increase in PMNs or epithelial cell permeability (Hazucha et al., 1996). In vitro studies suggest
that epithelial cells are the primary target of O3 in the lung and that O3 induces them to produce
many of the mediators found in the BALF of humans exposed to O3. Although O3 does not
induce AMs to produce these compounds in large quantities, it does directly impair the ability of
AMs to phagocytize and kill microorganisms.
A number of studies have found that O3 exposures increases epithelial cell permeability
through direct (technetium-99m labeled diethylene triamine pentaacetic acid, 99mTc-DTPA,
clearance) and indirect (e.g., increased BALF albumin, protein) techniques. Kehrl et al. (1987)
showed increased 99mTc-DTPA clearance in healthy young adults at 75 minutes postexposure
to 0.4 ppm O3 for 2 h. More recently, Foster and Stetkiewicz (1996) have shown that
increased 99mTc-DTPA clearance persists for at least 18-20 h post-O3 exposure (130 min to
average O3 concentration of 0.24 ppm), and the effect is greater at the lung apices than at the
base. Increased BALF protein, suggesting O3-induced changes in epithelial permeability, have
also been reported at 1 h and 18 h postexposure (Balmes et al., 1996; Devlin et al., 1997).
A recent meta-analysis of results from 21 publications (Mudway and Kelly, 2004), showed that
increased BALF protein is associated with total ozone dose (product of O3 concentration,
exposure duration, and VE ). Changes in permeability associated with acute inflammation may
provide increased access of inhaled antigens, particles, and other substances to the smooth
muscle, interstitial cells, and the blood.
In addition to affecting epithelial permeability and AM-mediated clearance in the
respiratory region of the lung, mucociliary clearance of the tracheobronchial airways is also
affected by O3 exposure. Only two studies (Foster et al., 1987; Gerrity et al., 1993) have
investigated the effect of O3 exposure on mucociliary particle clearance in humans. Foster et al.
(1987) measured clearance during and after a 2 h exposure to 0.4 ppm O3. Gerrity et al. (1993)
measured clearance at 2 h postexposure (0.4 ppm O3), by which time sRaw had returned to
baseline and FVC was within 5% of baseline (versus an 11% decrement immediately
postexposure). Foster et al. (1987) found a stimulatory effect of acute O3 exposure on
mucociliary clearance. Gerrity et al. (1993), who observed no effect on clearance, suggested that
transient clearance increases are coincident to pulmonary function responses. Investigators in
6-41
-------
both studies suggested that O3-induced increases in mucociliary clearance could be mediated by
cholinergic receptors.
6.10 EXTRAPULMONARY EFFECTS OF OZONE
Ozone reacts rapidly on contact with respiratory system tissue and is not absorbed or
transported to extrapulmonary sites to any significant degree as such. Human exposure studies
discussed in the previous O3 AQCDs (U.S. Environmental Protection Agency, 1986, 1996) failed
to demonstrate any consistent extrapulmonary effects. More recently, some human exposure
studies have attempted to identify specific markers of exposure to O3 in blood. Foster et al.
(1996) found a reduction in the serum levels of the free radical scavenger a-tocopherol after O3
exposure. Liu et al. (1997, 1999) used a salicylate metabolite, 2,3, dehydroxybenzoic acid
(DHBA), to indicate increased levels of hydroxyl radical which hydroxylates salicylate to
DHBA. Increased DHBA levels after exposure to 0.12 and 0.40 ppm suggest that O3 increases
production of hydroxyl radical. The levels of DHBA were correlated with spirometry changes.
Relative to preexposure, Corradi et al. (2002) observed increased levels of 8-OHdG (a biomarker
ROS-DNA interaction) in peripheral blood lymphocytes of healthy subjects at 18 h postexposure
to 0.1 ppmO3for2h.
Gong et al. (1998) observed a statistically significant O3-induced increase the alveolar-to-
arterial PO2 gradient in both healthy (n = 6) and hypertensive (n = 10) adult males (aged 41 to
78 yrs) exposed for 3 h with IE (VE « 30 L/min) to 0.3 ppm O3. The mechanism for the
decrease in arterial oxygen tension in the Gong et al. (1998) study could be due to an O3-induced
ventilation-perfusion mismatch. Foster et al. (1993) has demonstrated that even in relatively
young healthy adults (26.7 ± 7 yrs old), O3 exposure can cause ventilation to shift away from the
well perfused basal lung. This effect of O3 on ventilation distribution (and, by association, the
small airways) may persist beyond 24-h postexposure (Foster et al., 1997). Gong et al. (1998)
suggested that by impairing alveolar-arterial oxygen transfer, the O3 exposure could potentially
lead to adverse cardiac events by decreasing oxygen supply to the myocardium. The subjects in
the Gong et al. (1998) study had sufficient functional reserve so as to not experience significant
ECG changes or myocardial ischemia and/or injury.
6-42
-------
Effects of O3 exposure on alveolar-arterial oxygen gradients may be more pronounced in
patients with preexisting obstructive lung diseases. Relative to healthy elderly subjects, COPD
patients have reduced gas exchange and low SaO2. Any inflammatory or edematous responses
due to O3 delivered to the well-ventilated regions of the COPD lung could further inhibit gas
exchange and reduce oxygen saturation. In addition, O3-induced vasoconstriction could also
acutely induce pulmonary hypertension. Inducing pulmonary vasoconstriction and hypertension
in these patients would perhaps worsen their condition, especially if their right ventricular
function was already compromised.
6.11 EFFECTS OF OZONE MIXED WITH OTHER POLLUTANTS
Over the past 10 years only a handful of controlled-exposure human studies have examined
the effects of pollutant mixtures containing O3. The studies summarized in this section
complement the studies reviewed in the 1996 O3 AQCD (U.S. Environmental Protection
Agency, 1996). (The complexities ofO3 and co-pollutant exposures in animal studies are
discussed in Chapter 5, Section 5.4.4).
The results of a controlled study on children (Linn et al., 1997), designed to approximate
exposure conditions of an epidemiologic study (Neas et al., 1995) by matching the population
and exposure atmosphere (0.1 ppm O3, 0.1 ppm SO2 and 101 |ig/m2 h2SO4), did not support the
findings of this epidemiologic study. The study points out the difficulties in attempting to link
the outcomes of epidemiologic and controlled studies. Another vulnerable population,
asthmatics, demonstrated enhanced airway reactivity to house dust mite following exposures
to O3, NO2, and the combination of the two gases. Spirometric response, however, was impaired
only by O3, and O3 + NO2 at higher concentrations (Jenkins et al., 1999). Continuous exposure
to SO2 and NO2 increases inhaled bolus O3 absorption, while continuous exposure to O3
decreases O3 bolus absorption (Rigas et al., 1997). Inhalation of a mixture of PM25 and O3 by
healthy subjects increased brachial artery tone and reactivity (Brook et al., 2002). Since no other
cardiovascular endpoints were affected by the exposure, the pathophysiological importance of
this observation remains uncertain. However, acute pulmonary hypertension due to O3-induced
vasoconstriction could pose a risk to individuals with cardiovascular disease (see Section 6-10).
6-43
-------
All in all, the contention that air pollutant mixtures elicit stronger pathophysiologic effects
than individual pollutants of the mix is only weakly supported by human studies of either healthy
or at-risk individuals.
6.12 CONTROLLED STUDIES OF AMBIENT AIR EXPOSURES
A large amount of informative O3 exposure-effects data has been obtained in controlled
exposure studies of humans or laboratory animals under a variety of different experimental
conditions. However, laboratory simulation of the variable pollutant mixtures present in ambient
air is not practical. Thus, the exposure effects of one or several artificially generated pollutants
(i.e., a simple mixture) on pulmonary function and symptoms may not fully explain responses to
ambient air exposures where complex pollutant mixtures exist.
6.12.1 Mobile Laboratory Studies
Quantitatively useful information on the effects of acute exposure to photochemical
oxidants on pulmonary function responses and symptoms derived from field studies using a
mobile laboratory were presented in prior criteria documents (U.S. Environmental Protection
Agency, 1986, 1996). Relative to controlled exposure studies, mobile laboratory ambient air
studies suffer an additional limitation in terms of a dependence on outdoor ambient conditions.
Consistent with controlled exposure studies, mobile studies in California demonstrated that
pulmonary effects from exposure to ambient air in Los Angeles are related to O3 concentration
and level of exercise. Healthy subjects with a history of allergy also appeared to be more
responsive to O3 than "nonallergic" subjects (Linn et al., 1980, 1983b), although a standardized
evaluation of atopic status was not performed.
6.13 SUMMARY
Responses in humans exposed to ambient O3 concentrations include: decreased inspiratory
capacity; mild bronchoconstriction; rapid, shallow breathing pattern during exercise; and
symptoms of cough and pain on deep inspiration. Reflex inhibition of inspiration results in a
decrease in forced vital capacity (FVC) and, in combination with mild bronchoconstriction,
6-44
-------
contributes to a decrease in the forced expiratory volume in 1 s (FEVj). In addition to
physiological pulmonary responses and symptoms of breathing discomfort, O3 exposure also
results in airway hyperresponsiveness, inflammation, immune system activation, and epithelial
injury. With repeated O3 exposures over several days, spirometric and symptom responses
become attenuated, but this tolerance is lost after about a week without exposure. Airway
responsiveness also appears to be attenuated with repeated O3 exposures, but less so than FEVj.
Unlike spirometric and symptom responses, airway inflammation and small airways dysfunction
may not become attenuated by repeated O3 exposures.
Healthy young adults exposed to O3 concentrations >0.08 ppm develop significant
reversible, transient decrements in pulmonary function if minute ventilation (VE ) or duration of
exposure are increased sufficiently. The pattern of FEVj response appears to depend on the O3
exposure profile. Triangular exposure profiles can potentially lead to greater FEVj responses
than square wave exposures at equivalent average O3 doses. O3-induced decrements in FEVj do
not appear to depend on gender, race, body surface area, height, lung size, or baseline FVC in
healthy young adults. Healthy children experience similar spirometric responses but lesser
symptoms from O3 exposure relative to young adults. On average, spirometric and symptom
responses to O3 exposure appear to decline with increasing age beyond about 18 years of age.
There is a large degree of intersubject variability in physiologic and symptomatic responses of
heathy adults exposed to O3. However, responses tend to be reproducible within a given
individual over a period of several months. With increasing O3 concentration, the distribution of
FEVj decrements becomes asymmetrical, with a few individuals experiencing large decrements.
There is a tendency for slightly increased spirometric responses in mild asthmatics and
allergic rhinitics relative to healthy young adults. Spirometric responses in asthmatics appear to
be affected by baseline lung function, i.e., responses increase with disease severity. With
repeated daily O3 exposures, spirometric responses of asthmatics become attenuated; however,
airway responsiveness becomes increased in subjects with preexisting allergic airway disease
(with or without asthma). Possibly due to patient age, O3 exposure does not appear to cause
significant pulmonary function impairment or evidence of cardiovascular strain in patients with
cardiovascular disease or chronic obstructive pulmonary disease relative to healthy subjects.
Available information on recovery from O3 exposure indicates that an initial phase of
recovery in healthy individuals proceeds relatively rapidly, with acute spirometric and symptom
6-45
-------
responses resolving within about 2 to 4 h. Small residual lung function effects are almost
completely resolved within 24 h. Effects of O3 on the small airways, assessed by persistent
decrement in FEF25.75 and altered ventilation distribution, may be due in part to inflammation.
Indeed, a prolonged recovery of residual spirometric decrements following the initial rapid
recovery could be due to slowly resolving airway inflammation. In hyperresponsive individuals,
this recovery takes longer (as much as 48 hours) to return to baseline values. Persistent
spirometry changes observed for up to 48 h postexposure could plausibly be sustained by the
inflammatory mediators. Cellular responses (e.g., release of immunomodulatory cytokines)
appear to still be active as late as 20 h postexposure. More slowly developing inflammatory and
cellular changes may persist for up to 48 h, but the time course for these parameters in humans
has not been explored fully.
Soluble mediators of inflammation such as the cytokines (IL-6, IL-8) and arachidonic acid
metabolites (e.g., PGE2, PGF2(X, thromboxane, and leukotrienes [LTs] such as LTB4) have been
measured in the BAL fluid of humans exposed to O3. Many of these compounds have
bronchoconstrictive properties and may be involved in increased airway responsiveness
following O3 exposure. Some indicators of inflammation (e.g., PMN influx, IL-6, PGE2,
fibronectin) are attenuated with repeated O3 exposures. However, other markers (LDH, IL-8,
total protein, epithelial cells) do not show attenuation, thus indicating that tissue damage
probably continues to occur during repeated O3 exposure. There appears to be no strong
correlation between any of the measured cellular and biochemical changes and changes in lung
function measurements. A limited number of studies suggest that inflammatory responses may
be detected following O3 exposures that are insufficient to cause decrements in pulmonary
function. Whether airway reactivity or inflammatory responses to O3 are dependent on the age
of the exposed individual, such as spirometric responses, has not been determined.
Dietary antioxidant supplementation attenuates O3-induced spirometric responses, but not
the intensity of subjective symptoms nor inflammatory responses. Dietary antioxidants also
appear to afford partial protection to asthmatics by attenuating postexposure bronchial
hyperresponsiveness.
6-46
-------
REFERENCES
Adams, W. C. (1987) Effects of ozone exposure at ambient air pollution episode levels on exercise performance.
Sports Med. 4: 395-424.
Adams, W. C. (2000) Ozone dose-response effects of varied equivalent minute ventilation rates. J. Exposure Anal.
Environ. Epidemiol. 10: 217-226.
Adams, W. C. (2002) Comparison of chamber and face-mask 6.6-hour exposures to ozone on pulmonary function
and symptoms responses. Inhalation Toxicol. 14: 745-764.
Adams, W. C. (2003a) Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm ozone via
square-wave and triangular profiles on pulmonary responses. Inhalation Toxicol. 15: 265-281.
Adams, W. C. (2003b) Relation of pulmonary responses induced by 6.6-h exposures to 0.08 ppm ozone and 2-h
exposures to 0.30 ppm ozone via chamber and face-mask inhalation. Inhalation Toxicol. 15: 745-759.
Adams, W. C. (2006) Comparison of chamber 6.6-h exposures to 0.04-0.08 PPM ozone via square-wave and
triangular profiles on pulmonary responses. Inhalation Toxicol. 18: 127-136.
Adams, W. C.; Ollison, W. M. (1997) Effects of prolonged simulated ambient ozone dosing patterns on human
pulmonary function and symptomatology. Presented at: 90th annual meeting of the Air & Waste Management
Association; June; Toronto, Ontario, Canada. Pittsburgh, PA: Air & Waste Management Association; paper
no. 97-MP9.02.
Adams, W. C.; Schelegle, E. S. (1983) Ozone and high ventilation effects on pulmonary function and endurance
performance. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 55: 805-812.
Adams, W. C.; Savin, W. M.; Christo, A. E. (1981) Detection of ozone toxicity during continuous exercise via the
effective dose concept. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 51: 415-422.
Alexis, N.; Urch, B.; Tarlo, S.; Corey, P.; Pengelly, D.; O'Byrne, P.; Silverman, F. (2000) Cyclooxygenase
metabolites play a different role in ozone-induced pulmonary function decline in asthmatics compared to
normals. Inhalation Toxicol. 12: 1205-1224.
Alpert, S. E.; Walenga, R. W.; Jaspers, I.; Qu, Q.; Chen, L. C. (1997) Ozone inactivates Cyclooxygenase in human
tracheal epithelial cells without altering PGHS-2 mRNA or protein. Am. J. Physiol. 272: L879-L887.
Aris, R. M.; Christian, D.; Hearne, P. Q.; Kerr, K.; Finkbeiner, W. E.; Balmes, J. R. (1993) Ozone-induced airway
inflammation in human subjects as determined by airway lavage and biopsy. Am. Rev. Respir. Dis.
148: 1363-1372.
Aris, R. M.; Tager, L; Christian, D.; Kelly, T.; Balmes, J. R. (1995) Methacholine responsiveness is not associated
with O3-induced decreases inFEVj. Chest 107: 621-628.
Avissar, N. E.; Reed, C. K.; Cox, C.; Frampton, M. W.; Finkelstein, J. N. (2000) Ozone, but not nitrogen dioxide,
exposure decreases glutathione peroxidases in epithelial lining fluid of human lung. Am. J. Respir. Crit. Care
Med. 162: 1342-1347.
Balmes, J. R.; Chen, L. L.; Scannell, C.; Tager, L; Christian, D.; Hearne, P. Q.; Kelly, T.; Aris, R. M. (1996)
Ozone-induced decrements in FEVj and FVC do not correlate with measures of inflammation. Am. J. Respir.
Crit. Care Med. 153: 904-909.
Balmes, J. R.; Aris, R. M.; Chen, L. L.; Scannell, C.; Tager, I. B.; Finkbeiner, W.; Christian, D.; Kelly, T.; Hearne,
P. Q.; Ferrando, R.; Welch, B. (1997) Effects of ozone on normal and potentially sensitive human subjects.
part I: airway inflammation and responsiveness to ozone in normal and asthmatic subjects. Cambridge, MA:
Health Effects Institute. Research report no. 78; pp 1-37, 81-99.
Bascom, R.; Naclerio, R. M.; Fitzgerald, T. K.; Kagey-Sobotka, A.; Proud, D. (1990) Effect of ozone inhalation on
the response to nasal challenge with antigen of allergic subjects. Am. Rev. Respir. Dis. 142: 594-601.
Basha, M. A.; Gross, K. B.; Gwizdala, C. J.; Haidar, A. H.; Popovich, J., Jr. (1994) Bronchoalveolar lavage
neutrophilia in asthmatic and healthy volunteers after controlled exposure to ozone and filtered purified air.
Chest 106: 1757-1765.
Beckett, W. S.; McDonnell, W. F.; Horstman, D. H.; House, D. E. (1985) Role of the parasympathetic nervous
system in acute lung response to ozone. J. Appl. Physiol. 59: 1879-1885.
Bedi, J. F.; Drechsler-Parks, D. M.; Horvath, S. M. (1985) Duration of increased pulmonary function sensitivity to
an initial ozone exposure. Am. Ind. Hyg. Assoc. J. 46: 731-734.
Bergamaschi, E.; De Palma, G.; Mozzoni, P.; Vanni, S.; Vettori, M. V.; Broeckaert, F.; Bernard, A.; Mutti, A.
(2001) Polymorphism of quinone-metabolizing enzymes and susceptibility to ozone-induced acute effects.
Am. J. Respir. Crit. Care Med. 163: 1426-1431.
6-47
-------
Blomberg, A.; Helleday, R.; Pourazar, I; Stenfors, N.; Kelly, F. I; Frew, A. I; Holgate, S. T.; Sandstrom, T. (1997)
Early airway and peripheral blood cell responses to 0.20 ppm ozone in healthy human subjects. Eur. Respir. J.
10(suppl 25): 274S.
Blomberg, A.; Mudway, I. S.; Nordenhall, C.; Hedenstrom, H.; Kelly, F. J.; Frew, A. J.; Holgate, S. T.; Sandstrom,
T. (1999) Ozone-induced lung function decrements do not correlate with early airway inflammatory or
antioxidant responses. Eur. Respir. J. 13: 1418-1428.
Blomberg, A.; Mudway, I.; Svensson, M; Hagenbjork-Gustafsson, A.; Thomasson, L.; Helleday, R.; Dumont, X.;
Forsberg, B.; Nordberg, G.; Bernard, A. (2003) Clara cell protein as a biomarker for ozone-induced lung
injury in humans. Eur. Respir. J. 22: 883-888.
Bosson, J.; Stenfors, N.; Bucht, A.; Helleday, R.; Pourazar, J.; Holgate, S. T.; Kelly, F. J.; Sandstrom, T.; Wilson, S.;
Frew, A. J.; Blomberg, A. (2003) Ozone-induced bronchial epithelial cytokine expression differs between
healthy and asthmatic subjects. Clin. Exp. Allergy 33: 777-782.
Brook, R. D.; Brook, J. R.; Urch, B.; Vincent, R.; Rajagopalan, S.; Silverman, F. (2002) Inhalation of fine paniculate
air pollution and ozone causes acute arterial vasoconstriction in healthy adults. Circulation 105: 1534-1536.
Bush, M. L.; Asplund, P. T.; Miles, K. A.; Ben-Jebria, A.; Ultman, J. S. (1996) Longitudinal distribution of O3
absorption in the lung: gender differences and intersubject variability. J. Appl. Physiol. 81: 1651-1657.
Christian, D. L.; Chen, L. L.; Scannell, C. H.; Ferrando, R. E.; Welch, B. S.; Balmes, J. R. (1998) Ozone-induced
inflammation is attenuated with multiday exposure. Am. J. Respir. Crit. Care Med. 158: 532-537.
Coleridge, J. C. G.; Coleridge, H. M.; Schelegle, E. S.; Green, J. F. (1993) Acute inhalation of ozone stimulates
bronchial C-fibers and rapidly adapting receptors in dogs. J. Appl. Physiol. 74: 2345-2352.
Corradi, M.; Alinovi, R.; Goldoni, M.; Vettori, M.; Folesani, G.; Mozzoni, P.; Cavazzini, S.; Bergamaschi, E.; Rossi,
L.; Mutti, A. (2002) Biomarkers of oxidative stress after controlled human exposure to ozone. Toxicol. Lett.
134: 219-225.
David, G. L.; Romieu, I.; Sienra-Monge, J. J.; Collins, W. J.; Ramirez-Aguilar, M.; Del Rio-Navarro, B. E.;
Reyes-Ruiz, N. I.; Morris, R. W.; Marzec, J. M.; London, S. J. (2003) Nicotinamide adenine dinucleotide
(Phosphate) reduced:quinone oxidoreductase and glutathione s-transferase ml polymorphism and childhood
asthma. Am. J. Respir. Crit. Care Med. 168: 1199-1204.
Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the
lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
Devlin, R. B.; McDonnell, W. F.; Becker, S.; Madden, M. C.; McGee, M. P.; Perez, R.; Hatch, G.; House, D. E.;
Koren, H. S. (1996) Time-dependent changes of inflammatory mediators in the lungs of humans exposed to
0.4 ppm ozone for 2 hr: a comparison of mediators found in bronchoalveolar lavage fluid 1 and 18 hr after
exposure. Toxicol. Appl. Pharmacol. 138: 176-185.
Devlin, R. B.; Folinsbee, L. J.; Biscardi, F.; Hatch, G.; Becker, S.; Madden, M. C.; Robbins, M.; Koren, H. S. (1997)
Inflammation and cell damage induced by repeated exposure of humans to ozone. Inhalation Toxicol.
9:211-235.
Dimeo, M. J.; Glenn, M. G.; Holtzman, M. J.; Sheller, J. R.; Nadel, J. A.; Boushey, H. A. (1981) Threshold
concentration of ozone causing an increase in bronchial reactivity in humans and adaptation with repeated
exposures. Am. Rev. Respir. Dis. 124: 245-248.
Folinsbee, L. J.; Hazucha, M. J. (1989) Persistence of ozone-induced changes in lung function and airway
responsiveness. In: Schneider, T.; Lee, S. D.; Wolters, G. J. R.; Grant, L. D., eds. Atmospheric ozone research
and its policy implications: proceedings of the 3rd US-Dutch international symposium; May 1988; Nijmegen,
The Netherlands. Amsterdam, The Netherlands: Elsevier Science Publishers; pp. 483-492. (Studies in
environmental science 35).
Folinsbee, L. J.; Hazucha, M. J. (2000) Time course of response to ozone exposure in healthy adult females.
Inhalation Toxicol. 12: 151-167.
Folinsbee, L. J.; Horvath, S. M. (1986) Persistence of the acute effects of ozone exposure. Aviat. Space Environ.
Med. 57: 1136-1143.
Folinsbee, L. J.; Silverman, F.; Shephard, R. J. (1977) Decrease of maximum work performance following ozone
exposure. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 42: 531-536.
Folinsbee, L. J.; Drinkwater, B. L.; Bedi, J. F.; Horvath, S. M. (1978) The influence of exercise on the pulmonary
function changes due to exposure to low concentrations of ozone. In: Folinsbee, L. J.; Wagner, J. A.; Borgia,
J. F.; Drinkwater, B. L.; Gliner, J. A.; Bedi, J. F., eds. Environmental stress: individual human adaptations.
New York, NY: Academic Press; pp. 125-145.
6-48
-------
Folinsbee, L. I; Bedi, J. F.; Horvath, S. M. (1980) Respiratory responses in humans repeatedly exposed to low
concentrations of ozone. Am. Rev. Respir. Dis. 121: 431-439.
Folinsbee, L. J.; McDonnell, W. F.; Horstman, D. H. (1988) Pulmonary function and symptom responses after
6.6-hour exposure to 0.12 ppm ozone with moderate exercise. JAPCA 38: 28-35.
Folinsbee, L. J.; Horstman, D. H.; Kehrl, H. R.; Harder, S.; Abdul-Salaam, S.; Ives, P. J. (1994) Respiratory
responses to repeated prolonged exposure to 0.12 ppm ozone. Am. J. Respir. Crit. Care Med. 149: 98-105.
Folinsbee, L. J.; Devlin, R. B.; Robbins, M. K.; Biscardi, F. H.; Abdul-Salaam, S.; Koren, H. S. (1998) Repeated
exposure of humans to ozone: pulmonary function and symptom responses. Research Triangle Park, NC: U.S.
Environmental Protection Agency; National Center for Environmental Assessment; unpublished data.
Foster, W. M.; Stetkiewicz, P. T. (1996) Regional clearance of solute from the respiratory epithelia: 18-20 h
postexposure to ozone. J. Appl. Physiol. 81: 1143-1149.
Foster, W. M.; Costa, D. L.; Langenback, E. G. (1987) Ozone exposure alters tracheobronchial mucociliary function
in humans. J. Appl. Physiol. 63: 996-1002.
Foster, W. M.; Silver, J. A.; Groth, M. L. (1993) Exposure to ozone alters regional function and particle dosimetry in
the human lung. J. Appl. Physiol. 75: 1938-1945.
Foster, W. M.; Wills-Karp, M.; Tankersley, C. G.; Chen, X.; Paquette, N. C. (1996) Bloodborne markers in humans
during multiday exposure to ozone. J. Appl. Physiol. 81: 794-800.
Foster, W. M.; Weinmann, G. G.; Menkes, E.; Macri, K. (1997) Acute exposure of humans to ozone impairs small
airway function. Ann. Occup. Hyg. 41(suppl. 1): 659-666.
Foster, W. M.; Brown, R. H.; Macri, K.; Mitchell, C. S. (2000) Bronchial reactivity of healthy subjects: 18-20 h
postexposure to ozone. J. Appl. Physiol. 89: 1804-1810.
Fox, S. D.; Adams, W. C.; Brookes, K. A.; Lasley, B. L. (1993) Enhanced response to ozone exposure during the
follicular phase of the menstrual cycle. Environ. Health Perspect. 101: 242-244.
Foxcroft, W. J.; Adams, W. C. (1986) Effects of ozone exposure on four consecutive days on work performance and
VO2max. J. Appl. Physiol. 61: 960-966.
Frampton, M. W.; Morrow, P. E.; Torres, A.; Cox, C.; Voter, K. Z.; Utell, M. J.; Gibb, F. R.; Speers, D. M. (1997)
Ozone responsiveness in smokers and nonsmokers. Am. J. Respir. Crit. Care Med. 155: 116-121.
Frank, R.; Liu, M. C.; Spannhake, E. W.; Mlynarek, S.; Macri, K.; Weinmann, G. G. (2001) Repetitive ozone
exposure of young adults: evidence of persistent small airway dysfunction. Am. J. Respir. Crit. Care Med.
164: 1253-1260.
Freed, A. N.; Chou, C. L.; Fuller, S. D.; Croxton, T. L. (1996) Ozone-induced vagal reflex modulates airways
reactivity in rabbits. Respir. Physiol. 105: 95-102.
Gabrielson, E. W.; Yu, X.-Y.; Spannhake, E. W. (1994) Comparison of the toxic effects of hydrogen peroxide and
ozone on cultured human bronchial epithelial cells. Environ. Health Perspect. 102: 972-974.
Gerrity, T. R.; Bennett, W. D.; Kehrl, H.; DeWitt, P. J. (1993) Mucociliary clearance of inhaled particles measured
at 2 h after ozone exposure in humans. J. Appl. Physiol. 74: 2984-2989.
Gong, H., Jr.; Bradley, P. W.; Simmons, M. S.; Tashkin, D. P. (1986) Impaired exercise performance and pulmonary
function in elite cyclists during low-level ozone exposure in a hot environment. Am. Rev. Respir. Dis.
134: 726-733.
Gong, H., Jr.; Bedi, J. F.; Horvath, S. M. (1988) Inhaled albuterol does not protect against ozone toxicity in
nonasthmatic athletes. Arch. Environ. Health 43: 46-53.
Gong, H., Jr.; Shamoo, D. A.; Anderson, K. R.; Linn, W. S. (1997a) Responses of older men with and without
chronic obstructive pulmonary disease to prolonged ozone exposure. Arch. Environ. Health 52: 18-25.
Gong, H., Jr.; McManus, M. S.; Linn, W. S. (1997b) Attenuated response to repeated daily ozone exposures in
asthmatic subjects. Arch. Environ. Health 52: 34-41.
Gong, H., Jr.; Wong, R.; Sarma, R. J.; Linn, W. S.; Sullivan, E. D.; Shamoo, D. A.; Anderson, K. R.; Prasad, S. B.
(1998) Cardiovascular effects of ozone exposure in human volunteers. Am. J. Respir. Crit. Care Med.
158: 538-546.
Graham, D. E.; Koren, H. S. (1990) Biomarkers of inflammation in ozone-exposed humans: comparison of the nasal
and bronchoalveolar lavage. Am. Rev. Respir. Dis. 142: 152-156.
Graham, D.; Henderson, F.; House, D. (1988) Neutrophil influx measured in nasal lavages of humans exposed to
ozone. Arch. Environ. Health 43: 228-233.
Gwynn, R. C.; Thurston, G. D. (2001) The burden of air pollution: impacts among racial minorities. Environ. Health
Perspect. Suppl. 109(4): 501-506.
Hamilton, R. F.; Li, L.; Eschenbacher, W. L.; Szweda, L.; Holian, A. (1998) Potential involvement of
4-hydroxynonenal in the response of human lung cells to ozone. Am. J. Physiol. 274: L8-L16.
6-49
-------
Hasday, J. D.; Bascom, R.; Costa, J. I; Fitzgerald, T.; Dubin, W. (1999) Bacterial endotoxin is an active component
of cigarette smoke. Chest 115: 829-835.
Hazbun, M. E.; Hamilton, R.; Holian, A.; Eschenbacher, W. L. (1993) Ozone-induced increases in substance P and
8-epi-prostaglandin F2a in the airways of human subjects. Am. J. Respir. CellMol. Biol. 9: 568-572.
Hazucha, M. J. (1987) Relationship between ozone exposure and pulmonary function changes. J. Appl. Physiol.
62: 1671-1680.
Hazucha, M. J.; Sant'Ambrogio, G. (1993) Effects of ozone on the activity of slowly (SAR) and rapidly adapting
(RAR) receptors in cats. FASEB J. 7: 407A.
Hazucha, M. J.; Folinsbee, L. J.; Seal, E., Jr. (1992) Effects of steady-state and variable ozone concentration profiles
on pulmonary function. Am. Rev. Respir. Dis. 146: 1487-1493.
Hazucha, M. J.; Folinsbee, L. J.; Seal, E.; Bromberg, P. A. (1994) Lung function response of healthy women after
sequential exposures to NO2 and O3. Am. J. Respir. Crit. Care Med. 150: 642-647.
Hazucha, M. J.; Madden, M.; Pape, G.; Becker, S.; Devlin, R.; Koren, H. S.; Kehrl, H.; Bromberg, P. A. (1996)
Effects of cyclo-oxygenase inhibition on ozone-induced respiratory inflammation and lung function changes.
Eur. J. Appl. Physiol. Occup. Med. 73: 17-27.
Hazucha, M. J.; Folinsbee, L. J.; Bromberg, P. A. (2003) Distribution and reproducibility of spirometric response to
ozone by gender and age. J. Appl. Physiol. 95: 1917-1925.
Hiltermann, T. J. N.; Stolk, J.; Hiemstra, P. S.; Fokkens, P. H. B.; Rombout, P. J. A.; Sont, J. K.; Sterk, P. J.;
Dijkman, J. H. (1995) Effect of ozone exposure on maximal airway narrowing in non-asthmatic and asthmatic
subjects. Clin. Sci. 89: 619-624.
Hiltermann, T. J. N.; de Bruijne, C. R.; Stolk, J.; Zwinderman, A. H.; Spieksma, F. Th. M.; Roemer, W.;
Steerenberg, P. A.; Fischer, P. H.; vanBree, L.; Hiemstra, P. S. (1997) Effects of photochemical air pollution
and allergen exposure on upper respiratory tract inflammation in asthmatics. Am. J. Respir. Crit. Care Med.
156: 1765-1772.
Hiltermann, J. T. N.; Lapperre, T. S.; VanBree, L.; Steerenberg, P. A.; Brahim, J. J.; Sont, J. K.; Sterk, P. J.;
Hiemstra, P. S.; Stolk, J. (1999) Ozone-induced inflammation assessed in sputum and bronchial lavage fluid
from asthmatics: a new noninvasive tool in epidemiologic studies on air pollution and asthma. Free Radical
Biol. Med. 27: 1448-1454.
Holz, O.; Torres, R. A.; Timm, P.; Mticke, M.; Richter, K.; Koschyk, S.; Magnussen, H. (1999) Ozone-induced
airway inflammatory changes differ between individuals and are reproducible. Am. J. Respir. Crit. Care Med.
159: 776-784.
Holz, O.; Mticke, M.; Paasch, K.; Bohme, S.; Timm, P.; Richter, K.; Magnussen, H.; Torres, R. A. (2002) Repeated
ozone exposures enhance bronchial allergen responses in subjects with rhinitis or asthma. Clin. Exp. Allergy.
32: 681-689.
Holz, O.; Tal-Singer, R.; Kanniess, F.; Simpson, K. J.; Gibson, A.; Vessey, R. S. J.; Janicki, S.; Magnussen, H.;
Torres, R. A.; Richter, K. (2005) Validation of the human ozone challenge model as a tool for assessing
anti-inflammatory drugs in early development. J. Clin. Pharmacol. 45: 498-503.
Horstman, D. H.; Folinsbee, L. J.; Ives, P. J.; Abdul-Salaam, S.; McDonnell, W. F. (1990) Ozone concentration and
pulmonary response relationships for 6.6-hour exposures with five hours of moderate exercise to 0.08, 0.10,
and 0.12 ppm. Am. Rev. Respir. Dis. 142: 1158-1163.
Horstman, D. H.; Ball, B. A.; Brown, J.; Gerrity, T.; Folinsbee, L. J. (1995) Comparison of pulmonary responses of
asthmatic and nonasthmatic subjects performing light exercise while exposed to a low level of ozone.
Toxicol. Ind. Health 11: 369-385.
Horvath, S. M.; Gliner, J. A.; Matsen-Twisdale, J. A. (1979) Pulmonary function and maximum exercise responses
following acute ozone exposure. Aviat. Space Environ. Med. 50: 901-905.
Horvath, S. M.; Gliner, J. A.; Folinsbee, L. J. (1981) Adaptation to ozone: duration of effect. Am. Rev. Respir. Dis.
123:496-499.
Housley, D. G.; Eccles, R.; Richards, R. J. (1996) Gender difference in the concentration of the antioxidant uric acid
in human nasal lavage. Acta Oto-Laryngol. 116: 751-754.
Jenkins, H. S.; Devalia, J. L.; Mister, R. L.; Bevan, A. M.; Rusznak, C.; Davies, R. J. (1999) The effect of exposure
to ozone and nitrogen dioxide on the airway response of atopic asthmatics to inhaled allergen: dose- and
time-dependent effects. Am. J. Respir. Crit. Care Med. 160: 33-39.
Joad, J. P.; Kott, K. S.; Brie, J. M. (1996) The local C-fiber contribution to ozone-induced effects on the isolated
guinea pig lung. Toxicol. Appl. Pharmacol. 141: 561-567.
Torres, R.; Nowak, D.; Magnussen, H.; Speckin, P.; Koschyk, S. (1996) The effect of ozone exposure on allergen
responsiveness in subjects with asthma or rhinitis. Am. J. Respir. Crit. Care Med. 153: 56-64.
6-50
-------
Torres, R. A.; Holz, O.; Zachgo, W.; Timm, P.; Koschyk, S.; Mtiller, B.; Grimminger, F.; Seeger, W.; Kelly, F. J.;
Dunster, C.; Frischer, T.; Lubec, G.; Waschewski, M.; Niendorf, A.; Magnussen, H. (2000) The effect of
repeated ozone exposures on inflammatory markers inbronchoalveolar lavage fluid and mucosal biopsies.
Am. J. Respir. Crit. Care Med. 161: 1855-1861.
Kehrl, H. R.; Hazucha, M. J.; Solic, J. J.; Bromberg, P. A. (1985) Responses of subjects with chronic obstructive
pulmonary disease after exposures to 0.3 ppm ozone. Am. Rev. Respir. Dis. 131: 719-724.
Kehrl, H. R.; Vincent, L. M.; Kowalsky, R. J.; Horstman, D. H.; O'Neil, J. J.; McCartney, W. H.; Bromberg, P. A.
(1987) Ozone exposure increases respiratory epithelial permeability in humans. Am. Rev. Respir. Dis. 135:
1124-1128.
Kehrl, H. R.; Peden, D. B.; Ball, B. A.; Folinsbee, L. J.; Horstman, D. H. (1999) Increased specific airway reactivity
of persons with mild allergic asthma after 7.6 hours of exposure to 0.16 ppm ozone. J. Allergy. Clin.
Immunol. 104: 1198-1204.
Kleinman, M. T.; Mautz, W. J.; Bjarnason, S. (1999) Adaptive and non-adaptive responses in rats exposed to ozone,
alone and in mixtures, with acidic aerosols. Inhalation Toxicol. 11: 249-264.
Koren, H. S.; Devlin, R. B.; Graham, D. E.; Mann, R.; McGee, M. P.; Horstman, D. H.; Kozumbo, W. J.; Becker, S.;
House, D. E.; McDonnell, W. F.; Bromberg, P. A. (1989) Ozone-induced inflammation in the lower airways
of human subjects. Am. Rev. Respir. Dis. 139: 407-415.
Koren, H. S.; Devlin, R. B.; Becker, S.; Perez, R.; McDonnell, W. F. (1991) Time-dependent changes of markers
associated with inflammation in the lungs of humans exposed to ambient levels of ozone. Toxicol. Pathol.
19:406-411.
Kozumbo, W. J.; Hanley, N. M.; Agarwal, S.; Thomas, M. J.; Madden, M. C. (1996) Products of ozonized
arachidonic acid potentiate the formation of DNA single strand breaks in cultured human lung cells. Environ.
Mol. Mutagen. 27: 185-195.
Kreit, J. W.; Gross, K. B.; Moore, T. B.; Lorenzen, T. J.; D'Arcy, J.; Eschenbacher, W. L. (1989) Ozone-induced
changes in pulmonary function and bronchial responsiveness in asthmatics. J. Appl. Physiol. 66: 217-222.
Krishna, M. T.; Springall, D.; Meng, Q.-H.; Withers, N.; Macleod, D.; Biscione, G.; Frew, A.; Polak, J.; Holgate, S.
(1997) Effects of ozone on epithelium and sensory nerves in the bronchial mucosa of healthy humans. Am. J.
Respir. Crit. Care Med. 156: 943-950.
Kulle, T. J.; Sauder, L. R.; Kerr, H. D.; Farrell, B. P.; Bermel, M. S.; Smith, D. M. (1982) Duration of pulmonary
function adaptation to ozone in humans. Am. Ind. Hyg. Assoc. J. 43: 832-837.
Kulle, T. J.; Milman, J. H.; Sauder, L. R.; Kerr, H. D.; Farrell, B. P.; Miller, W. R. (1984) Pulmonary function
adaptation to ozone in subjects with chronic bronchitis. Environ. Res. 34: 55-63.
Kulle, T. J.; Sauder, L. R.; Hebel, J. R.; Chatham, M. D. (1985) Ozone response relationships in healthy
nonsmokers. Am. Rev. Respir. Dis. 132: 36-41.
Larsen, R. L; McDonnell, W. F.; Horstman, D. H.; Folinsbee, L. J. (1991) An air quality data analysis system for
interrelating effects, standards, and needed source reductions: part 11. a lognormal model relating human lung
function decrease to O3 exposure. J. Air Waste Manage. Assoc. 41: 455-459.
Linder, J.; Herren, D.; Monn, C.; Wanner, H.-U. (1988) Die Wirkung von Ozon auf die korperliche
Leistungsfahigkeit [The effect of ozone on physical activity]. Schweiz Z. Sportmed. 36: 5-10.
Linn, W. S.; Jones, M. P.; Bachmayer, E. A.; Spier, C. E.; Mazur, S. F.; Avol, E. L.; Hackney, J. D. (1980)
Short-term respiratory effects of polluted ambient air: a laboratory study of volunteers in a high-oxidant
community. Am. Rev. Respir. Dis. 121: 243-252.
Linn, W. S.; Fischer, D. A.; Medway, D. A.; Anzar, U. T.; Spier, C. E.; Valencia, L. M.; Venet, T. G.; Hackney, J.
D. (1982a) Short-term respiratory effects of 0.12 ppm ozone exposure in volunteers with chronic obstructive
pulmonary disease. Am. Rev. Respir. Dis. 125: 658-663.
Linn, W. S.; Medway, D. A.; Anzar, U. T.; Valencia, L. M.; Spier, C. E.; Tsao, F. S.-D.; Fischer, D. A.; Hackney, J.
D. (1982b) Persistence of adaptation to ozone in volunteers exposed repeatedly for six weeks. Am. Rev.
Respir. Dis. 125:491-495.
Linn, W. S.; Shamoo, D. A.; Venet, T. G.; Spier, C. E.; Valencia, L. M.; Anzar, U. T.; Hackney, J. D. (1983a)
Response to ozone in volunteers with chronic obstructive pulmonary disease. Arch. Environ. Health
38: 278-283.
Linn, W. S.; Avol, E. L.; Hackney, J. D. (1983b) Effects of ambient oxidant pollutants on humans: a movable
environmental chamber study. In: Lee, S. D.; Mustafa, M. G.; Mehlman, M. A., eds. International symposium
on the biomedical effects of ozone and related photochemical oxidants; March 1982; Pinehurst, NC.
Princeton, NJ: Princeton Scientific Publishers, Inc.; pp. 125-137.
6-51
-------
Linn, W. S.; Avol, E. L.; Shamoo, D. A.; Spier, C. E.; Valencia, L. M.; Venet, T. G.; Fischer, D. A.; Hackney, J. D.
(1986) A dose-response study of healthy, heavily exercising men exposed to ozone at concentrations near the
ambient air quality standard. Toxicol. Ind. Health 2: 99-112.
Linn, W. S.; Gong, H., Jr.; Shamoo, D. A.; Anderson, K. R.; Avol, E. L. (1997) Chamber exposures of children to
mixed ozone, sulfur dioxide, and sulfuric acid. Arch. Environ. Health 52: 179-187.
Liu, L.; Leech, J. A.; Urch, R. B.; Silverman, F. S. (1997) In vivo salicylate hyroxylation: a potential biomarker for
assessing acute ozone exposure and effects in humans. Am. J. Respir. Crit. Care Med. 156: 1405-1412.
Liu, L.; Leech, J. A.; Urch, R. B.; Poon, R.; Zimmerman, B.; Kubay, J. M.; Silverman, F. S. (1999) A comparison of
biomarkers of ozone exposure in human plasma, nasal lavage, and sputum. Inhalation Toxicol. 11: 657-674.
Marthan, R.; Roux, E.; Savineau, J.-P. (1996) Human bronchial smooth muscle responsiveness after in vitro
exposure to oxidizing pollutants. Cell Biol. Toxicol. 12: 245-249.
McBride, D. E.; Koenig, J. Q.; Luchtel, D. L.; Williams, P. V.; Henderson, W. R., Jr. (1994) Inflammatory effects of
ozone in the upper airways of subjects with asthma. Am. J. Respir. Crit. Care Med. 149: 1192-1197.
McDonnell, W. F. (1996) Individual variability in human lung function responses to ozone exposure. Environ.
Toxicol. Pharmacol. 2: 171-175.
McDonnell, W. F.; Horstman, D. H.; Hazucha, M. J.; Seal, E., Jr.; Haak, E. D.; Salaam, S. A.; House, D. E. (1983)
Pulmonary effects of ozone exposure during exercise: dose-response characteristics. J. Appl. Physiol.: Respir.
Environ. Exercise Physiol. 54: 1345-1352.
McDonnell, W. F., Ill; Horstman, D. H.; Abdul-Salaam, S.; House, D. E. (1985) Reproducibility of individual
responses to ozone exposure. Am. Rev. Respir. Dis. 131: 36-40.
McDonnell, W. F.; Kehrl, H. R.; Abdul-Salaam, S.; Ives, P. J.; Folinsbee, L. J.; Devlin, R. B.; O'Neil, J. J.;
Horstman, D. H. (1991) Respiratory response of humans exposed to low levels of ozone for 6.6 hours. Arch.
Environ. Health 46: 145-150.
McDonnell, W. F.; Muller, K. E.; Bromberg, P. A.; Shy, C. M. (1993) Predictors of individual differences in acute
response to ozone exposure. Am. Rev. Respir. Dis. 147: 818-825.
McDonnell, W. F.; Stewart, P. W.; Andreoni, S.; Seal, E., Jr.; Kehrl, H. R.; Horstman, D. H.; Folinsbee, L. J.; Smith,
M. V. (1997) Prediction of ozone-induced FEVj changes: effects of concentration, duration, and ventilation.
Am. J. Respir. Crit. Care Med. 156: 715-722.
McDonnell, W. F.; Stewart, P. W.; Smith, M. V.; Pan, W. K.; Pan, J. (1999) Ozone-induced respiratory symptoms:
exposure-response models and association with lung function. Eur. Respir. J. 14: 845-853.
McKenzie, D. C.; Stirling, D. R.; Fadl, S.; Allen, M. (1987) The effects of salbutamol on pulmonary function in
cyclists exposed to ozone: a pilot study. Can. J. Sport Sci. 12: 46-48.
Messineo, T. D.; Adams, W. C. (1990) Ozone inhalation effects in females varying widely in lung size: comparison
with males. J. Appl. Physiol. 69: 96-103.
Michelson, P. H.; Dailey, L.; Devlin, R. B.; Peden, D. B. (1999) Ozone effects on the immediate-phase response to
allergen in the nasal airways of allergic asthmatic subjects. Otolaryngol. Head Neck Surg. 120: 225-232.
Mohammed, S. P.; Higenbottam, T. W.; Adcock, J. J. (1993) Effects of aerosol-applied capsaicin, histamine and
prostaglandin-E2 on airway sensory receptors of anaesthetized cats. J. Physiol. Lond. 469: 51-66.
Molfino, N. A.; Wright, S. C.; Katz, I.; Tarlo, S.; Silverman, F.; McClean, P. A.; Szalai, J. P.; Raizenne, M.; Slutsky,
A. S.; Zamel, N. (1991) Effect of low concentrations of ozone on inhaled allergen responses in asthmatic
subjects. Lancet 338(8761): 199-203.
Mudway, I. S.; Kelly, F. J. (2000) Ozone and the lung: a sensitive issue. Mol. Aspects. Med. 21: 1-48.
Mudway, I. S.; Kelly, F. J. (2004) An investigation of inhaled ozone dose and the magnitude of airway inflammation
in healthy adults. Am. J. Respir. Crit. Care Med. 169: 1089-1095.
Mudway, I. S.; Blomberg, A.; Frew, A. J.; Holgate, S. T.; Sandstrom, T.; Kelly, F. J. (1999) Antioxidant
consumption and repletion kinetics in nasal lavage fluid following exposure of healthy human volunteers to
ozone. Eur. Respir. J. 13: 1429-1438.
Mudway, I. S.; Stenfors, N.; Blomberg, A.; Helleday, R.; Dunster, C.; Marklund, S. L.; Frew, A. J.; Sandstrom, T.;
Kelly, F. J. (2001) Differences in basal airway antioxidant concentrations are not predictive of individual
responsiveness to ozone: a comparison of healthy and mild asthmatic subjects. Free Radical Biol. Med.
31:962-974.
Nayak, A. S. (2003) The asthma and allergic rhinitis link. Allergy Asthma Proc. 24: 395-402.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol.
141: 111-122.
6-52
-------
Newson, E. I; Krishna, M. T.; Lau, L. C. K.; Howarth, P. H.; Holgate, S. T.; Frew, A. J. (2000) Effects of
short-term exposure to 0.2 ppm ozone onbiomarkers of inflammation in sputum, exhaled nitric oxide, and
lung function in subjects with mild atopic asthma. J. Occup. Environ. Med. 42: 270-277.
Nightingale, J. A.; Rogers, D. F.; Chung, K. F.; Barnes, P. J. (2000) No effect of inhaled budesonide on the response
to inhaled ozone in normal subjects. Am. J. Respir. Crit. Care Med. 161: 479-486.
Nikasinovic, L.; Momas, L; Seta, N. (2003) Nasal epithelial and inflammatory response to ozone exposure: a review
of laboratory-based studies published since 1985. J. Toxicol. Environ. Health B 6: 521-568.
Otto-Knapp, R.; Jurgovsky, K.; Schierhorn, K.; Kunkel, G. (2003) Antioxidative enzymes in human nasal mucosa
after exposure to ozone. Possible role of GSTM1 deficiency. Inflamm. Res. 52: 51-55.
Passannante, A. N.; Hazucha, M. J.; Bromberg, P. A.; Seal, E.; Folinsbee, L.; Koch, G. (1998) Nociceptive
mechanisms modulate ozone-induced human lung function decrements. J. Appl. Physiol. 85: 1863-1870.
Peden, D. B. (2001a) Air pollution in asthma: effect of pollutants on airway inflammation. Ann. Allergy Asthma
Immunol. 87(suppl. 3): 12-17.
Peden, D. B. (200 Ib) Effect of pollutants in rhinitis. Curr. Allergy Asthma Rep. 1: 242-246.
Peden, D. B.; Setzer, R. W., Jr.; Devlin, R. B. (1995) Ozone exposure has both a priming effect on allergen-induced
responses and an intrinsic inflammatory action in the nasal airways of perennially allergic asthmatics. Am. J.
Respir. Crit. Care Med. 151: 1336-1345.
Peden, D. B.; Boehlecke, B.; Horstman, D.; Devlin, R. (1997) Prolonged acute exposure to 0.16 ppm ozone induces
eosinophilic airway inflammation in asthmatic subjects with allergies. J. Allergy Clin. Immunol.
100: 802-808.
Riediker, M.; Monn, C.; Koller, T.; Stahel, W. A.; Wuthrich, B. (2001) Air pollutants enhance rhinoconjunctivitis
symptoms in pollen-allergic individuals. Ann. Allergy Asthma Immunol. 87: 311-318.
Rigas, M. L.; Ben-Jebria, A.; Ultman, J. S. (1997) Longitudinal distribution of ozone absorption in the lung: effects
of nitrogen dioxide, sulfur dioxide, and ozone exposures. Arch. Environ. Health 52: 173-178.
Romieu, L; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. L; Estela del
Rio-Navarro, B.; Hernandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and
antioxidant supplementation influence lung function in relation to ozone exposure in asthmatic children in
Mexico City. Thorax 59: 8-10.
Samet, J. M.; Hatch, G. E.; Horstman, D.; Steck-Scott, S.; Arab, L.; Bromberg, P. A.; Levine, M.; McDonnell, W.
F.; Devlin, R. B. (2001) Effect of antioxidant supplementation on ozone-induced lung injury in human
subjects. Am. J. Respir. Crit. Care Med. 164: 819-825.
Scannell, C.; Chen, L.; Aris, R. M.; Tager, I.; Christian, D.; Ferrando, R.; Welch, B.; Kelly, T.; Balmes, J. R. (1996)
Greater ozone-induced inflammatory responses in subjects with asthma. Am. J. Respir. Crit. Care Med.
154: 24-29.
Schelegle, E. S.; Adams, W. C. (1986) Reduced exercise time in competitive simulations consequent to low level
ozone exposure. Med. Sci. Sports Exercise 18: 408-414.
Schelegle, E. S.; Adams, W. C.; Siefkin, A. D. (1987) Indomethacin pretreatment reduces ozone-induced pulmonary
function decrements in human subjects. Am. Rev. Respir. Dis. 136: 1350-1354.
Schelegle, E. S.; Siefkin, A. D.; McDonald, R. J. (1991) Time course of ozone-induced neutrophilia in normal
humans. Am. Rev. Respir. Dis. 143: 1353-1358.
Schelegle, E. S.; Carl, M. L.; Coleridge, H. M.; Coleridge, J. C. G.; Green, J. F. (1993) Contribution of vagal
afferents to respiratory reflexes evoked by acute inhalation of ozone in dogs. J. Appl. Physiol. 74: 2338-2344.
Schelegle, E. S.; Eldridge, M. W.; Cross, C. E.; Walby, W. F.; Adams, W. C. (2001) Differential effects of airway
anesthesia on ozone-induced pulmonary responses in human subjects. Am. J. Respir. Crit. Care Med. 163:
1121-1127.
Schonfeld, B. R.; Adams, W. C.; Schelegle, E. S. (1989) Duration of enhanced responsiveness upon re-exposure to
ozone. Arch. Environ. Health 44: 229-236.
Schwartz, L. W.; Dungworth, D. L.; Mustafa, M. G.; Tarkington, B. K.; Tyler, W. S. (1976) Pulmonary responses of
rats to ambient levels of ozone: effects of 7-day intermittent or continuous exposure. Lab. Invest.
Seal, E., Jr.; McDonnell, W. F.; House, D. E.; Salaam, S. A.; Dewitt, P. J.; Butler, S. O.; Green, J.; Raggio, L.
(1993) The pulmonary response of white and black adults to six concentrations of ozone. Am. Rev. Respir.
Dis. 147: 804-810.
Seal, E., Jr.; McDonnell, W. F.; House, D. E. (1996) Effects of age, socioeconomic status, and menstrual cycle on
pulmonary response to ozone. Arch. Environ. Health 51: 132-137.
Solic, J. J.; Hazucha, M. J.; Bromberg, P. A. (1982) The acute effects of 0.2 ppm ozone in patients with chronic
obstructive pulmonary disease. Am. Rev. Respir. Dis. 125: 664-669.
6-53
-------
Solway, I; Leff, A. R. (1991) Sensory neuropeptides and airway function. J. Appl. Physiol. 71: 2077-2087.
Spannhake, E. W.; Reddy, S. P. M; Jacoby, D. B.; Yu, X.-Y.; Saatian, B.; Tian, J. (2002) Synergismbetween
rhinovirus infection and oxidant pollutant exposure enhances airway epithelial cell cytokine production.
Environ. Health Perspect. 110: 665-670.
Stenfors, N.; Pourazar, J.; Blomberg, A.; Krishna, M. T.; Mudway, I.; Helleday, R.; Kelly, F. J.; Frew, A. J.;
Sandstrom, T. (2002) Effect of ozone on bronchial mucosal inflammation in asthmatic and healthy subjects.
Respir. Med. 96: 352-358.
Tepper, J. S.; Costa, D. L.; Lehmann, J. R.; Weber, M. F.; Hatch, G. E. (1989) Unattenuated structural and
biochemical alterations in the rat lung during functional adaptation to ozone. Am. Rev. Respir. Dis.
140: 493-501.
Tepper, J. S.; Wiester, M. J.; Weber, M. F.; Menache, M. G. (1990) Measurements of cardiopulmonary response in
awake rats during acute exposure to near-ambient concentrations of ozone. J. Appl. Toxicol. 10: 7-15.
Torres, A.; Utell, M. J.; Morow, P. E.; Voter, K. Z.; Whitin, J. C.; Cox, C.; Looney, R. J.; Speers, D. M.; Tsai, Y.;
Frampton, M. W. (1997) Airway inflammation in smokers and nonsmokers with varying responsiveness to
ozone. Am. J. Respir. Crit. Care Med. 156: 728-736.
Trenga, C. A.; Koenig, J. Q.; Williams, P. V. (2001) Dietary antioxidants and ozone-induced bronchial
hyperresponsiveness in adults with asthma. Arch. Environ. Health 56: 242-249.
U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA;
PB87-142949.
U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
http://cfpub2.epa.gov/ncea/.
Ultman, J. S.; Ben-Jebria, A.; Arnold, S. F. (2004) Uptake distribution of ozone in human lungs: intersubject
variability in physiologic response. Boston, MA: Health Effects Institute; research report no. 125. Available:
http://www.healtheffects.org/Pubs/Ultman.pdft29 July, 2005].
Vagaggini, B.; Carnevali, S.; Macchioni, P.; Taccola, M.; Fornai, E.; Bacci, E.; Bartoli, M. L.; Cianchetti, S.; Dente,
F. L.; Di Franco, A.; Giannini, D.; Paggiaro. P. L. (1999) Airway inflammatory response to ozone in subjects
with different asthma severity. Eur. Respir. J. 13: 274-280.
Vagaggini, B.; Taccola, M.; Conti, L; Carnevali, S.; Cianchetti, S.; Bartoli, M. L.; Bacci, E.; Dente, F. L.; Di Franco,
A.; Giannini, D.; Paggiaro, P. L. (2001) Budesonide reduces neutrophilic but not functional airway response
to ozone in mild asthmatics. Am. J. Respir. Crit. Care Med. 164: 2172-2176.
Vagaggini, B.; Taccola, M.; Cianchetti, S.; Carnevali, S.; Bartoli, M. L.; Bacci, E.; Dente, F. L.; Di Franco, A.;
Giannini, D.; Paggiaro, P. L. (2002) Ozone exposure increases eosinophilic airway response induced by
previous allergen challenge. Am. J. Respir. Crit. Care Med. 166: 1073-1077.
Voter, K. Z.; Whitin, J. C.; Torres, A.; Morrow, P. E.; Cox, C.; Tsai, Y.; Utell, M. J.; Frampton, M. W. (2001)
Ozone exposure and the production of reactive oxygen species by bronchoalveolar cells in humans. Inhalation
Toxicol. 13:465-483.
Wang, G.; Umstead, T. M.; Phelps, D. S.; Al-Mondhiry, H.; Floras, J. (2002) The effect of ozone exposure on the
ability of human surfactant protein A variants to stimulate cytokine production. Environ. Health Perspect.
110:79-84.
Weinmann, G. G.; Weidenbach-Gerbase, M.; Foster, W. M.; Zacur, H.; Frank, R. (1995) Evidence for
ozone-induced small-airway dysfunction: lack of menstrual-cycle and gender effects. Am. J. Respir. Crit.
Care Med. 152: 988-996.
Yang, I. A.; Holz, O.; Torres, R. A.; Magnussen, H.; Barton, S. J.; Rodriguez, S.; Cakebread, J. A.; Holloway, J. W.;
Holgate, S. T. (2005) Association of tumor necrosis factor-a polymorphisms and ozone-induced change in
lung function. Am. J. Respir. Crit. Care Med. 171: 171-176.
Yeadon, M.; Wilkinson, D.; Darley-Usmar, V.; O'Leary, V. J.; Payne, A. N. (1992) Mechanisms contributing to
ozone-induced bronchial hyperreactivity in guinea-pigs. Pulm. Pharmacol. 5: 39-50.
Ying, R. L.; Gross, K. B.; Terzo, T. S.; Eschenbacher, W. L. (1990) Indomethacin does not inhibit the
ozone-induced increase in bronchial responsiveness in human subjects. Am. Rev. Respir. Dis. 142: 817-821.
Yu, M.; Pinkerton, K. E.; Witschi, H. (2002) Short-term exposure to aged and diluted sidestream cigarette smoke
enhances ozone-induced lung injury inB6C3Fl mice. Toxicol. Sci. 65: 99-106.
Zhang, L.-Y.; Levitt, R. C.; Kleeberger, S. R. (1995) Differential susceptibility to ozone-induced airways
hyperreactivity in inbred strains of mice. Exp. Lung Res. 21: 503-518.
6-54
-------
7. EPIDEMIOLOGIC STUDIES OF
HUMAN HEALTH EFFECTS ASSOCIATED
WITH AMBIENT OZONE EXPOSURE
7.1 INTRODUCTION
This chapter evaluates current epidemiologic literature on health and physiological effects
of ambient ozone (O3) exposure. Epidemiologic studies linking community ambient O3
concentrations to health effects were reported in the 1996 Ozone Air Quality Criteria Document
(O3 AQCD; U.S. Environmental Protection Agency, 1996a). Many of those studies reported that
pulmonary function decrements, respiratory symptoms, and hospital and emergency department
admissions in human populations were associated with ambient levels of O3. Numerous more
recent epidemiologic studies discussed in this chapter evaluate the relationship of ambient O3 to
morbidity and mortality, and thereby provide an expanded basis for assessment of health effects
associated with exposures to O3 at concentrations currently encountered in the United States.
As discussed elsewhere in this document (Chapters 5 and 6), a substantial amount of
experimental evidence links O3 exposure unequivocally with respiratory effects in laboratory
animals and humans. These include structural changes in the bronchiolar-alveolar transition
(centriacinar) region of the lung, biochemical evidence of acute cellular and tissue injury,
inflammation, increased frequency and severity of experimental bacterial infection, and
temporary reductions in mechanical lung function. These effects have been observed with
exposure to O3 at ambient or near-ambient concentrations. Thus, many of the reported
epidemiologic associations of ambient O3 with respiratory health effects have considerable
biological credibility. Accordingly, the new epidemiologic studies of ambient O3 assessed here
are best considered in combination with information from the other chapters on ambient O3
concentration and exposure (Chapter 3), and toxicological effects of O3 in animals and humans
(Chapters 5 and 6, respectively). The epidemiologic studies constitute important information on
associations between health effects and exposures of human populations to "real-world" O3 and
also help to identify susceptible subgroups and associated risk factors.
7-1
-------
A wide variety of oxidants in both the gaseous and particulate phases have not been
examined in relation to health outcomes in the literature. Therefore, discussion in this chapter is
limited to studies of human health effects associated with ambient O3 exposure. Ambient
concentrations of the most abundant non-O3 oxidants (i.e., peroxyacetyl nitrate, peroxypropionyl
nitrate, and H2O2) have not been shown to cause adverse health effects in toxicologic studies.
However, as constituents of ambient air mixes, these oxidants may contribute to some of the
effects attributed to O3. Therefore, health effect associations observed in relation to ambient O3
concentrations may represent O3 effects, per se, or O3 may be serving as a surrogate measure for
the overall photochemical oxidant mix.
7.1.1 Approach to Identifying Ozone Epidemiologic Studies
Numerous O3 epidemiologic papers have been published since completion of the 1996 O3
AQCD. The U.S. Environmental Protection Agency (NCEA-RTP) has implemented a
systematic approach to identify relevant epidemiologic studies for consideration in this chapter.
In general, an ongoing search has been employed in conjunction with other strategies to
identify O3 epidemiologic literature pertinent to developing criteria for the O3 National Ambient
Air Quality Standards (NAAQS). A publication base was established using Medline, Pascal,
BIOSIS, NTIS, and Embase, as well as a set of search terms proven by prior use to identify
pertinent literature. The search strategy was later reexamined and modified to enhance
identification of published papers. PubMed was added to the search regime. New studies
accepted for publication through December 2004, as identified using the approaches above, have
been included in this AQCD.
While the above search regime provided good coverage of the relevant literature,
additional approaches augmented the traditional search methods. First, a Federal Register
Notice was issued requesting information and published papers from the public at large. Next,
non-EPA chapter authors, expert in this field, identified literature on their own. NCEA-RTP
staff also identified publications as an element of their assessment and interpretation of the
literature. Finally, additional potentially relevant publications were included following external
review as a result of comments from both the public and the Clean Air Scientific Advisory
Committee (CASAC). More recent studies accepted in 2005 and 2006 for publication also were
included if they added significantly to the existing body of data on critically important topics,
7-2
-------
such as those discussing methodological issues (e.g., O3 exposure misclassification, potential
confounding by copollutants and meteorological factors), the O3-mortality relationship, and
cardiovascular health outcomes. The combination of these approaches is believed to have
produced a comprehensive ascertainment of studies appropriate for review and assessment here.
The main criterion for selecting literature for the present assessment is to include those identified
studies that evaluated relationships between measured ambient O3 levels and human health
outcomes, which are pertinent to the evaluation of scientific bases useful for derivation of O3
NAAQS for the United States.
7.1.2 Approach to Assessing Epidemiologic Evidence
Definitions of the various types of epidemiologic studies assessed here were noted in an
earlier paniculate matter (PM) AQCD (U.S. Environmental Protection Agency, 1996b). Briefly,
epidemiologic studies are generally divided into two groups, morbidity studies and mortality
studies. Morbidity studies evaluate O3 effects on a wide range of health endpoints, including:
changes in pulmonary function, respiratory symptoms, self-medication in asthmatics, and airway
inflammation; changes in cardiovascular physiology/functions; and cardiopulmonary emergency
department visits and hospital admissions. Mortality studies investigate O3 effects on total
(nonaccidental) mortality and cause-specific mortality, providing evidence related to a clearly
adverse endpoint. The epidemiologic strategies most commonly used in O3 health studies are
cross-sectional studies, prospective cohort studies, ecologic studies, time-series semi-ecologic
studies, and case-crossover studies. All of these are observational rather than experimental
studies.
The approach to assessing epidemiologic evidence has been stated most recently in the
2004 PM AQCD (U.S. Environmental Protection Agency, 2004) and is summarized here. The
critical assessment of epidemiologic evidence presented in this chapter is conceptually based
upon consideration of salient aspects of the evidence of associations so as to reach fundamental
judgments as to the likely causal significance of the observed associations (see Hill, 1965). The
general evaluation of the strength of the epidemiologic evidence reflects consideration not only
of the magnitude and precision of reported O3 effect estimates and their statistical significance,
but also of the robustness of the effects associations. Statistical significance corresponds to the
allowable rate of error (Type I error) in the decision problem constructed from assuming that a
-------
simple null hypothesis of no association is true. It is a conditional probability; for statistical
significance, typically there is a less than 0.05 chance of rejecting the null hypothesis given that
it is true. Robustness of the associations is defined as stability in the effect estimates after
considering a number of factors, including alternative models and model specifications, potential
confounding by copollutants, as well as issues related to the consequences of measurement error.
Consideration of the consistency of the effects associations, as discussed in the following
sections, involves looking across the results of multiple- and single-city studies conducted by
different investigators in different places and times. Relevant factors are known to exhibit much
variation across studies, including, for example, the presence and levels of copollutants, the
relationships between central measures of O3 and exposure-related factors, relevant demographic
factors related to sensitive subpopulations, and climatic and meteorological conditions. Thus, in
this case, consideration of consistency and the related heterogeneity of effects are appropriately
understood as an evaluation of the similarity or general concordance of results, rather than an
expectation of finding quantitative results within a very narrow range.
Looking beyond the epidemiologic evidence, evaluation of the biological plausibility of
the O3-health effects associations observed in epidemiologic studies reflects consideration of
both exposure-related factors and dosimetric/toxicologic evidence relevant to identification of
potential biological mechanisms. Similarly, coherence of health effects associations reported in
the epidemiologic literature reflects consideration of information pertaining to the nature of the
various respiratory- and cardiac-related mortality and morbidity effects and biological markers
evaluated in toxicologic and human clinical studies. These broader aspects of the assessment are
only touched upon here but are more fully integrated in the discussion presented in Chapter 8.
In assessing the relative scientific quality of epidemiologic studies reviewed here and to
assist in interpreting their findings, the following considerations were taken into account:
(1) To what extent are the aerometric data/exposure metrics used of adequate quality and
sufficiently representative to serve as credible exposure indicators, well-reflecting
geographic or temporal differences in study population pollutant exposures in the
range(s) of ambient pollutant concentrations evaluated?
(2) Were the study populations well defined and adequately selected so as to allow for
meaningful comparisons between study groups or for meaningful temporal analyses
of health effects results?
7-4
-------
(3) Were the health endpoint measurements meaningful and reliable, including clear
definition of diagnostic criteria utilized and consistency in obtaining dependent
variable measurements?
(4) Were the statistical analyses used appropriate, as well as properly performed and
interpreted?
(5) Were likely important covariates (e.g., potential confounders or effect modifiers)
adequately controlled for or taken into account in the study design and statistical
analyses?
(6) Were the reported findings internally consistent, biologically plausible, and coherent
in terms of consistency with other known facts?
These guidelines provide benchmarks for judging the relative quality of various studies and
in assessing the body of epidemiologic evidence. Detailed critical analysis of all epidemiologic
studies on O3 health effects, especially in relation to all of the above questions, is beyond the
scope of this document. As discussed in the upcoming Sections 7.1.3 and 7.1.4, considerations
in the interpretation and presentation of the epidemiologic evidence led to emphasis being placed
on certain studies in the main chapter text, tables, and figures. Additional studies in the O3
epidemiologic literature are presented in Chapter 7 Annex Tables (see Annex Section AX7.1).
Of most importance are those studies which provide useful qualitative or quantitative
information on concentration-response relationships for health effects associated with ambient
air levels of O3 likely to be encountered among healthy and susceptible populations in the
United States.
7.1.3 Considerations in the Interpretation of Epidemiologic Studies of
Ozone Health Effects
Prior to discussing results from recent O3 epidemiologic studies, issues and questions
arising from the study designs and analysis methods used in the assessment of O3 effect
estimates will be briefly presented. Study design can restrict the health effect parameters that
can be estimated. Separate considerations need to be made for acute versus chronic effect
studies, as well as individual- versus aggregate-level analyses. Time-series studies and panel
studies are most frequently conducted in air pollution epidemiologic research. Aggregate-level
exposure and/or outcome data are often used in these types of studies. Analyses using
administrative health outcome data (e.g., numbers of deaths and emergency hospital admissions)
7-5
-------
have inherent limitations as well as strengths (Virnig and McBean, 2001). The impact of study
design or the loss of information due to aggregation depends on the source of exposure
(Sheppard et al., 2005).
This section mainly focuses on the topics of exposure assessment and model specification
in air pollution epidemiologic studies. Potential biases that may result from O3 exposure
measurement error and from the choice of exposure index and lag period are discussed first.
Model specification issues and potential confounding by temporal factors, meteorological
effects, seasonal trends, and copollutants are then discussed.
7.1.3.1 Exposure Assessment and Measurement Error in Epidemiologic Studies
In many air pollution epidemiologic studies, especially time-series studies with
administrative data on mortality and hospitalization outcomes, data from central ambient
monitoring sites generally are used as the estimate of exposure. Personal exposures of individual
study participants generally are not directly measured in epidemiologic studies. The use of O3
concentrations from ambient monitors as surrogate measures for personal O3 exposures was
discussed previously (Chapter 3, Section 3.9). Routinely collected ambient monitor data, though
readily available and convenient, may not represent true personal exposure, which includes both
ambient and non-ambient (i.e., indoor) source exposures.
In several studies focused on evaluating exposure to O3, measurements were made in a
variety of indoor environments, including homes (Lee et al., 2004), schools (Linn et al., 1996),
and the workplace (Liu et al., 1995). Indoor O3 concentrations were, in general, approximately
one-tenth of the outdoor concentrations in these studies. Few indoor sources of O3 exist,
possible sources being office equipment (e.g., photocopiers, laser printers) and air cleaners.
As described in Section 3.8 of this document, O3 in the indoor environment is largely dependent
on the outdoor ambient O3 concentration. Other factors that influence indoor O3 concentrations
include air exchange rate, outdoor infiltration, indoor circulation rate, and O3 removal processes.
Sheppard (2005) described the relationship between panel studies and time-series studies
and discussed the use of personal exposure measurements versus ambient concentrations in these
designs, with the main focus being on nonreactive pollutants (e.g., PM). She noted that non-
ambient exposures typically varied across individuals but were not likely to have strong temporal
correlations. In contrast, ambient concentrations across individuals should be highly correlated,
7-6
-------
as they tend to vary over time similarly for everyone because of changes in source generation,
weather, and season. The independence of ambient and non-ambient exposure sources has
important implications for selection of study designs that are most effective for estimating health
effects (Sheppard, 2005). A simulation study by Sheppard et al. (2005) examining nonreactive
pollutants found that there was no noticeable difference between effect estimates using either
total personal exposure or ambient concentration data when non-ambient source exposures were
independent of ambient source exposures in time-series studies. In the case of O3, there are
limited non-ambient sources; thus, the non-ambient source exposures are likely to be
independent from ambient source exposures. However, unlike PM, O3 is a reactive pollutant.
In applying these conclusions to O3, an additional assumption needs to be made, i.e., that its
chemical reactivity does not introduce strong temporal correlations.
Other complications for O3 in the relationship between personal exposures and ambient
concentrations include expected strong seasonal variation of personal behaviors and building
ventilation practices that can modify exposure. In addition, the relationship may be affected by
temperature (e.g., high temperature may increase air conditioning use, which may reduce O3
penetration indoors), further complicating the role of temperature as a confounder of O3 health
effects. It should be noted that the pattern of exposure misclassification error and influence of
confounders may differ across the outcomes of interest as well as in susceptible populations. For
example, those who may be suffering from chronic cardiovascular or respiratory conditions may
be in a more protective environment (i.e., with less exposure to both O3 and its confounders,
such as temperature and PM) than those who are healthy.
As discussed thoroughly in the 2004 PM AQCD (Section 8.4.5), the resulting exposure
measurement error and its effect on the estimates of relative risk must be considered. In theory,
there are three components to exposure measurement error in time-series studies as described by
Zeger et al. (2000): (1) the use of average population rather than individual exposure data;
(2) the difference between average personal ambient exposure and ambient concentrations at
central monitoring sites; and (3) the difference between true and measured ambient
concentrations. The first error component, having aggregate rather than individual exposure
data, is a Berksonian measurement error, which in a simple linear model increases the standard
error, but does not bias the risk estimate. The second error component resulting from the
difference between average personal ambient exposure and outdoor ambient concentration level
7-7
-------
has the greatest potential to introduce bias. If the error is of a fixed amount (i.e., absolute
differences do not change with increasing concentrations), there is no bias. However, if the error
is not a fixed difference, this error will likely attenuate the O3 risk estimate as personal O3
exposures are generally lower than ambient O3 concentrations. The third error component, the
instrument measurement error in the ambient levels, is referred to as nondifferential
measurement error and is unlikely to cause substantial bias.
The impact of exposure measurement error on O3 effect estimates was demonstrated in a
study by Navidi et al. (1999). In this study, two statistical designs (bidirectional case-crossover
design and multilevel analytic design) were considered. Simulations were conducted using data
from the University of Southern California Children's Health Study of the long-term effects of
air pollutants on children. The effect estimate from computed "true" O3 exposure was compared
to effect estimates from exposure determined by using several methods: (1) ambient stationary
monitors; (2) the microenvironmental approach (multiply O3 concentrations in various
microenvironments by time present in each microenvironment); and (3) personal sampling.
Effect estimates based on all three exposure measures were biased toward the null. The bias that
results when using the microenvironmental and personal sampling approach is due to
nondifferential measurement error. Use of ambient monitors to determine exposure will
generally overestimate true personal O3 exposure (because their use implies that subjects are
outdoors 100% of their time and not in close proximity to sources that reduce O3 levels such as
NO emissions from mobile sources); thus, generally, their use can result in effect estimates that
are biased toward the null if the error is not of a fixed amount.
Zidek (1997) noted that a statistical analysis must balance bias and imprecision (error
variance). Ignoring measurement error in air pollution epidemiologic studies often results in
underestimated risk estimates. For example, in a reanalysis of the study by Burnett et al. (1994)
on the acute respiratory effects of ambient air pollution, Zidek et al. (1998) noted that accounting
for measurement error, as well as making a few additional changes to the analysis, resulted in
qualitatively similar conclusions. However, while the original analysis by Burnett et al.
observed a 4.0% (95% CI: 1.9, 6.0) excess risk of daily respiratory admissions in the summer
months attributable to a 40 ppb increase in 1-h max O3 (combined effect of a 1- and 3-day lag
of O3), results from Zidek et al. indicated that a 9.4% (95% CI: 7.0, 11.8) increase was observed
for a 40 ppb increase (a change from 10 to 50 ppb) in 1-h max O3 (combined effect of a 2- and
-------
3-day lag of O3). Although the different lag periods examined in the two analyses make direct
comparisons of the effect estimates difficult, it appears that correcting the measurement error
resulted in a larger effect estimate. Available data and analysis limit our ability to weigh the
importance of uncertainty due to measurement error relative to other sources in the various
studies reviewed in this chapter.
In addition to overestimation of exposure and the resulting underestimation of effects, the
use of ambient O3 concentrations may obscure the presence of thresholds in epidemiologic
studies at the population level. Using PM2 5 as an example, Brauer et al. (2002) examined the
relationship between ambient concentrations and mortality risk in a simulated population with
specified common individual threshold levels. They found that no population threshold was
detectable when a low threshold level was specified. Even at high specified individual threshold
levels, the apparent threshold at the population level was much lower than specified. Brauer
et al. (2002) concluded that surrogate measures of exposure (i.e., those from centrally-located
ambient monitors) that were not highly correlated with personal exposures obscured the presence
of thresholds in epidemiologic studies at the population level, even if a common threshold exists
for individuals within the population.
As discussed in Section 3.9, O3 concentrations measured at central ambient monitors may
explain, at least partially, the variance of individual personal exposures; however, this
relationship is influenced by factors such as air exchange rates in housing and time spent
outdoors, which may vary by city. Other studies conducted in various cities observed that the
daily averaged personal O3 exposures from the population were well correlated with monitored
ambient O3 concentrations, although substantial variability existed among the personal
measurements. Thus, there is supportive evidence that ambient O3 concentrations from central
monitors may serve as valid surrogate measures for mean personal O3 exposures experienced by
the population, which is of most relevance to time-series studies. This is especially likely true
for respiratory hospital admission studies for which much of the response is attributable to O3
effects on asthmatics. In children, for whom asthma is more prevalent, ambient monitors are
likely to correlate reasonably well with personal exposure to O3 of ambient origin because
children spend more time outdoors in the warm season. However, of some concern for mortality
and hospitalization time-series studies is the extent to which ambient O3 concentrations are
representative of personal O3 exposures in another particularly susceptible group of individuals,
7-9
-------
the debilitated elderly, as the correlation between the two measurements has not been examined
in this population. A better understanding of the relationship between ambient concentrations
and personal exposures, as well as of the factors that affect the relationship will improve the
interpretation of ambient concentration-population health response associations observed.
Existing epidemiologic models may not fully take into consideration all the biologically
relevant exposure history or reflect the complexities of all the underlying biological processes.
Using ambient concentrations to determine exposure generally overestimates true personal O3
exposures (by approximately 2- to 4- fold in the various studies described in Section 3.9),
resulting in biased descriptions of underlying concentration-response relationships (i.e., in
attenuated risk estimates). The implication is that the effects being estimated occur at fairly low
exposures and the potency of O3 is greater than these effect estimates indicate. As very few
studies evaluating O3 health effects with personal O3 exposure measurements exist in the
literature, effect estimates determined from ambient O3 concentrations must be evaluated and
used with caution to assess the health risks of O3.
The ultimate goal of the O3 NAAQS is to set a standard for the ambient level., not personal
exposure level, of O3. Until more data on personal O3 exposure become available, the use of
routinely monitored ambient O3 concentrations as a surrogate for personal exposures is not
generally expected to change the principal conclusions from O3 epidemiologic studies.
Therefore, population health risk estimates derived using ambient O3 levels from currently
available observational studies (with appropriate caveats taking into account personal exposure
considerations) remain useful.
7.1.3.2 Ozone Exposure Indices Used
The O3-related effect estimates for mortality and morbidity health outcomes are usually
presented in this document as a relative risk, i.e., the risk rate relative to a baseline mortality or
morbidity rate. Relative risks are based on an incremental change in exposure. To enhance
comparability between studies, presenting these relative risks by a uniform exposure increment
is needed. However, determining a standard increment is complicated by the use of different O3
exposure indices in the existing health studies. The three daily O3 exposure indices that most
often appear in the literature are the maximum 1-h average within a 24-h period (1-h max), the
maximum 8-h average within a 24-h period (8-h max), and 24-h average (24-h avg)
7-10
-------
concentrations. As levels are lower and less variable for the longer averaging times, relative
risks of adverse health outcomes for a specific numeric concentration range are not directly
comparable across metrics. Using the nationwide distributional data for O3 monitors in U.S.
Metropolitan Statistical Areas, increments representative of a low-to-high change in O3
concentrations were approximated on the basis of annual mean to 95th percentile differences
(Langstaff, 2003), as follows:
Daily Exposure Index
1-h max O3
8-h max O3
24-h avg O3
Exposure Increment (ppb)
40
30
20
In the following chapter sections, efforts were made to standardize the O3 risk estimates using
these increments, except as noted. The specified incremental change for each daily O3 exposure
index ensures that risk estimates are comparable across the different metrics. For example, a 2%
excess risk for a 20 ppb change in mean 1-h max O3 was standardized and presented as a 4%
excess risk for a 40 ppb change in mean 1-h max O3. This standardized risk estimate is
approximately comparable to the excess risks observed for a 30 ppb change in mean 8-h max O3
or a 20 ppb change in mean 24-h avg O3. Thus, the different increments for each daily O3
exposure index do not represent inconsistencies; rather, they are appropriately scaled to facilitate
comparisons between the various studies that used different indices. Note that in the Chapter 7
Annex Tables (see Annex Section AX7.1), effect estimates are not standardized; there, the
results are presented in the tables as reported in the published papers.
7.1.3.3 Lag Time: Period between Ozone Exposure and Observed Health Effect
Exposure lags may reflect the distribution of effects across time in a population and the
potential mechanisms of effects. The choice of lag days for the relationship between exposure
and health effects depends on the hypothesis being tested and the mechanism involved in the
expression of the outcome. Effects can occur acutely with exposure on the same or previous
day, cumulatively over several days, or after a delayed period of a few days. With knowledge
of the mechanism of effect, the choice of lag days can be determined prior to analysis.
7-11
-------
As one example, one could expect cough to occur acutely after exposure with a lag of 0 or 1 day,
given that O3 can act as a short-term irritant. However, an O3-related inflammatory response
may not lead to asthma exacerbation until several days later. An asthmatic may be impacted
by O3 on the first day of exposure, have further effects triggered on the second day, and then
report to the emergency room for an asthmatic attack three days after exposure. Further, within
a population of asthmatics, exacerbation of asthma symptoms may be observed over a period of
several days, since each asthmatic may have varying individual aspects of the disease and may
be affected by the exposure differently depending on his/her sensitivity and disease severity.
The results from controlled human studies may be useful in assessing the adequacy of lags for
some respiratory health outcomes.
Some studies attempted to examine the overall impact of O3 through distributed lag
models. The single-day lag model calculates a risk estimate that assumes dependence only on
exposure from the specified day. In contrast, the distributed lag model provides an estimate that
is a summary measure of the cumulative distributed lag effect from all included lag days. The
standard error of the cumulative sum of the individual distributed lag coefficients takes into
consideration the variance-covariance of the multiple lags, and it is generally larger than the
standard error of the single-day lag coefficient due to positive auto-correlation. Thus, if the
underlying O3-health outcome relationship was a single-day effect, then modeling the
relationship with a distributed lag model would make the estimate less significant. On the other
hand, if the effect of O3 on health outcomes persisted over several days, then applying a single-
day lag model would result in an underestimation of the multiday effects, although the single-
day estimate may still reflect a portion of the multiday effect, to the extent that daily O3 levels
are auto-correlated. Choosing a lag model requires balancing variance and bias, and the
usefulness (i.e., ease of interpretation, comparability across studies) of the model, as well as
knowledge about mechanisms underlying effects for a given health outcome. Sufficient
information is not available to judge the adequacy of a given model for different outcomes.
As the parameters estimated from single-day lag versus multiday lag models are not the
same, interpretation and comparison of these results will be difficult. When comparing the
impacts of these different models, the nuance of increments used in calculating the estimates is
different depending on the model. For example, an excess percent mortality risk "per 20 ppb
increase in 24-h avg O3" in a distributed lag model including lag 0- through 6-days tacitly means
7-12
-------
a 20 ppb increase in each of the 7 days. The difference in the exposure scenarios in the single-
day versus multiday lag model (i.e., 20 ppb increase in one day versus several consecutive days)
complicates a simple comparison of risk estimates from two different models using "the same
increment." Thus, it should be recognized that these are distinct models that summarize the
underlying process differently.
Only a limited number of studies have hypothesized a priori the lag structure to be
examined. Most of the O3 time-series studies examined relatively small numbers of single-day
lag models, typically lags of 0 through 3 days. Sheppard et al. (1999) noted that when
considering single-day lag estimates it is important to consider the effect estimate in the context
of the pattern of adjacent lags as these estimates contain information from the adjacent days
owing to serial correlation of the pollutant series. In many cases, a pattern of positive
associations across several lag days was reported. For the respiratory and cardiovascular
outcomes investigated, the "most significant" lags were generally 0- or 1-day lags, suggesting
that the majority of the single-day associations are immediate, not a random pattern in which
associations can be observed on any of the lags examined with equal probabilities. For example,
two recent meta-analyses of O3-mortality effects observed that the combined estimate from
0-day lag models was larger than the estimate from longer lag days (Bell et al., 2005; Levy
et al., 2005).
Bias resulting from the selection of lags has not been examined specifically for O3 effects.
However, the issue of lags has been investigated for PM and the results of this analysis are most
likely of relevance for O3. Lumley and Sheppard (2000) performed a simulation study to
examine model selection bias in air pollution epidemiology using PM2 5 as an example.
Sheppard et al. (1999; reanalysis Sheppard, 2003) had investigated the association between
asthma hospital admissions and ambient PM2 5 concentrations over an eight-year period in
Seattle, WA. Note that the results from Lumley and Sheppard (2000) and Sheppard et al. (1999)
were based on Generalized Additive Model (GAM) analyses using default convergence criteria
(see Section 7.1.3.7). A negative control analysis, using simulated data with no association
between PM exposure and the health outcome, and a positive control analysis, using a specified
non-zero excess risk added to the simulation, were performed for comparison. The bias from
selection of the best of seven lags (0 to 6 days) and residual seasonal confounding in the negative
control analysis (median log relative risk of 0.0013) was approximately half the log relative risk
7-13
-------
estimated from the observed data (0.0027), after adjusting for season and temperature. In the
positive control model (true log relative risk of 0.0083), the bias was small (median log relative
risk of 0.0080). Results from these simulations indicate that bias from selection of lags may be
small, but of the same magnitude as the estimated health impacts.
Selection of lag periods should depend on the hypothesis of the study and the potential
mechanism of the effect. When the mechanism of the health effect is unknown, investigating the
association between outcome and exposure using cumulative distributed lag models may be
informative. Analyzing a large number of lags and simply choosing the largest and most
significant results may bias the air pollution risk estimates away from the null. Most studies
have shown that O3 has a fairly consistent, immediate effect on health outcomes, including
respiratory hospitalizations and mortality. Several studies also observed significant O3 effects
over longer cumulative lag periods, suggesting that in addition to single-day lags, multiday lags
should be investigated to fully capture a delayed O3 effect on health outcomes. In this document,
discussion largely focuses on effect estimates from 0- and 1-day lags, with some consideration of
cumulative, multiday lag effects. It is not straightforward to compare and contrast results from
single-day versus multiday lag models, because the parameters estimated from these models are
not the same. These complications need to be taken into consideration when interpreting results
from various lag models.
7.1.3.4 Model Specification to Adjust for Temporal Trends and Meteorologic Effects
Several challenges are encountered with respect to designing and interpreting time-series
studies. The principal challenge facing the analyst in the daily time-series context is avoiding
bias due to confounding by short-term temporal factors operating over time scales from days to
seasons. In the current regression models used to estimate short-term effects of air pollution,
two major potential confounders generally need to be considered: (1) seasonal trend and other
"long-wave" temporal trends; and (2) weather effects. Both of these variables tend to predict a
significant fraction of fluctuations in time-series. Unfortunately, as O3 has strong seasonal
cycles and is formed more at higher temperatures, both terms are also highly correlated with O3.
The correlation of O3 with these confounding terms tends to be higher than that for PM or other
gaseous pollutants. In the United States, the mass concentration of PM2 5 generally does not
have strong seasonal cycles like O3, because PM2 5 tends to reflect both primary emissions
7-14
-------
(throughout the year, but often higher in winter in most U.S. cities) and secondary aerosols
(higher in summer). Therefore, PM2 5 and O3 effect estimates from studies primarily designed to
examine PM2 5 health effects may not be comparable, because model specifications that may be
appropriate for PM2 5 may not necessarily be adequate for O3.
An examination of recent time-series studies indicates that several types of fitting
approaches have been used to adjust for temporal trends and weather effects. The use of
parametric and nonparametric smoothers with varying degrees of freedom per year has emerged
as the prevailing approach. The use of larger degrees of freedom to adjust for potential
confounding by time-varying factors may inadvertently result in ascribing more effects to these
unmeasured potential confounders and mask the air pollution effect. Often smaller pollution
effect estimates are observed when more degrees of freedom are used. Currently, the degrees of
freedom used to adjust for temporal trends in time-series studies generally range from 4 to
12 degrees of freedom per year using either nonparametric or parametric smoothers. Statistical
diagnostics such as Akaike's Information Criteria, residual autocorrelation, or dispersion of the
regression model often are used to choose or to evaluate the adequacy of the degrees of freedom
for temporal trend. However, these diagnostics do not guarantee "adequate" control for temporal
confounding, as the appropriate extent of smoothing is not identifiable from the data and the
proper selection of smoothing parameters requires prior knowledge of the nature of the
confounding (e.g., shape and duration of influenza epidemics).
The issue of model specifications to adjust for temporal trends and weather variables in
time-series studies was a consideration of several researchers that conducted sensitivity analyses
of PM data (Health Effects Institute, 2003). The sensitivity of O3 coefficients to model
specifications for temporal trend adjustment has not been as well-studied. Recent multicity
studies examined the sensitivity of O3 coefficients to the extent of smoothing for adjustment of
temporal trends and meteorologic factors (Bell et al., 2004; Huang et al., 2005; Ito et al., 2005).
Most, if not all, O3 studies used the same model specifications to estimate the excess risks for
PM and other gaseous pollutants. The model specification designed to control confounding by
meteorological and temporal factors for PM may not necessarily be adequate for O3. As noted
above, O3 is expected to have the strongest correlation with both temporal (seasonal) trend and
weather effects. The strong annual cycle in O3 concentrations presents a unique problem in
time-series analyses where time trends are fitted simultaneously with pollution and other model
7-15
-------
terms (i.e., co-adjustment). In this setting, the annual O3 cycle itself may compete with the
smooth function of time to explain some of the annual, cyclic behavior in the health outcome,
which can result in biased effect estimates for O3 when data for all seasons are analyzed
together.
Current weather models used in time-series analyses can be classified by their use of:
(1) quantile (e.g., quartile, quintile) indicators; (2) parametric functional forms such as V- or
U-shape functions; and (3) parametric (e.g., natural splines) or nonparametric (e.g., locally
estimated smoothing splines [LOESS]) smoothing functions. More recent studies tend to use
smoothing functions. While these methods provide flexible ways to fit health outcomes as a
function of temperature and other weather variables, there are two major issues that need further
examination to enable more meaningful interpretation of O3 morbidity and mortality effects.
The first issue is the interpretation of weather or temperature effects. Most researchers
agree about the morbidity and mortality effects of extreme temperatures (i.e., heat waves or cold
spells). However, as extreme hot or cold temperatures, by definition, happen rarely, much of the
health effects occur in the mild or moderate temperature range. Given the significant correlation
between O3 and temperature, ascribing the association between temperature and health outcomes
solely to temperature effects may underestimate the effect of O3. The second issue is that
weather model specifications are fitted for year-round data in most studies. Such models will
ignore the correlation structure that can change across seasons, resulting in inefficiency
and model mis-specification. This is particularly important for O3, which appears to change
its relationship with temperature as well as with other pollutants across seasons.
This changing relationship between O3 and temperature, as well as between O3 and other
pollutants across seasons, and its potential implications for health effects modeling have not
been examined thoroughly in the time-series literature. Even the flexible smoother-based
adjustments for seasonal and other time-varying variables cannot fully take into account these
complex relationships. One obvious way to alleviate or avoid this complication is to analyze
data by season. While this practice reduces sample size, its extent would not be as serious as for
PM (which is collected only every sixth day in most locations) because O3 is collected daily,
though only in warm seasons in some states. An alternative approach is to include separate O3
concentration variables for each season (by multiplying O3 concentrations by a season indicator
7-16
-------
variable). However, this approach assumes that all effects in the model that are not indicated to
be season-specific do not vary seasonally.
In locations where seasonal variability may be a factor, O3 effect estimates calculated using
year-round data can be misleading, as the changing relationship between O3, temperature, and
other pollutants across seasons may have a significant influence on the estimates. Analyses have
indicated that confounding from seasonal variability may be controlled effectively by stratifying
the data by season.
7.1.3.5 Confounding Effects of Copollutants
Extensive discussion of issues related to confounding effects among air pollutants in time-
series studies are provided in Section 8.4.3 of the 2004 PM AQCD (U.S. Environmental
Protection Agency, 1996b). Since the general issues discussed there are applicable to all
pollutants, such discussions are not repeated here. What was not discussed in the 2004 PM
AQCD was the issue of changing relationships among air pollutants across seasons. Compared
to other pollutants, O3 has strong seasonal cycles. Ambient O3 levels are typically higher in the
summer or warm season, often referred to as the O3 season. In the winter or colder months, O3
levels tend to be much lower than in the summer months. During the winter in some urban
locations, O3 mainly comes from the free troposphere and can be considered a tracer for
relatively clean air (i.e., cold, clear air coming down from the upper atmosphere), as discussed in
Chapter 3 of this document. The clean air is associated with the passage of cold fronts and the
onset of high-pressure conditions, which occur with colder temperatures. Thus, sunny clear
winter days following a high-pressure system are the days when air pollution levels from
primary emissions (e.g., NO2, SO2, and PM from local sources) tend to be lower and O3 is
relatively higher. This can lead to negative correlations between O3 and the primary pollutants
in the winter. As shown in Figure 3-21 in Chapter 3, the relationship between O3 and PM2 5 was
U-shaped for the year-round data in Fort Meade, MD. The negative PM2 5/O3 slope was in the
range of O3 concentrations less than 30 ppb, providing supporting evidence of the
aforementioned winter phenomenon. Thus, the correlation between O3 and PM for year-round
data may be misleading. The high reactivity of O3 with certain copollutants further complicates
the analysis and interpretation. For example, the reaction between NO (emitted from motor
vehicles) and O3 results in reduced O3 levels but increased NO2 levels during high traffic periods.
7-17
-------
Multipollutant regression models often are used to assess potential confounding by
copollutants; however, there are limitations to these models. Zidek et al. (1996) examined,
through simulation, the joint effects of multicollinearity and measurement error in a Poisson
regression model. The results illustrated the transfer of effects from the "causal" variable to the
confounder. However, in order for the confounder to have a larger effect size than the true
predictor, the correlation between the two covariates had to be very high (r > 0.9), with moderate
error (a > 0.5, where a is the probability of rejecting the null hypothesis given that the null is
true) for the true predictor and no error for the confounder in their scenarios. The transfer-of-
causality effect was lessened when the confounder also became subject to error. Another
interesting finding was the behavior of the standard errors of the coefficients. When the
correlation between the covariates was high (r = 0.9) and both covariates had no error, the
standard errors for both coefficients were inflated by a factor of two; however, this phenomenon
disappeared when the confounder had error. The effect of multicollinearity is generally even
more complex when analyzing real data. For further discussion, see the 2004 PM AQCD
(Sections 8.4.3 and 8.4.5).
Uncertainty remains as to the use of multipollutant regression models in assessing the
independent health effects of pollutants that are correlated. Particularly in the case of O3,
concern remains as to whether multipollutant regression models for year-round data can adjust
for potential confounding adequately because of the changing relationship between O3 and other
pollutants. Despite these limitations, use of multipollutant models is still the prevailing
approach employed in most, if not all, studies of O3 health effects; and it serves as an important
tool in addressing the issue of confounding by copollutants, especially in season-stratified
analyses.
7.1.3.6 Hypothesis Testing and Model Selection in Ozone Epidemiologic Studies
Epidemiologic studies investigated the association between various measures of O3 (e.g.,
multiple lags, different metrics, etc.) and various health outcomes using different model
specifications. Statistically testing a null hypothesis (i.e., there is no effect of O3) requires one to
calculate the value of a test statistic (i.e., the t-value). If the observed test statistic exceeds a
critical value (oftentimes the 95th percentile) or is outside a range of values, the null hypothesis
is rejected. However, when multiple testing is done using a critical value determined for a single
7-18
-------
test, the chance that at least one of the hypotheses is significant may be greater than the specified
error rate. Procedures are available to ensure that the rejection error rate does not exceed the
expected error rate (usually 5%) when conducting multiple hypothesis testing. However, often
the multiple hypotheses being tested are not statistically independent, thus some corrections,
such as the Bonferroni adjustment, may be overly conservative.
Multiple hypothesis testing and model selection also contribute to publication bias.
Publication bias is the tendency of investigators to submit and/or editors to accept manuscripts
for publication based on the strength of the study findings. Although publication bias commonly
exists for many topics of research, it may be present to a lesser degree in the air pollution
literature. Several air pollutants often are examined in a single study, leading to the publication
of significant, as well as nonsignificant, individual pollutant results. For example, many air
pollution papers with a focus on PM health effects also published O3 results. Ozone was largely
considered a potentially confounding copollutant of PM; thus, O3 effect estimates were often
presented regardless of the statistical significance of the results. Another aspect of publication
bias is only selecting the largest or statistically strongest effect estimate to report and not the
array of models evaluated. Bell et al. (2005) conducted a meta-analysis of O3 health effects
studies and found that studies reporting a single estimate compared to multiple estimates
generally showed larger effects for the same lag. They also examined the presence of this bias
by comparing the results from the meta-analysis to the multicity National Morbidity, Mortality,
and Air Pollution Study (NMMAPS). When comparing the marginal posterior distributions of
the overall effect under the meta-analysis and in NMMAPS for the eight U.S. cities common to
both the approaches, the meta-analysis effect estimate of same-day O3 concentration was nearly
2-fold that of the NMMAPS pooled effect estimate.
Testing multiple hypotheses may, at times, be appropriate. For example, without an a
priori biologic model that definitively specifies which of several lag periods or meteorologic-
control models should take priority, multiple hypotheses may need to be developed for
researchers to explore more thoroughly potential associations for an O3-related health effect
(Goodman, 2005). In this case, it would be useful to state which hypotheses are confirmatory
versus exploratory. The key concern is the reporting of a single estimate or estimates from one
model. Presentation of an array of models, exposure definitions, and outcome measures may be
more appropriate in communicating the results. Sensitivity analyses, which are critical for
7-19
-------
model validation, also involve multiple testing. When conducting sensitivity analyses, one
should guard against multiple testing errors by restricting the inferences to consistency of
the effect.
One method to address model uncertainty and multiple hypothesis testing is Bayesian
model averaging. In Bayesian model averaging, predictions and inferences are based on a set of
models rather than a single model, and each model contributes proportionally to the support it
receives from the observed data (Clyde, 1999). In addition to the uncertainty of effect
estimation, Bayesian model averaging can incorporate uncertainty regarding the choice of
confounding variables, pollutants, and lags. Koop and Tole (2004) used Bayesian model
averaging to analyze the effect of various air pollutants, including O3, SO2, CO, NO, NO2,
PM10_2 5, and PM2 5, on mortality in Toronto, Canada. The 50+ explanatory variables required the
fitting of an enormous number of potential models. Although the point estimates for all
pollutants were positive, very small effects were found. However, in the context of the many
interaction terms, meteorological variables, smoothing surfaces, and the relatively loose
posterior distribution, the analysis results cannot be interpreted meaningfully. Clyde (2000) and
Clyde et al. (2000) also used Bayesian model averaging to analyze the relationship between PM
and mortality. Clyde (2000) noted that Bayesian model averaging did not take into consideration
factors that might bias the estimated effect toward the null. For example, measurement error in
the exposure variables was not considered. In addition, the Poisson model (similar to many
other regression models) assumed that all individuals in a population had equal risks, including
potentially susceptible and vulnerable populations such as those with respiratory illnesses and
outdoor workers. While Bayesian model averaging can theoretically be used to take into account
uncertainty, claims of causality based on observational studies may be highly sensitive to the
choice of prior distributions and class of models under consideration (Clyde et al., 2000).
Another limitation of Bayesian model averaging is that the estimated posterior effects may be
diluted (i.e., result in smaller coefficients) when variables are highly correlated, as may be the
case for air pollution studies (George, 1999 in comments to Hoeting et al., 1999).
Additional methods to control for multiple hypothesis testing are by deciding a priori
which hypotheses are confirmatory and exploratory, and then limiting the number of
confirmatory tests. For example, Dominici et al. (2003) used a minimum number of tests in the
U.S. 90 cities study, which reduced the uncertainty associated with multiple testing. In addition,
7-20
-------
they performed sensitivity analyses to examine the consistency and robustness of the effects.
Another approach is to partition the data into two sets, one for model identification and a second
for model confirmation.
The summary of health effects in this chapter is vulnerable to the errors of publication bias
and multiple testing. Efforts have been made to reduce the impact of multiple testing errors on
the conclusions in this document. To address multiple hypothesis testing, emphasis will be
placed in this chapter on a priori hypotheses. As identifying a priori hypotheses is difficult in the
majority of the studies, the most common hypotheses will be considered. For example, although
many studies examined multiple single-day lag models, priority would be given to effects
observed at 0- or 1-day lags rather than at longer lags. Both single- and multiple-pollutant
models that include O3 will be considered and examined for robustness of results. Analyses of
multiple model specifications for adjustment of temporal or meteorological trends will be
considered sensitivity analyses. Sensitivity analyses shall not be granted the same inferential
weight as the original hypothesis-driven analysis; however, these analyses will be discussed in
this chapter as appropriate given their valuable insights that may lead scientific knowledge in
new directions.
7.1.3.7 Impact of Generalized Additive Models Convergence Issue on Ozone Risk
Estimates
Generalized Additive Models (GAM) have been widely utilized for epidemiologic analysis
of the health effects attributable to air pollution. The impact of the GAM convergence issue was
thoroughly discussed in Section 8.4.2 of the 2004 PM AQCD. Reports have indicated that using
the default convergence criteria in the Splus software package for the GAM function can lead to
biased regression estimates for PM and an underestimation of the standard error of the effect
estimate (Dominici et al., 2002; Ramsay et al., 2003). GAM default convergence criteria has a
convergence precision of 10~3 and a maximum number of 10 iterations. The more stringent
convergence criteria refers to increased stringency of both the convergence precision and
number of iterations. The use of default convergence criteria was found to be a problem when
the estimated relative risks were small and two or more nonparametric smoothing curves were
included in the GAM (Dominici et al., 2002). The magnitude and direction of the bias depend in
part on the concurvity of the independent variables in the GAM and the magnitude of the risk
7-21
-------
estimate. Recent focus has been on the influence of the GAM function on effect estimates for
PM. However, because O3 covaries more strongly with both weather and time factors than does
PM, the issue of GAM convergence criteria for O3 also needs to be considered.
A meta-analysis by Stieb et al. (2003) found some difference in O3-mortality risk estimates
between the GAM and non-GAM studies. GAM studies were defined as studies that analyzed
effect estimates using nonparametric smoothing functions of time or weather. Non-GAM studies
were all other studies, including those using Generalized Linear Models (GLM) and Generalized
Estimating Equations (GEE) in their analysis. In the single-pollutant models, the O3-mortality
risk estimates for the non-GAM studies (10 estimates) and GAM studies (15 estimates) were
1.8% (95% CI: 0.5, 3.1) and 2.2% (95% CI: 1.4, 2.8), respectively, per 40 ppb daily 1-h
max O3. In the multipollutant models, the pooled risk estimate was 1.0% (95% CI: -0.5, 2.6)
for non-GAM studies (7 estimates) and 0.5% (95% CI: -1.0, 1.9) for GAM studies
(4 estimates).
Results from recent meta-analyses of O3-mortality effects suggest that there are no
substantial differences between GAM-affected estimates and non-GAM-affected estimates (Bell
et al., 2005; Ito et al., 2005; Levy et al., 2005). GAM-affected studies included those that used
default convergence criteria. Non-GAM-affected studies included GAM studies that used
stringent convergence criteria or those that used other modeling techniques. Ito et al. (2005)
found that the single-pollutant combined estimate for the GAM-affected studies (15 estimates)
and non-GAM-affected studies (28 estimates) were 1.92% (95% CI: 1.02, 2.81) and 1.40%
(95% CI: 0.78, 2.02), respectively, per 20 ppb increase in 24-h avg O3. In the analysis by Levy
et al. (2005), the single-pollutant combined estimate for the GAM-affected studies (29 estimates)
and non-GAM-affected studies (17 estimates) were 1.56% (95% CI: 1.01, 2.11) and 1.80%
(95% CI: 1.17, 2.43), respectively, per 40 ppb increase in 1-h max O3. Bell et al. (2005) also
reported that the pooled estimate was larger for the studies that were not GAM-affected.
A few GAM studies reanalyzed O3 risk estimates using more stringent convergence criteria
or GLM. Reanalysis of an asthma hospital admissions study in Seattle, WA (Sheppard et al.,
1999; reanalysis Sheppard, 2003) indicated that there were only slight changes in the risk
estimates when using more stringent convergence precision (10~8) in GAM. The original GAM
analysis indicated an excess risk of 9% (95% CI: 3, 17), whereas the stringent GAM analysis
found an excess risk of 11% (95% CI: 3, 19) per 30 ppb increase in 8-h max O3 at a 2-day lag.
7-22
-------
Similar results were found using GLM with natural splines: 11% (95% CI: 2, 20). In the
reanalysis of Santa Clara County, CA data, Fairley (1999; reanalysis Fairley, 2003) used the
same methods as the original analysis except the convergence precision (e) was increased
from 10~4 to 10~12 and the maximum number of iterations (M) were increased from 10 to 107.
The O3-mortality risk estimate slightly increased from 2.8% (95% CI not provided) using default
GAM parameters to 2.9% (95% CI: -0.3, 6.0) using stringent GAM parameters per 30 ppb
increase in 8-h max O3 at a 0-day lag. The O3-mortality risk estimates further increased to 3.0%
(95% CI: -0.3, 6.3) using GLM with natural cubic splines. In the reanalysis of the Netherlands
data by Hoek et al. (2000; reanalysis Hoek, 2003), the O3 nonaccidental mortality risk estimates
increased from 1.3% (95% CI: 0.8, 1.9) using default GAM to 1.5% (95% CI: 1.0, 2.1) using
stringent GAM (e = 10~8, M = 103) and 1.6% (95% CI: 0.9, 2.4) using GLM with natural splines
per 30 ppb increase in 8-h avg O3 (12 p.m.-8 p.m.) at a 1-day lag.
In the limited number of studies that have reanalyzed O3 risk estimates, there is little
evidence that default GAM analyses resulted in positively biased estimates, as was observed for
PM. Generally, it appears that the use of default convergence criteria in GAM tends to bias risk
estimates toward the null, in addition to underestimating the standard errors. However, one
study by Cifuentes et al. (2000) in Santiago, Chile observed a large difference in the O3-
mortality excess risks calculated using default GAM (0.9% [95% CI: 0.2, 1.6] per 40 ppb
increase in 1-h max O3) and GLM (0.1% [95% CI: -0.6, 0.8]). The GAM convergence problem
appears to vary depending on data sets, and likely depends upon the intercorrelation among
covariates and the magnitude of the risk estimate; thus, its impact on the results of individual
studies cannot be known without a reanalysis. Consistent with the approach used in the 2004
PM AQCD, the results from studies that analyzed data using GAM with default convergence
criteria and at least two nonparametric smoothing terms are generally not considered in this
chapter, with some exceptions as noted.
7.1.3.8 Summary of Considerations in the Interpretation of Ozone Epidemiologic Studies
The previous sections discussed the topics of exposure assessment and model specification
in O3 epidemiologic studies. Also examined were the issues of hypothesis testing and model
selection, as well as the impact of the GAM convergence issue on O3 risk estimates.
7-23
-------
• Exposure measurement error may result from the use of stationary ambient monitors
as an indicator of personal exposure in population studies. The relationship between
individual personal O3 exposures and ambient O3 concentrations was found to be
affected by factors such as air exchange rates in housing and time spent outdoors,
which varied by individual or by city. However, ambient O3 concentrations, though
they do tend to overestimate personal exposures, were found in general to be well
correlated to daily averaged personal O3 exposures for the populations studied.
Therefore population risk estimates derived from ambient O3 concentrations remain
useful and appropriate, but should be evaluated and used with caution in O3 health
risk assessments.
• The three daily O3 exposure indices used most often in O3 epidemiologic studies are
1-h max O3, 8-h max O3, and 24-h avg O3. To facilitate comparison of health effects
calculated using the different O3 metrics, risk estimates are standardized or scaled by
specified exposure increments: 40 ppb for 1-h max O3, 30 ppb for 8-h max O3, and
20 ppb for 24-h avg O3.
• The lag period between O3 exposure and the observed health effect may reflect the
distribution of effects across time in a population and the potential mechanisms of
effects. Among the single-day lags examined, the strongest health effects were
observed with a 0- or 1-day lag in O3 exposure. Multiday lags of O3 exposure also
were investigated. As the parameters estimated from single-day versus multiday
models are not the same, interpreting results from these distinct lag models
requires caution.
• In time-series studies estimating short-term effects of air pollution, temporal trends
and weather effects are two major potential confounders that require consideration.
Of all the air pollutants, O3 is expected to have the strongest correlation with both
temporal and meteorological factors. The use of parametric and nonparametric
smoothers with varying degrees of freedom per year has been the prevailing approach.
Confounding from seasonal variability may be controlled effectively by stratifying
analyses by season.
• A major methodological issue affecting O3 epidemiologic studies concerns the
evaluation of the extent to which other air pollutants may confound or modify
O3-related effect estimates. The changing relationship between O3 and copollutants
across seasons further complicates the issue. The use of multipollutant regression
models is the prevailing approach for controlling potential confounding
by copollutants in O3 health effects studies.
• In air pollution epidemiologic studies, multiple hypotheses are often tested and certain
models are selected for presentation. In many cases, the multiple hypotheses being
tested area not independent and, thus, some correlations (e.g., the Bonferroni
adjustment) may be overly conservative. To reduce the potential impact of multiple
testing errors, emphasis should be placed on a limited number of a priori hypotheses.
Additional analyses of varying model specifications considered along with sensitivity
analyses.
7-24
-------
GAM is widely used in epidemiologic analyses of air pollution health effects.
Recent reports have indicated that the commonly used default convergence criteria
lead to biased regression estimates for PM. The limited number of studies that have
evaluated this issue for O3 observed little evidence that default GAM analyses resulted
in positively biased estimates.
7.1.4 Approach to Presenting Ozone Epidemiologic Evidence
To produce a thorough appraisal of newly available evidence, key information (including
study design, analysis, mean O3 concentrations, and health outcome results) from important new
studies is presented in summary tables in the Chapter 7 Annex (Annex Section AX7.1). Each
section of this chapter starts by concisely highlighting important points derived from the 1996 O3
AQCD assessment. In the main body of the chapter, particular emphasis is focused on studies
and analyses that provide pertinent information for the critical assessment of health risks from O3
exposure. Not all studies are accorded equal weight in the overall interpretive assessment of
evidence regarding O3-associated health effects. Among well-conducted studies with adequate
control for confounding, increasing scientific weight is accorded in proportion to the precision of
their effect estimates. Small-scale studies without a wide range of exposures generally produce
less precise estimates compared to larger studies with a broad exposure gradient. The size of the
study, as indicated by the length of the study period and total number of events, and the
variability of O3 exposures are important components that help to determine the precision of the
health effect estimates. In evaluating the epidemiologic evidence in this chapter, more weight is
accorded to estimates from studies with narrow confidence bands.
Emphasis is placed in the text on the discussion of (1) new multicity studies that employ
standardized methodological analyses for evaluating O3 effects across several or numerous cities
and often provide overall effect estimates based on combined analyses of information pooled
across multiple cities; (2) studies that consider O3 as a component of a complex mixture of air
pollutants including PM and other gaseous criteria pollutants (CO, NO2, SO2); and (3) North
American studies conducted in the United States or Canada. Multicity studies are of particular
interest and value due to their evaluation of a wider range of O3 exposures and large numbers of
observations. They generally provide more precise effect estimates than most smaller scale
studies of single cities. Compared to meta-analyses of multiple "independent" studies, a
potential advantage of multicity studies is their consistency in data handling and model
7-25
-------
specifications, which eliminates variation due to analysis approach. Also, unlike meta-analyses,
they do not suffer from potential omission of nonsignificant results due to "publication bias."
Furthermore, geographic patterns of air pollution effects have the potential to provide especially
valuable evidence regarding relative homogeneity and/or heterogeneity of O3 health effects
relationships across geographic locations. Due to the potential for confounding by copollutants,
preference is given to studies with effect estimates from multipollutant models, i.e., models with
both O3 and PM rather than O3-only models. The potential impacts of different health care
systems and the underlying health status of populations also need to be accounted for in the
assessment (Hubbell et al., 2005; Levy et al., 2001); thus, U.S. studies are emphasized over
non-U.S. studies. In accordance to the emphasis placed on the O3 epidemiologic studies in this
chapter, the tables in the Chapter 7 Annex Section AX7.1 were organized by region, with
multicity studies in each region presented first.
In the coming sections, field/panel studies and studies of emergency department visits and
hospital admissions, which contributed to the establishment of the revised 1997 NAAQS for O3,
are presented first. This is followed by a discussion of O3-related mortality and effects of
chronic exposures to O3. The chapter ends with an integrative discussion providing a summary
and conclusions.
7.2 FIELD STUDIES ADDRESSING ACUTE EFFECTS OF OZONE
7.2.1 Summary of Key Findings on Field Studies of Acute Ozone Effects
from the 1996 Ozone AQCD
In the 1996 O3 AQCD, individual-level camp and exercise studies provided useful
quantitative information on the concentration-response relationships linking human lung
function declines with ambient O3 concentrations. The available body of evidence supported a
dominant role of O3 in the observed lung function decrements. Extensive epidemiologic
evidence of pulmonary function responses to ambient O3 has been derived from camp studies.
Six studies from three separate research groups provided a combined database on individual
concentration-response relationships for 616 children (mostly healthy, nonasthmatic) ranging in
age from 7 to 17 years, each with at least six sequential measurements of FEVj (forced
expiratory volume in 1 second) while attending summer camps (Avol et al., 1990; Higgins et al.,
7-26
-------
1990; Raizenne et al., 1987, 1989; Spektor et al., 1988a, 1991). In the combined reanalysis by
Kinney et al. (1996a) using consistent analytical methods, these data yielded an average
relationship between afternoon FEVj and concurrent-hour O3 concentration of -0.50 mL/ppb
(95% CI: -0.63, -0.36), with study-specific slopes ranging from -1.29 to -0.19 mL/ppb.
At the time of the afternoon lung function measurement, mean 1-h avg O3 concentrations ranged
from 53 to 123 ppb (maximum range 95 to 245 ppb). Exposure in camp studies usually extended
for multiple hours. Although the regression results noted above were based on 1-h O3 levels,
single- and multiple-hour averages were observed to be highly correlated; thus, these results
might represent, to some extent, the influence of multihour exposures. In addition to the camp
study results, two studies involving lung function measurements before and after well-defined
exercise events in adults yielded concentration-response slopes of -0.4 mL/ppb (95% CI: -0.7,
-0.1)(Selwynetal., 1985) and-1.35 mL/ppb (95% CI: -2.04, -0.66) (Spektor et al., 1988b).
Ozone concentrations during exercise events of approximately !/2-hour duration ranged from 4 to
135 ppb in these studies.
Results from other field/panel studies also supported a consistent relationship between
ambient O3 exposure and acute respiratory morbidity in the population. Respiratory symptoms
(or exacerbation of asthma) and decrements in peak expiratory flow (PEF) occurred with
increased ambient O3 concentrations, especially in asthmatic children (Lebowitz et al., 1991;
Krzyzanowski et al., 1992). The results showed greater responses in asthmatic than in
nonasthmatic individuals (Lebowitz et al., 1991; Krzyzanowski et al., 1992), indicating that
asthmatics might constitute a sensitive group in epidemiologic studies of oxidant air pollution.
Since the 1996 O3 AQCD, new research has examined a broad scope of field studies as
discussed next.
7.2.2 Introduction to Recent Field Studies of Acute Ozone Effects
Numerous field studies carried out over the past decade have tested for and, in many cases,
observed acute associations between O3 concentrations and measures of respiratory ill-health and
O3 concentrations in groups of subjects (Table AX7-1 in Chapter 7 Annex Section AX7.1).
Acute field studies are distinguished from time-series studies in that they are designed to recruit
and collect data from individual human subjects instead of utilizing administrative data on
aggregate health outcomes such as daily mortality, hospital admissions, or emergency
7-27
-------
department visits. Although individual-level health outcome data are collected in field studies,
ambient O3 concentrations from centrally located monitoring stations are generally used to assess
exposure. Because of the logistical burden associated with direct data collection from individual
subjects, field/panel studies tend to be small in both numbers of subjects and in duration of
follow-up. While this may limit the statistical power of field studies as compared with time-
series studies, the ability to determine individual-level information on health outcomes and
potentially confounding factors adds scientific value.
The most common outcomes measured in acute field studies on the effects of air pollution
exposure are lung function and various respiratory symptoms. Other respiratory outcomes
examined on a limited basis include inflammation and generation of hydroxyl radicals in the
upper airways, as well as numbers of school absences. Also, several studies examined
cardiovascular outcomes, including heart rate variability (HRV) and risk of myocardial
infarctions (MI). The first group of studies provided varying degrees of evidence supporting the
conclusion that elevated O3 levels have negative impacts on lung function and symptoms,
confirming and adding to the body of knowledge that was presented in the 1996 O3 AQCD.
Some emphasis has been placed in examining the independent role of O3 in the presence of PM
and other pollutants. The other new studies contribute information regarding possible O3-related
cardiopulmonary outcomes that have not previously been as well documented.
7.2.3 Effects of Acute Ozone Exposure on Lung Function
As discussed in the 1996 O3 AQCD and in the earlier chapter of this document on
controlled human exposure studies (Chapter 6), a large body of literature from clinical and field
studies has clearly and consistently demonstrated reversible decrements in pulmonary function
following acute O3 exposure. Significant O3-induced spirometric and symptom responses have
been observed in clinical studies of exercising healthy young adults (see Section 6.2) and in
some potentially susceptible subpopulations, namely asthmatics and children (see Sections 6.3.2
and 6.5.1). Field studies of acute O3 exposure that examine pulmonary function fall into two
distinct groupings, those that conduct spirometry (measuring FEVb FVC [forced vital capacity],
and other spirometric indices) and those that measure PEF using peak flow meters. Results from
the previous O3 AQCD and Chapter 6 of this document support the conclusion that the
spirometric parameter, FEVl3 is a strong and consistent measure of lung function and may be
7-28
-------
used in the assessment of asthma (Fuhlbrigge et al., 2001). PEF is a closely related but different
metric of lung function. PEF measurements have been shown to be more variable than FEVj in
some studies (Vaughan et al., 1989; Cross and Nelson, 1991) and can have an element of
uncertain reliability when self-administered by study subjects. However, Lippmann and Spektor
(1998) state that PEF measurements from small inexpensive flow meters, which are more
feasible to use in field studies, can produce similar results to PEF measured spirometrically.
Studies of FEVj are assessed here first, followed by a discussion of PEF studies. Other
dividing aspects within these two major types of lung function studies include health status of
subjects (e.g., healthy, mildly asthmatic, severely asthmatic), age group, time spent outdoors,
and exertion levels. Several studies brought these factors together to produce informative data.
Some FEVj studies involved both increased outdoor O3 exposure and higher exertion levels.
The results from this group of subjects are relatively comparable to those seen with exercising
subjects in the clinical studies discussed in Chapter 6.
7.2.3.1 Spirometry (FEVt) Studies in Outdoor Worker, Exercise, Children, Elderly, and
Asthmatic Panels
Studies published over the past decade have provided some new insights on the acute
effects of O3 on FEVj. The results of all studies that investigated quantitative O3-related effects
on FEVj are summarized in the following tables. Tables 7-la,b,c present changes in FEVj
associated with O3 exposure in adults, whereas Tables 7-2a,b,c present effects in children.
Borsboom et al. (1999) examined the circadian variation in spirometric parameters and observed
that FEVl3 FVC, and PEF, in general, increased from early morning until around noon, then
decreased afterwards. Average variations in FEVl3 FVC, and PEF were 2.8%, 4.8%, and 3.1%,
respectively. To take into account the circadian variation, several studies stratified their analyses
by time of day, examining the effect of O3 on lung function separately for morning and afternoon
measurements. Results from these studies are shown in Tables 7-lb and 7-2b. In other
studies, O3 exposure was related to the cross-day change in spirometric measurements (i.e., the
difference between same day morning and afternoon measurements). Results from these studies
are presented in Tables 7-lc and 7-2c. Studies that did not provide quantitative O3 data (Cuijpers
et al., 1994; Delfino et al., 2004; Frischer et al., 1997) are not included in the tables. The data
presented in Hoppe et al. (1995a) were further analyzed in a subsequent paper (Hoppe et al.,
2003); results from the latter paper are included in the tables. In general, the O3 effect estimates
7-29
-------
Table 7-la. Field Studies that Investigated the Association Between Acute Ambient O3
Exposure and Changes in FEVt in Adults
Reference
Korricketal. (1998)
Braueretal. (1996)
Schindleretal. (2001)
Hoppe et al. (2003)
Romieuetal. (1998)
Study Location
Mount Washington, NH
Fraser Valley, British
Columbia, Canada
Eight communities in
Switzerland
Munich, Germany
Mexico City
Study Period
Summers 1991, 1992
Jun-Aug 1993
May-Sep 1991
Apr-Sep 1992-1995
Mar-May 1996;
Jun-Aug 1996
Mean O3 (SD)
Level, ppb
40 (12)
40.3 (15.2)
46.6
(1.5-127.6)b
65.9- 70.4 c
123 (40)
O3 Index
8-h avg a
1-hmax
8-h avg
!/2-h max
1-hmax
a Average of the hourly O3 concentrations during each hike. Hikes averaged 8 hours in duration.
b Range of 8-h avg concentrations is presented by Schindler et al. (2001).
0 Range of mean i/2-h max O3 concentrations on high O3 days is presented for Hoppe et al. (2003).
showed decrements for FEVj across studies, especially in children. The studies presented in the
tables are discussed in further detail below, starting with O3 effects on individuals with elevated
exertion levels and increased exposure due to time spent outdoors, followed by O3 effects on
persons in other potential risk groups.
Outdoor worker and exercise panels
A very important part of the basis for the current 8-h NAAQS for O3 was the results from
controlled human exposure studies, as discussed in Chapter 6. These field studies with subjects
at elevated exertion levels are of particular interest due to their similarities to the human
chamber studies. The majority of human chamber studies have examined the effects of O3
exposure in subjects performing continuous or intermittent exercise for variable periods of time
(see Chapter 6 of this document).
A study by Brauer et al. (1996) reported unusually large O3 effects on lung function among
outdoor workers. This study presented O3 effects observed during an extended outdoor exposure
period combined with elevated levels of exertion. The investigators repeatedly measured
spirometric lung function before and after outdoor summer work shifts over 59 days among a
7-30
-------
Table 7-lb. Percent Changes in FEVt (95% CI) Associated with Acute Ambient O3
Exposures in Adults, Ordered by Size of the Estimate a
Reference
Study Population/Analysis
N
% Change in FEVj
1 Braueretal. (1996)b
2 Braueretal. (1996)b
3 Romieuetal. (1998)c
4 Schindleretal. (2001)
5 Romieuetal. (1998)c
6 Hoppe et al. (2003)b
7 Romieuetal. (1998)c
8 Romieuetal. (1998)c
9 Romieuetal. (1998)c
10 Romieuetal. (1998)c
11 Hoppe et al. (2003)
12 Hoppe et al. (2003)
13 Romieuetal. (1998)c
14 Hoppe et al. (2003)b
15 Hoppe et al. (2003)b
16 Hoppe et al. (2003)b
17 Hoppe et al. (2003)b
18 Hoppe et al. (2003)
19 Romieuetal. (1998)c
20 Hoppe et al. (2003)
21 Hoppe et al. (2003)
22 Hoppe et al. (2003)
23 Hoppe et al. (2003)
Berry pickers, next morning 58
Berry pickers, afternoon 58
Street workers on placebo, 1st phase (lag 0-1) 19
Adults who never smoked (lag 0) 3,912
Street workers on placebo, 1st phase (lag 0) 19
Athletes, afternoon (lag 0) 43
Street workers on supplement, 1st phase 22
(lag 0-1)
Street workers on supplement, 1st phase (lag 0) 22
Street workers on placebo, 2nd phase (lag 0) 23
Street workers on placebo, 2nd phase (lag 0-1) 23
Elderly, morning (lag 2) 41
Athletes, morning (lag 0) 43
Street workers on supplement, 2nd phase (lag 0) 19
Athletes, afternoon (lag 2) 43
Athletes, afternoon (lag 1) 43
Athletes, morning (lag 2) 43
Athletes, morning (lag 1) 43
Elderly, afternoon (lag 0) 41
Street workers on supplement, 2nd phase 19
(lag 0-1)
Elderly, afternoon (lag 1) 41
Elderly, morning (lag 0) 41
Elderly, morning (lag 1) 41
Elderly, afternoon (lag 2) 41
-6.36 (-8.02,-4.70)
-5.40 (-6.51,-4.28)
-3.55 (-6.28,-0.82)
-2.96 (-5.11,-0.76)
-2.17 (-3.45,-0.89)
-1.26 (-2.63, 0.10)
-1.25 (-4.36, 1.86)
-0.53 (-
0.40 (-
-0.36 (-
-0.22(-
0.01 (-
0.18 (-
0.24 (-
0.48 (-
0.62 (-
0.71 (-
0.75 (-
0.82 (-
2.08, 1.01)
1.94,1.14)
2.93,2.20)
3.86, 3.42)
0.10,0.09)
0.72, 1.08)
0.64,1.12)
0.97, 1.94)
0.45, 1.68)
0.65, 2.07)
2.08, 3.58)
0.77, 2.42)
1.16 (-1.26, 3.58)
1.68 (-3.72, 7.07)
1.82 (-2.19, 5.84)
2.88 (-0.24, 6.00)
aChange in FEVl is per standard unit ppb O3 (40 ppb for i/2-h max O3 and 1-h max O3, 30 ppb for 8-h max O3,
and 20 ppb for 24-hr avg O3).
bBrauer et al. (1996) and Hoppe et al. (2003) studies also included children. The study population for
Brauer et-al. ranged in age from 10 to 69 years (mean age 44 years). For Hoppe et al. (2003), the athletes
ranged in age from 13 to 38 years (mean age 18 years).
GRomieu et al. (1998) present change in FEVl (mL). The data from Romieu et al. (1998) were transformed
to percent change by dividing the estimates by 3,300 mL (average FEVj for 40 year old Mexican-American
males by Hankinson et al., 1999).
7-31
-------
Table 7-lc. Cross-day Percent Changes in FEVt (95% CI) Associated with Acute
Ambient O3 Exposures in Adults, Ordered by Size of the Estimate
Reference
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
Korrick et
al
al
al
al
al
al
al
al
al
al
al
al
al
al
Brauer et al.
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
. (1998)
(1996) b
Study Population/Analysis
Hikers
Hikers
Hikers
Hikers
Hikers
Hikers
with wheeze or asthma (post-pre-hike)
who hiked 8-12 hours (post-pre-hike)
age 28-37 years (post-pre-hike)
who never smoked (post-pre-hike)
male (post-pre-hike)
age 38-47 years (post-pre-hike)
All hikers (post-pre-hike)
All hikers, with PM25 and acidity in model
(post-pre-hike)
Hikers
Hikers
Hikers
age 18-27 years (post-pre-hike)
female (post-pre-hike)
age 48-64 years (post-pre-hike)
Hikers without wheeze or asthma
(post-pre-hike)
Hikers
Hikers
who hiked 2-8 hours (post-pre-hike)
who formerly smoked (post-pre-hike)
Berry pickers (post-pre-work shift)
N
40
265
185
405
375
142
530
530
135
155
68
490
265
125
58
Cross-day % Change
in FEVj
-4
-2.
-2
-1
-1.
-1
-1.
-1.
-1
-1.
-1.
-1
-0
-0
0
.47
.07
.01
.77
.65
.59
.53
.44
.29
.17
.14
.08
.99
.72
.00
(-7.
(-3.
(-3.
(-3.
(-3.
(-3.
(-2.
(-3.
(-2.
(-3.
(-3.
(-2.
(-2.
(-3.
(-1
.65,
.78,
.42,
.24,
.12,
.12,
.82,
.32,
.88,
.46,
.08,
.49,
.70,
.07,
.66,
-1.29)
-0.36)
-0.60)
-0.30)
-0.18)
-0.06)
-0.24)
0.44)
0.30)
1.12)
0.80)
0.33)
0.72)
1.63)
1.66)
"Change in FEVj is per standard unit ppb O3 (40 ppb for i/2-h max O3 and 1-h max O3, 30 ppb for 8-h max O3,
and 20 ppb for 24-h avg O3).
bBrauer et al. (1996) study also included children. The study population ranged in age from 10 to 69 years
(mean age 44 years).
group of 58 berry pickers in Fraser Valley, British Columbia, Canada. The subjects, both male
and female native Punjabi-speakers, ranged in age from 10 to 69 years old, with a mean age of
44 years. Outdoor work shifts averaged 11 hours in duration. The mean 1-h max O3
concentration was 40.3 ppb (SD 15.2). Exertion levels were estimated using portable heart rate
monitors carried over a period of four or more hours by a representative subset of subjects
during 16 work shifts. Heart rates over the work shift averaged 36% higher than resting levels.
Post-shift FEVj and FVC decreased as a function of O3 concentration, and the effects of O3
7-32
-------
Table 7-2a. Field Studies that Investigated the Association Between Acute Ambient O3
Exposure and Changes in FEVt in Children
Reference
Linnetal. (1996)
Scarlett etal. (1996)
Hoppe et al. (2003)
Ulmer etal. (1997)
Castillejos etal. (1995)
Romieu et al. (2002)
Chen etal. (1999)
Study Location
Rubidoux, Upland,
and Torrance, CA
Surrey, England
Munich, Germany
Freudenstadt and
Villingen, Germany
SW Mexico City
Mexico City
Sanchun, Taihsi, and
Linyuan, Taiwan
Study Period
Fall-spring
1992-1993,
1993-1994
Jun-Jul 1994
Apr-Sep 1992-1995
Mar-Oct 1994
Aug 1990-Oct 1991
Octl998-Apr2000
May 1995-Jan 1996
Mean O3 (SD)
Level, ppb
23 (12)
50.7 (24.48)
65.9- 70.4 a
Freudenstadt:
50.6 (22.5-89.7) b
Villingen:
32.1 (0.5-70.1)b
112.3 (0-365) c
102 (47)
19.7-110.3°
O3 Index
24-h avg
8-hmax
Vi-h max
Vi-h max
i/2-h max
1-h max
1-hmax
1-h max
aRange of mean i/2-h max O3 concentrations on high O3 days is presented for Hoppe et al. (2003).
bMedian and 90th percentile interval are presented for Ulmer et al. (1997).
°Range of 1-h max O3 concentrations are presented by Castillejos et al. (1995) and Chen et al. (1999).
remained significant after adjusting for PM2 5 in the analysis (see Table 7-lb). Declines in
lung function also were observed on the morning following high O3 exposure. Ozone was not
significantly associated with a cross-day change in lung function (i.e., difference between
afternoon and morning spirometry measurements) (see Table 7-lc). One explanation for the
lack of an association may be that each spirometric measurement incorporated the impact of the
previous day's O3 in addition to the O3 effect on that particular day. The effects seen in this
study are larger than have been reported previously in studies with briefer exposure durations.
For example, a change of -3.8 mL (95% CI: -4.6, -3.0) in afternoon FEVj was shown per
1 ppb increase in O3 concentrations, compared to the decline of 0.4 mL/ppb and 1.35 mL/ppb
observed in the earlier adult exercise studies (Spektor et al., 1988b; Selwyn et al., 1985). These
results are consistent with the interpretation that extended exposures to O3 produce more marked
effects on lung function. Further, when data were restricted to days with 1-h max O3
concentrations under 40 ppb, the O3 effects on afternoon FEVj did not change in magnitude and
remained significant. However, a possible role of copollutants cannot be completely excluded.
7-33
-------
Table 7-2b. Percent Changes in FEVt (95% CI) Associated with Acute Ambient O3
Exposures in Children, Ordered by Size of the Estimate a
Reference
Study Population/Analysis
N % Change in FEVj
1 Ulmeretal. (1997)b
2 Ulmeretal. (1997)b
3 Ulmeretal. (1997)b
4 Ulmeretal. (1997)b
5 Hoppe et al. (2003)c
6 Chen etal. (1999)
7 Chen etal. (1999)
8 Hoppe et al. (2003)
9 Romieu et al. (2002)b
10 Romieu et al. (2002)b
11 Chen etal. (1999)
12 Ulmeretal. (1997)b
13 Chen etal. (1999)
14 Hoppe et al. (2003)c
15 Linn et al. (1996)b
16 Linn et al. (1996)b
17 Romieu et al. (2002)b
18 Hoppe et al. (2003)
19 Hoppe et al. (2003)c
20 Romieu et al. (2002)b
21 Romieu et al. (2002)b
22 Scarlett etal. (1996)d
23 Romieu et al. (2002)b
24 Hoppe et al. (2003)c
School children in Freudenstadt (lag 1)
School boys in Freudenstadt and Villingen
(lag 1)
School children in Freudenstadt and Villingen
(lag 1)
School girls in Freudenstadt and Villingen
(lag 1)
Asthmatics, afternoon (lag 2)
Children, with NO2 in model (lag 1)
Children (lag 1)
Children, morning (lag 0)
Moderate to severe asthmatic children on
placebo (lag 1)
Moderate to severe asthmatic children on
placebo, with NO2 and PM10 in model (lag 1)
Children (lag 2)
School children in Villingen (lag 1)
Children (lag 7)
Asthmatics, afternoon (lag 1)
School children, next morning
School children, afternoon
All asthmatic children on placebo (lag 1)
Children, afternoon (lag 0)
Asthmatics, afternoon (lag 0)
Moderate to severe asthmatic on supplement
(lag 1)
Moderate to severe asthmatic on supplement,
with NO2 and PM10 in model (lag 1)
School children (lag 1)
All asthmatic children on supplement (lag 1)
Asthmatics, morning (lag 1)
57 -4.60 (-7.54,-1.67)
67 -3.23 (-6.47, 0.00)
135 -2.98 (-5.33,-0.63)
68 -2.32 (-5.53, 0.88)
43 -2.08 (-6.24, 2.08)
895 -1.97 (-3.51,-0.43)
895 -1.48 (-2.84,-0.12)
44 -1.45 (-4.27, 1.38)
35 -0.99 (-1.80,-0.18)
35 -0.97 (-1.87,-0.07)
895 -0.93 (-2.56, 0.71)
78 -0.79 (-3.93, 2.34)
895 -0.72 (-1.81, 0.37)
43 -0.56 (-4.61, 3.50)
269 -0.27 (-0.79, 0.24)
269 -0.19 (-0.73, 0.35)
78 -0.19 (-0.71, 0.33)
44 -0.14 (-2.71, 2.42)
43 -0.10 (-6.59, 6.39)
47 -0.04 (-0.80, 0.72)
47 -0.01 (-0.82, 0.80)
154 0.01 (-0.20, 0.22)
80 0.04 (-0.52, 0.60)
43 0.30 (-3.93, 4.53)
7-34
-------
Table 7-2b (cont'd). Percent Changes in FEVt (95% CI) Associated with Acute Ambient
O3 Exposures in Children, Ordered by Size of the Estimate a
Reference Study Population/Analysis N % Change in FEVj
25 Hoppe et al. (2003) Children, morning (lag 1) 44 0.83 (-0.53,2.20)
26 Hoppe et al. (2003) Children, afternoon (lag 1) 44 0.93 (-0.80,2.66)
27 Hoppe et al. (2003) Children, morning (lag 2) 44 1.17 (-0.36,2.70)
28 Hoppe et al. (2003) Children, afternoon (lag 2) 44 1.20 (-0.12,2.52)
29 Hoppe et al. (2003)c Asthmatics, morning (lag 2) 43 1.40 (-3.69,6.49)
30 Hoppe et al. (2003)c Asthmatics, morning (lag 0) 43 3.41 (-2.50,9.33)
"Change in FEV{ is per standard unit ppb O3 (40 ppb for !/2-h max O3 and 1-h max O3, 30 ppb for 8-h max O3,
and 20 ppb for 24-h avg O3).
bLinn et al. (1996), Romieu et al. (2002), and Ulmer et al. (1997) present change in FEV; (mL). The data were
transformed to percent change by dividing the estimates by 1,900 mL (average FEVl among 8 to 10 year olds
by Hankinson et al., 1999).
°H6ppe et al. (2003) study also included young adults. The study population age for the asthmatics ranged
from 12 to 23 years (mean age 15 years).
dFEV0 75 results are presented in Scarlett et al. (1996).
Table 7-2c. Cross-day Percent Changes in FEVt (95% CI) Associated with Acute
Ambient O3 Exposures in Children, Ordered by Size of the Estimate a
Cross-day % Change
Reference Study Population/Analysis N in FEVj
1 Linn et al. (1996)b School children (p.m.-a.m.) 269 -0.61 (-1.09,-0.14)
2 Castillejos et al. (1995) Private primary school (post-pre-exercise) 40 -0.48 (-0.72,-0.24)
"Change in FEV{ is per standard unit ppb O3 (40 ppb for i/2-h max O3 and 1-h max O3, 30 ppb for 8-h max O3,
and 20 ppb for 24-h avg O3).
bLinn et al. (1996) present change in FEV{ (mL). The data were transformed to percent change by dividing the
estimates by 1,900 mL (average FEV{ among 8 to 10 year olds by Hankinson et al., 1999).
In a Mexico City study of 47 outdoor street workers (Romieu et al., 1998), spirometry was
performed repeatedly at the end of the work shift over a 2-month period. Subjects were exposed
to outdoor ambient O3 for a mean of 7.4 h during the workday. The mean 1-h max O3
concentration during the study period was 123 ppb (SD 40). Among those who had never taken
an antioxidant supplement (subjects who received a placebo during the first phase of the study),
7-35
-------
same day O3 concentrations were associated with decreases in FEVj. A mean change of
-71.6 mL (95% CI: -113.9, -29.3) (approximately a 4% decline) was observed per 40 ppb
increase in 1-h max O3. The results from this study, in addition to those from the Canadian study
of berry pickers (Brauer et al., 1996), indicate that outdoor workers are a potentially vulnerable
population that may need protection from O3 exposures.
Hoppe et al. (1995a) examined forestry workers (n = 41) for changes in pulmonary
function attributable to O3 exposure in Munich, Germany. In addition, athletes (n = 43) were
monitored in the afternoon following a 2-h outdoor training period. Pulmonary function tests
were conducted on days of both "high" (mean lA>-h max O3 of 64 to 74 ppb) and "low" (mean
l/2-h max O3 of 32 to 34 ppb) ambient O3 concentrations. Ventilation rates were estimated from
the average activity levels. Athletes, who had a fairly high ventilation rate of 80 L/min,
experienced a significant decrease of 60.8 mL (95% CI: 6.4, 115.2) in FEVj per 40 ppb increase
in mean l/2-h max O3. Among the forestry workers, a similar O3-related decline in FEVj also was
observed (-56.0 mL [95% CI: -118.4, 6.4]). In a subsequent study, Hoppe et al. (2003)
reanalyzed the results of the athletes after stratifying the spirometric data by time of day
(morning versus afternoon) and at different lag periods (lags of 0 to 2 days). The reanalysis
indicated that O3-related decrements were observed only with the afternoon FEVj at a 0-day lag,
-1.26% (95% CI: -2.63, 0.10) change in FEVj per 40 ppb increase in mean V2-h max O3. The
mean lA>-h max O3 in the afternoon (1 p.m. to 4 p.m.) was 65.9 ppb (range 51 to 86) on O3 days.
One FEVj study clearly demonstrated small but measurable effects of multihour O3
exposures on adults exercising outdoors. In Korrick et al. (1998), adult hikers (n = 530) of
Mount Washington, NH performed spirometry before and after hiking for a mean of 8 hours
(range 2 to 12). The mean hourly O3 concentration during each hike ranged from 21 to 74 ppb.
After the hike, all subjects combined experienced a small mean decline of 1.5% (95% CI: 0.2,
2.8) in FEVi and 1.3% (95% CI: 0.5, 2.1) in FVC per 30 ppb increase in the mean of the hourly
O3 concentration during the hike. In addition, Korrick et al. (1998) compared hikers who hiked 8
to 12 hours to those who hiked 2 to 8 hours. Among those who hiked longer, the percent change
in FEVj was more than 2-fold greater per ppb exposure compared to those who hiked only for 2
to 8 hours. Each hour hiked, which may reflect dose, was associated with a decline of 0.3%
(p = 0.05) in FEVj, after adjusting for O3.
7-36
-------
In a Mexico City study, the O3 effect attributable to exercise was determined using a group
of school children (n = 40) who were chronically exposed to moderate to high levels of O3
(Castillejos et al., 1995). Children were tested up to 8 times between August 1990 and October
1991. Spirometry was performed by the children before and after a 1-h intermittent exercise
session outdoors. Outdoor O3 levels ranged up to 365 ppb, with a mean of 112.3 ppb. Linear
trend analyses indicated a relationship between quintiles of O3 and percent change in lung
function. However, stratified analyses indicated that significant changes were observed only
with higher quintiles of O3 exposure (72 to 125 ppb and 183 to 365 ppb). Therefore, children
exercising at higher O3 levels experienced declines in pulmonary function despite the repeated
daily exposure to moderate and high levels of O3 in Mexico City.
Collectively, the above studies confirm and extend clinical observations that prolonged
exposure periods, combined with elevated levels of exertion or exercise, may magnify the effect
of O3 on lung function. The most representative data come from the Korrick et al. (1998) hiker
study. This U.S. study provided outcome measures stratified by several factors (e.g., gender,
age, smoking status, presence of asthma) within a population capable of more than normal
exertion.
Panel studies of children, elderly, and asthmatics
Hoppe et al. (1995a,b) examined several potentially susceptible populations for changes in
pulmonary function attributable to O3 exposure in Munich, Germany. The forestry workers and
athletes were discussed in the previous section. Senior citizens (n = 41) and juvenile asthmatics
(n = 43) were also monitored on "low" O3 (mean %-h max O3 of 32 to 34 ppb) and "high" O3
(mean !/2-h max O3 of 64 to 74 ppb) days. Subjects were requested to stay outdoors for at least
2 hours just before the afternoon pulmonary function test. Clerks (n = 40) were considered the
nonrisk control group. Although clerks spent the majority of their time indoors, their outdoor
exposures on high O3 days were similar to that of the four other risk groups. The results showed
no significant O3 effects on the senior citizens. Clinical studies also have consistently shown
that seniors are less responsive to O3 (Bedi et al., 1989; Drechsler-Parks, 1995). Asthmatics and
clerks experienced slight reductions in FEVj on high O3 days. Among all risk groups, juvenile
asthmatics experienced the largest O3-related decline in FEVj, -84.0 mL (95% CI: -196.4,
28.4) per 40 ppb increase in mean %-h max O3. To further examine their hypotheses on
7-37
-------
characteristics of O3 risk groups, Hoppe et al. (2003) conducted a different analysis on a more
expanded data base than utilized in the earlier study. Children were examined as an additional
risk group. Mean lA>-h max O3 ranged from 65.9 ppb to 70.4 ppb on O3 days for each risk group.
Hoppe et al. (2003) presented both group mean values and analyses on an individual basis. For
the group mean values, consistent O3 effects were not detectable. On an individual basis, a
potential pattern of O3 sensitivity was observed (see Table AX7-1 in Annex Section AX7.1 for
details). About 20% of the children and asthmatics were regarded as O3 responders (i.e.,
individuals that had greater than 10% change in FEVj) compared to only 5% of the elderly and
athletes. These results indicated that while the majority of the population did not react to O3
exposure, a small group of susceptible individuals experienced health effects from O3. The
sample size limits quantitative extrapolation to larger populations, but may allow cautious first
estimates.
Several other panel studies performed spirometry in children, another potentially
susceptible group (Avol et al., 1998; Chen et al., 1999; Cuijpers et al., 1994; Frischer et al.,
1997; Linn et al., 1996; Romieu et al., 2002; Scarlett et al., 1996; Ulmer et al., 1997).
All studies, with the exception of Avol et al. (1998) and Scarlett et al. (1996), observed a
decrease in FEVj associated with O3 exposure. In a cohort of 154 children in Surrey, England,
Scarlett et al. (1996) observed no association between ambient O3 concentrations and FEV075
(0.2 mL [95% CI: -3.6, 3.9] increase per 30 ppb increase in 8-h max O3 at a 1-day lag), but
noted a small effect of PM10 on lung function. The mean 8-h max O3 concentration was 50.7 ppb
(range 6.8 to 128 ppb). The study by Avol et al. (1998) examined three groups of children,
asthmatic (n = 53), wheezy (n = 54), and healthy (n = 103). Ozone levels were reported as being
very low (values not provided). The authors advised that noncompliance by the subjects might
have been a problem, and further noted limitations in the analysis methods and other aspects of
the study design.
One large study measured spirometric lung function in 895 school children in three towns
in Taiwan (Chen et al., 1999). The 1-h max O3 concentrations ranged from 19.7 to 110.3 ppb.
Lung function was measured only once for each subject. The authors reported significant
associations between diminished FEVj and FVC with a 1-day lag of O3 concentrations. Effect
sizes were typical of those observed in past studies, i.e., 0.5 to 1.0 mL decline in FEVj per ppb
7-38
-------
increase in O3 concentration. Ozone was the only air pollutant associated with changes in lung
function in multipollutant models including SO2, CO, PM10, and NO2.
Linn et al. (1996) repeatedly measured spirometric lung function among 269 school
children in three southern California communities (Rubidoux, Upland, and Torrance). Lung
function was measured over 5 consecutive days, once in each of 3 seasons over 2 school years.
Between-week variability was controlled in the analysis by seasonal terms in the model.
Statistical power was limited by the relatively narrow range of exposures that were experienced
within each week. In addition, the study was restricted to the school year, eliminating most of
the "high" O3 season from consideration. During the study period, 24-h avg O3 levels at the
central monitoring site ranged up to 53 ppb (mean 23 ppb), whereas personal measurements
ranged up to 16 ppb (mean 5 ppb). A mean change of -11.6 mL (95% CI: -20.6, -2.6)
(approximately a 1% decline) in FEVj was observed from morning to afternoon per 20 ppb
increase in 24-h avg O3. Other associations (involving individual morning or afternoon FVC
and FEVj measurements) went in the plausible direction, but the O3 effect estimates were
much smaller.
Ulmer et al. (1997) examined 135 children aged 8 to 11 years in two towns in Germany
from March to October 1994 for O3 effects on pulmonary function at four time periods. The
cross-sectional results at each of the four time points showed limited FVC and no FEVj
associations. However, the longitudinal analysis, which combined data from all four periods
yielded a mean change of -87.5 mL (95% CI: -143.2, -31.7) (approximately a 5% decline)
in FEVj per 40 ppb increase in l/2-h max O3 for the town with the higher O3 levels (median Va-h
max of 50.6 ppb versus 32.1 ppb). In the cross-sectional analysis, only between-person
variability was analyzed. The longitudinal analysis, in which the subjects provided multiple
days of measurements, provided information on both between- and within-subject responses.
There are a limited number of new epidemiologic studies examining the effects of O3
on FEVj; however, results from these studies indicate that acute exposure to O3 is associated
with declines in FEVj in children. These results further support the negative effects of O3 on
lung function observed in the meta-analysis on children attending summer camp (Kinney et al.,
1996a) and in the clinical literature.
7-39
-------
7.2.3.2 Peak Flow Meter (PEF) Studies in Asthmatics and Healthy Individuals
Many studies of the acute effect of O3 on PEF examined self-administered PEF levels
daily, both in the morning and afternoon. PEF follows a circadian rhythm, with the highest
values found during the afternoon and lowest values during the night and early morning
(Borsboom et al., 1999). Due to the diurnal variation in PEF, most studies analyzed their data
after stratifying by time of day. The peak flow studies examined both asthmatic panels and
healthy individuals. The asthma panels are discussed first.
Asthma panels
The effects of acute O3 exposure on PEF in asthmatics were examined in several panel
studies. Figures 7-1 and 7-2 present percent changes in morning and afternoon PEF outcomes
from seven panel studies of children, mostly asthmatic, ranging in age from 5 to 13 years. The
effect estimates from all single-day and multiday lag models are presented. Only single-city
results with analyses stratified by morning and afternoon are included in the figure. Studies that
examined cross-day changes and daily variability in PEF (e.g., Just et al., 2002; Thurston et al.,
1997) are not included in the figure since such outcomes are not directly comparable.
Collectively, nearly all of the studies indicated decrements of peak flow but most of the
individual estimates were not statistically significant. The results from the individual studies are
further discussed below.
In Mexico City, two studies of asthmatic school children were carried out simultaneously
in the northern (Romieu et al., 1996) and southwestern sections of the city (Romieu et al., 1997).
In the northern study, 71 mildly asthmatic school children aged 5 to 13 years old, were followed
over time for daily morning (before breakfast) and afternoon (bedtime) PEF. The mean 1-h
max O3 was 190 ppb (SD 80). In single-pollutant models, O3 concentrations at 0-, 1-, and 2-day
lags were associated with diminished morning and afternoon PEF, but only the 0-day lag
morning effect was significant. The O3 effect became nonsignificant when PM2 5 was added to
the model. In the southwestern study, 65 mildly asthmatic children aged 5 to 13 years old were
followed during the summer and winter for daily morning and afternoon PEF. The mean 1-h
max O3 was 196 ppb (SD 78). Ozone concentrations at a 0- and 1-day lag were associated with
afternoon PEF, with larger effects at a 1-day lag. Associations involving O3 were stronger than
7-40
-------
HI
o.
a)
O)
c
re
£
O
1
3 -
2 -
1 -
0 -
1 -
2 -
3 -
4 -
5 -
-
7 -
Q
(Single-day Lags
|
1 • ^
c a
to O
§ o
fc .a
/R v
1
I Jc
T
Asthmatic
Yes O
No X
lO O
CD -*r
u ii
s. s.
^ >>
o o
o o
_o ^
'•^ "•*'
1 Multiday Lags |
<
'
) (
5 (
(
)
°
' C) 0
;;
cocotD-*CMT--<3-co cr
? 0
T™cotocoo>^— o^r LO ^"
T- ,- || || n T- T- T- .<- ||
1 " Z Z z " " " " Z
2 Z " — ' 2 Z "Z. 2 ' "
• ' ^ ^ — ,j ' • - ' • ^_s -^
O=!ojQ|^> •§•_ §
Q" ^- ._ ^ Q
£/)" D) c v» O QJ F ^ ,£S ,9
•= (0 ffi /1% ±±. 1— rti ^ f™ -V-
0)
T3
J?
IE
Q.
CO
cu
O
o
o
CM
CD
0)
Q.
Q.
Q
C
CO
IB
E
o
IT
o
O
CO
w
'S
CD
O
o
"5
m
0)
T3
ra
CO
LU
0)
O5
CT>
o
O
J)
CD
b
Figure 7-1. Percent change (95% CI) in morning PEF in children per standardized
increment (see Section 7.1.3.2). For single-day lag models, previous day O3
effects are shown. For multiday lag models, the cumulative effects of a 1- to
5-day lag are shown for Mortimer et al. (2002) and Neas et al. (1999), and the
effect of a 1- to 10-day lag is shown for Gold et al. (1999).
those involving PM10. Several additional studies, both in the United States and in other
countries, reported significant associations between O3 exposure and decrements in PEF among
asthmatics (Gielen et al., 1997; Jalaludin et al., 2000; Just et al., 2002; Ross et al., 2002;
Thurston et al., 1997).
7-41
-------
y_
m
a.
O)
c
n
x:
O
t
3 -
2 -
1 -
1 -
2 -
3 -
4 -
5 -
fi -
7 -
(Single-day Lags)
-
I
-It
! ,
^ t t
Asthmatic
Yes O
No X
|Multiday Lags)
' *
05
oo
II
z
O
"c
IO
CO
CO
CO
to
(O
II
z
•q
II
z.
c
a
r^
ii
E
CO
"S
IB
0)
a
0,
D.
(O
en
o
'E
o
CC
83
II
Z.
O
CO
0)
D
ffl
"E
o
U.
o
II
.
o
*
2
o
O
a.
o>
0)
o
0
ffi
b
Figure 7-2. Percent change (95% CI) in afternoon PEF in children per standardized
increment (see Section 7.1.3.2). For single-day lag models, current day O3
effects are shown. For multiday lag models, the cumulative effect of a 1- to
5-day lag is shown for Neas et al. (1999) and a 1- to 9-day lag is shown for
Gold et al. (1999).
Other epidemiologic studies did not find a significant O3 effect on the lung function of
asthmatics. Delfino et al. (1997a) examined morning and evening PEF among 22 asthmatics
ranging in age from 9 to 46 years, living in Alpine, CA. Daily ambient 12-h avg O3 (8 a.m. to
8 p.m.) concentrations ranged from 34 to 103 ppb, with a mean value of 64 ppb. Unique to this
study, personal O3 exposures were measured using 12-h passive O3 samplers that were worn by
7-42
-------
the subjects. The personal 12-h avg O3 (8 a.m.-8 p.m.) concentrations, which had a mean value
of 18 ppb, were much lower than the fixed-site ambient levels. Quantitative O3 results were not
reported, but the researchers stated that no O3 effects were observed on morning and evening
PEF. In Hiltermann et al. (1998), 60 nonsmoking adults aged 18 to 55 years in Bilthoven, the
Netherlands, were followed between July and October 1995 with morning and afternoon PEF
measurements. Although negative associations were observed between O3 and cross-day
changes in PEF, the results were not significant. The mean ambient 8-h max O3 was 41.3 ppb
(range 3 to 49).
Mortimer et al. (2002) examined 846 asthmatic children from the National Cooperative
Inner-City Asthma Study (NCICAS) for O3-related changes in PEF. Children from eight U. S.
urban areas (St. Louis, MO; Chicago, IL; Detroit, MI; Cleveland, OH; Washington, DC;
Baltimore, MD; East Harlem, NY; and Bronx NY) were monitored from June through August
1993. Median 8-h avg O3 (10 a.m.-6 p.m.) concentrations ranged from 34 ppb in Chicago to
58 ppb in Washington, DC. The mean 8-h avg O3 level across the eight cities was 48 ppb. This
study provides representative data for the United States, in so much as children from multiple
cities throughout the East and Midwest were examined. Asthmatic children from urban areas are
an important subgroup of potentially at-risk populations. Study children either had physician-
diagnosed asthma and symptoms in the past 12 months or respiratory symptoms consistent with
asthma that lasted more than 6 weeks during the previous year.
Mortimer et al. (2002) examined O3-related changes in PEF for single-day lags from 1 to
6 days and a multiday lag period of 5 days. Of all the pollutants examined, including O3, PM10,
NO2, and SO2, none were associated with evening PEF. Only O3 was found to be associated
with morning PEF. The effect estimates of the association between O3 and morning PEF for the
single-day and multiday lags are depicted as error density curves in Figure 7-3 (for description of
error density curves, see Annex Section AX7.2). Small morning effects were observed at 1- and
2-day lags. The effect of O3 on morning outcomes increased over several days. A strong
association between O3 and PEF also was found with a multiday lag period (cumulative lag of
1 to 5 days). Unrestricted lag models suggested that the O3 exposure from 3 to 5 days prior had a
greater impact on morning % PEF than more immediate exposures. Mortimer et al. discussed
biological mechanisms for delayed effects on pulmonary function, which included increased
bronchial reactivity secondary to airway inflammation associated with irritant exposure. Animal
7-43
-------
-3 -2.5 -2 -1.5 -1 -0.5
1.5
0)
Q
2.0
1.5
1.0
0.5
0.0 E-
1 - to 5-day
Multiday Lag
-3 -2.5 -2 -1.5 -1 -0.5 0
% Change in PEF
0.5
1.5
Figure 7-3. Percent changes in PEF per 30 ppb increase in 8-h avg O3 in urban children.
Single-day lags (1-, 2-, 3-, 4-, 5-, and 6-day) are shown in the upper panel.
The cumulative multiday lag (1- to 5-day) is shown in the lower panel.
Source: Derived from Mortimer et al. (2002).
toxicology and human chamber studies (see Chapters 5 and 6) provide further evidence that
exposure to O3 may augment cellular infiltration and cellular activation, enhance release of
cytotoxic inflammatory mediators, and alter membrane permeability.
Figure 7-4 illustrates the probability density curves of the results from the individual-cities
analysis and that from the pooled analysis of all eight cities. The error density curve for the
all-cities analysis is a graphical presentation of the all-cities regression analysis presented by
Mortimer et al. (2002), a change in morning PEF of-1.18% (95% CI: -2.10, -0.26) per 30 ppb
7-44
-------
1
0.8
.^ 0.6
(A
Q °'4
0.2
0
- - Individual-Cities Analysis
— All-Cities Analysis
-4
-2 -1
% Change in PEF
Figure 7-4. Percent change in PEF per 30 ppb increase in 8-h avg O3 with a cumulative lag
of 1 to 5 days. The error density curve is shown for the pooled analysis of the
eight NCICAS cities and the summary density curve is presented for the
analyses from the individual cities. Note that 99% and 78% of the areas
under the curves are less than zero for the pooled-cities analysis and
individual-cities analysis, respectively.
Source: Derived from Mortimer et al. (2002).
increase in 8-h avg O3 with a cumulative lag of 1 to 5 days. The summary density curve for the
individual-cities analysis was calculated by summing together eight normal distribution
functions, one for each of the study cities, then taking the derivative of the summed function (see
Annex Section AX7.2 for further explanation of summary density curves). The area under the
density curve and to the left of a value on the x-axis is an estimate of the probability that the
effect estimate will be less than or equal to that value. For example, the area under the density
curve to the left of 0% change in PEF is 99% in the all-cities analysis. A wider distribution was
observed in the individual-cities analysis, with only 78% of the area less than zero. The
summary density curve of the individual-cities analysis has a larger standard error than the error
density curve of the all-cities analysis, because the summary density does not average the city
effects. The regression analysis by Mortimer et al. (2002) suggested a lack of heterogeneity by
city, as indicated by the nonsignificant interaction term between O3 effect and city. As shown in
Figure 7-4, the summary density curve of the individual-cities analysis has a peak at about the
same value as the curve of the all-cities analysis, suggesting a common O3 effect for all eight
7-45
-------
cities and small variation among them. The unimodal shape of the density curve of the
individual-cities analysis also indicates the absence of outlying cities.
Mortimer et al. (2002) further noted that small declines in morning PEF may be of
uncertain clinical significance; thus they calculated the incidence of >10% declines in PEF.
A 5 to 15% change in FEVj has been expressed as having clinical importance to asthma
morbidity (American Thoracic Society, 1991; Lebowitz et al., 1987; Lippmann, 1988).
Although greater variability is expected in PEF measurements, a > 10% change in PEF also may
have clinical significance. In Mortimer et al. (2002), O3 was associated with an increased
incidence of >10% declines in morning PEF (odds ratio of 1.30 [95% CI: 1.04, 1.61] per 30 ppb
increase in 8-h avg O3 for a 5-day cumulative lag). This finding suggests that exposure to O3
might be related to clinically important changes in PEF in asthmatic children. This study also
observed that excluding days when 8-h avg O3 levels were greater than 80 ppb provided effect
estimates that were similar to those when all days were included in the analysis, indicating that
the negative effect of O3 on morning PEF persisted at levels below 80 ppb. There is some
concern, however, regarding the lack of an association between O3 and afternoon PEF.
Results from the multicities study by Mortimer et al. (2002), as well as those from several
regional studies, provide evidence of a significant relationship between O3 concentrations and
PEF among asthmatics. Collectively, these studies indicate that O3 may be associated with
declines in lung function in this potentially susceptible population.
Panels of healthy subjects
The effect of O3 on PEF in healthy subjects also was investigated in several studies.
A study of 162 children (9 years of age) in England examined the relationship between O3 and
PEF in the winter and summer seasons (Ward et al., 2002). The median 24-h avg O3
concentrations were 13.0 ppb in the winter and 22.0 ppb in the summer. The O3 effect estimates
were generally positive in the winter and negative in the summer. Single-day lags of 0- to
3-days were examined; however, the strongest association was found with a multiday lag period.
During the summer, a decline of 11.10 L/min (95% CI: 0.18, 21.98) was observed in morning
PEF per 20 ppb increase in 24-h avg O3 with a 7-day cumulative lag. Smaller O3 effects were
observed on afternoon PEF.
7-46
-------
During the summer of 1990, Neas et al. (1995) examined 83 children in Uniontown, PA
and reported twice daily PEF measurements. Researchers found that evening PEF was
associated with O3 levels weighted by hours spent outdoors. Using a similar repeated measures
design, Neas et al. (1999) saw evidence for effects due to ambient O3 exposure among
156 children attending two summer day camps in the Philadelphia, PA area. The mean daytime
12-h avg O3 (9 a.m. to 9 p.m.) levels were 57.5 ppb at the southwestern camp and 55.9 ppb at the
northeastern camp. Negative associations were found between the preceding 12-h avg O3 and
afternoon PEF (recorded before leaving camp), as well as morning PEF (recorded upon arrival at
camp). However, the only significant relationship was between O3 and both morning and
afternoon PEF considered jointly when a multiday lag period was used. Naeher et al. (1999), in
a sample of 473 nonsmoking women (age 19 to 43 years) living in Vinton, VA, also showed the
strongest association between O3 and evening PEF with a 5-day cumulative lag exposure.
Another study in southwestern Mexico City analyzed morning and afternoon PEF data
collected from 40 school children aged 8 to 11 years (Gold et al., 1999). Subjects provided
measurements upon arriving and before departing from school each day. The mean 24-h avg O3
was 52.0 ppb (IQR 25). A negative effect of O3 on PEF was observed, -1.60 mL/s (95% CI:
-3.56, 0.36) and -1.80 mL/s (95% CI: -3.76, 0.16) per 20 ppb increase in 24-h avg O3 on the
same day afternoon and next day morning PEF, respectively. A greater effect was observed for
PEF regressed on O3 concentrations with a cumulative 10-day lag period (-3.50 mL/s [95% CI:
-5.52, -1.49] on same day afternoon). These results suggest a longer, cumulative effect of O3
on PEF. Alternatively, the associations observed at the 10-day lag period may reflect
confounding by other time-varying factors or be a chance finding from an exploratory analysis.
In a recent study of 43 mail carriers in Taichung City, Taiwan, PEF was monitored twice
daily during a six-week period (Chan and Wu, 2005). The mean 8-h avg O3 (9 a.m.-5 p.m.)
concentration during their work shift was 35.6 ppb (SD 12.1). Associations were observed
between evening PEF and O3 concentrations at lags of 0, 1 and 2 days. The greatest effect was
observed at a lag of 1 day, a 2.07% decline in PEF per 30 ppb increase in 8-h avg O3
(quantitative results for 95% CI not provided). Similar O3 effects on morning PEF were
observed. The effect of O3 on PEF was robust to adjustment for copollutants; no association
with PEF was observed for PM10 and NO2 in multipollutant models.
7-47
-------
7.2.4 Respiratory Symptoms
Studies published over the past decade represent an improved new body of data on the
symptom effects of O3. Respiratory symptoms in acute air pollution field studies are usually
measured using questionnaire forms (or "daily diaries") that are filled out by study subjects,
usually without the direct supervision of research staff. Questions address the daily experience
of coughing, wheezing, shortness of breath (or difficulty breathing), production of phlegm, and
others. While convenient and potentially useful in identifying acute episodes of morbidity,
measurements of daily symptoms are prone to a variety of errors. These include, for example,
misunderstanding of the meaning of symptoms, variability in individual interpretation of
symptoms, inability to remember symptoms if not recorded soon after their occurrence, reporting
bias if days of high air pollution levels are identifiable by subjects, and the possibility of falsified
data. In spite of these potential problems, the ease of data collection has made daily symptom
assessment a common feature of field studies. Many of the studies reviewed above for lung
function results also included measurements of daily symptoms. Pearce et al. (1998) reported
that one advantage associated with the study of asthma panels is that the population is usually
already familiar with symptom terms, e.g., wheezing and cough. Delfmo et al. (1998a) further
states that the use of repeated daily symptom diaries has additional advantages of reducing recall
bias, given the proximity of events, and allowing for health effects to be modeled with each
subject serving as their own control over time. Also, study design can blind the participants
from the air pollution aspect of the study. Careful efforts by study staff can help ensure that the
symptom diaries provide information that is less affected by the potential problems noted above.
Similar to studies of lung function, respiratory symptom studies can be divided into two
groups, asthma panels or healthy subjects. Asthma panel studies are discussed first.
Asthma panels
Most studies examining respiratory symptoms related to O3 exposure focused on asthmatic
children. Among the health outcomes, of particular interest were those associated with asthma,
including cough, wheeze, shortness of breath, and increased medication use. Figures 7-5 and 7-6
present the odds ratios for O3-related cough and medication use among asthmatic children from
six studies (Gielen et al., 1997; Jalaludin et al., 2004; Just et al., 2002; Ostro et al., 2001; Romieu
et al., 1996, 1997). For consistency, only single-city or single-region studies that presented odds
7-48
-------
>.o
!> 1.8
O
O
•5 1.6
-------
t.o
4.0 -
IV
= 3.5^
O
re
| 3.0-
05
2
4= 25 -
X, Z'°
LiJ
0 2.0 -
ts
t£
% 1.5-
o
1.0 -
n ^
| Single-day Lags |
i
* 1 * - 1
I * |
| Multlday Lags |
i
<
-------
ratios are included in the figure. Studies that presented change in severity of symptoms or some
other informative health outcome are not included in the figure, since such symptom outcomes
differ from indicating simple presence or absence of symptoms. The study by Gent et al. (2003)
also is not included in this figure, as odds ratios for cough and medication use were analyzed for
quintiles of O3 concentrations using the lowest quintile as the reference. These studies are
discussed separately.
The various effect estimates for the association between O3 concentrations and cough are
depicted in Figure 7-5. Despite the variability in the individual effect estimates, there is some
consistency in the O3 effects. In general, the majority of the odds ratios appear to be greater
than one among the single-day lag models, suggesting an association between acute exposure
to O3 and increased cough among asthmatic children. Figure 7-6 presents the odds ratios
for O3-associated bronchodilator use. The results for medication use are less consistent than
those for cough; one study by Just et al. (2002) observed strong positive associations, but had
wide confidence intervals.
Among the studies reporting results for daily symptoms and asthma medication use,
several observed associations with O3 concentrations that appeared fairly robust (Delfmo et al.,
2003; Desqueyroux et al., 2002a,b; Gent et al., 2003; Hiltermann et al., 1998; Just et al., 2002;
Mortimer et al., 2000, 2002; Newhouse et al., 2004; Romieu et al., 1996, 1997; Ross et al., 2002;
Thurston et al., 1997). Mortimer et al. (2002) reported morning symptoms in 846 asthmatic
children from 8 U.S. urban areas to be most strongly associated with a cumulative 1- to 4-day lag
of O3 concentrations in the NCICAS. The NCICAS used standard protocols that included
instructing caretakers of the subjects to record symptoms in the daily diary by observing or
asking the child (Mitchell et al., 1997). Symptoms reported included cough, chest tightness, and
wheeze. In the analysis pooling individual subject data from all eight cities, the odds ratio for
the incidence of symptoms was 1.35 (95% CI: 1.04, 1.69) per 30 ppb increase in 8-h avg O3
(10 a.m.-6 p.m.). The mean 8-h avg O3 was 48 ppb across the 8 cities. Excluding days when 8-h
avg O3 was greater than 80 ppb (less than 5% of days), the odds ratio was 1.37 (95% CI: 1.02,
1.82) for incidence of morning symptoms. Figure 7-7 presents the probability density curves of
the odds ratios for the incidence of symptoms with a 1- to 4-day cumulative lag from the
individual-cities analysis and the all-cities analysis. This figure confirms the regression results
indicating that there is a significant increase in odds for incidence of symptoms, as the area
7-51
-------
Individual-Cities Analysis
All-Cities Analysis
1 1.5
Odds Ratio for Incidence of Symptoms
Figure 7-7. Odds ratio for the incidence of symptoms per 30 ppb increase in 8-h avg O3
with a cumulative lag of 1 to 4 days. The error density curve is shown for the
pooled analysis of the eight NCICAS cities and the summary density curve is
presented for the analyses from the individual cities. Note that 99% and 76%
of the areas under the curves are less than zero for the pooled-cities analysis
and individual-cities analysis, respectively.
Source: Derived from Mortimer et al. (2002).
under the density curve with an odds ratio greater than one is 99%. Mortimer et al. (2002)
did not observe significant interactions among the eight cities, indicating that there was no
heterogeneity among the city-specific estimates. The unimodal distribution of the city-stratified
summary density curve also suggests a lack of significant heterogeneity in O3 effects among the
cities. It should be noted that other pollutants, including PM10 (monitored in 3 cities), NO2
(in 7 cities), and SO2 (in all 8 cities), also were associated with increased incidence of morning
symptoms. In multipollutant models, the O3 effect was shown to be slightly diminished. The
odds ratios for the incidence of symptoms per 30 ppb increase in 8-h avg O3 were 1.23 (95% CI:
0.94, 1.61) with SO2 and 1.14 (95% CI: 0.85, 1.59) with NO2. In the three urban areas
with PM10 data, the odds ratios were 1.21 (95% CI: 0.61, 2.40) in the O3-only model and
1.08 (95% CI: 0.41, 2.40) when PM10 also was included in the model.
Another one of the larger studies was that of Gent and colleagues (2003), where
271 asthmatic children under age 12 and living in southern New England were followed over
6 months (April through September) for daily symptoms in relation to O3 and PM25. Mean 1-h
max O3 and 8-h max O3 concentrations were 58.6 ppb (SD 19.0) and 51.3 ppb (SD 15.5),
7-52
-------
respectively. The data were analyzed for two separate groups of subjects, 130 who used
maintenance asthma medications during the follow-up period and 141 who did not. The need for
regular medication was considered to be a proxy for more severe asthma. Not taking any
medication on a regular basis and not needing to use a bronchodilator would suggest the
presence of very mild asthma. Effects of 1-day lag O3 were observed on a variety of respiratory
symptoms only in the medication user group. Both daily 1-h max and 8-h max O3 concentrations
were similarly related to symptoms such as chest tightness and shortness of breath. Effects
of O3, but not PM2 5, remained significant and even increased in magnitude in two-pollutant
models. Some of the associations were noted at 1-h max O3 levels below 60 ppb. In contrast, no
effects were observed among asthmatics not using maintenance medication. In terms of person-
days of follow-up, this is one of the larger studies currently available that address symptom
outcomes in relation to O3, and provides supportive evidence for effects of O3 independent
of PM2 5. Study limitations include limited control for meteorological factors and the post-hoc
nature of the population stratification by medication use.
Some international studies have reported significant associations of respiratory symptoms
with O3. The incidence of asthma attacks was associated with O3 concentrations in a group of
60 severe asthmatics (mean age 55 years) followed over a 13-month period in Paris, France
(Desqueyroux et al., 2002a). In a similar study, Desqueyroux et al. (2002b) observed O3-
associated exacerbation of symptoms in 39 adult patients (mean age 67 years) with chronic
obstructive pulmonary disease (COPD). Interestingly, in contrast to results from controlled
human exposure studies (see Section 6.3.1, Subjects with COPD), the O3 effect appeared larger
in this study among subjects who smoked and among those with more severe COPD. However,
the low O3 concentrations experienced during this study (summer mean 8-h max O3 of 21 ppb
[SD 9]) raise plausibility questions. In a study of 60 nonsmoking asthmatic adults (aged 18 to
55 years) in Bilthoven, the Netherlands, Hiltermann and colleagues (1998) reported associations
between O3 and daily symptoms of shortness of breath and pain upon deep inspiration. The
mean 8-h max O3 was 41 ppb (range 3 to 49). The O3 associations were stronger than those
for PM10, NO2, SO2, and black smoke (BS). No differences in response were evident between
subgroups of subjects defined on the basis of steroid use or airway hyperresponsiveness. Daily
use of bronchodilators or steroid inhalers was not found to be associated with O3 in this study.
7-53
-------
Other studies showed only limited or a lack of evidence for symptom increases being
associated with O3 exposure (Avol et al., 1998; Chen et al., 1998; Delfmo et al., 1996, 1997a,
1998a; Gielen et al., 1997; Jalaludin et al., 2004; Ostro et al., 2001; Taggart et al., 1996). Avol
et al. (1998) studied symptoms in asthmatic, wheezy, and healthy children aged 10 to 12 years in
southern California. Some symptom associations were noted but they were inconsistent. For
example, children with wheeze were at increased risk for difficulty of breathing and for
wheezing at low O3 concentrations (1-h max O3 of 30 ppb), but not at higher O3 concentrations
(1-h max O3 of 120 ppb). The authors noted that O3 concentrations were relatively low and that
children studied did not spend substantial time engaged in physical activities outdoors . Ostro
et al. (2001) reported no associations between daily symptoms and ambient O3 concentrations in
a cohort of 138 African-American children with asthma followed over 3 months (August to
October) in Central Los Angeles and Pasadena, CA. However, the use of extra asthma
medication was associated with 1-h max O3 concentrations at a 1-day lag. Mean 1-h max O3
concentrations were 59.5 ppb in Los Angeles and 95.8 ppb in Pasadena. Delfmo and colleagues
(1996) followed 12 asthmatic teens living in San Diego, CA for respiratory symptoms over a
2-month period and saw no relationship with central site ambient O3 (mean 12-h avg O3 of
43 ppb [SD 17]). Personal O3 exposures (mean 12-h avg O3 of 11.6 ppb [SD 11.2]) were
associated with the composite symptom score and p2-agonist inhaler use, but the relationship
with symptom score disappeared when weekday/weekend differences were controlled in the
statistical analysis. Study power was likely compromised by the small sample size. This
observation of stronger associations with O3 levels from personal monitors implies that gains in
power may be achieved if exposure misclassification is reduced through the use of personal
exposure measurements rather than central site ambient O3 concentrations. A similar study of
22 asthmatics in Alpine, CA observed no effects of O3 on symptoms when personal O3 exposure
(mean 12-h avg O3 of 18 ppb) was used as the exposure metric (Delfmo et al., 1997a). However,
a later study in the same location involving 24 subjects (Delfmo et al., 1998a) did find an
association between respiratory symptoms and ambient O3 concentrations (mean 1-h max O3 of
90 ppb), with stronger O3 effects experienced by asthmatics not on anti-inflammatory
medication. In this study, a binary symptom score was used, whereas the earlier study used a
linear symptom score of 0 through 6.
7-54
-------
The multicities study by Mortimer et al. (2002), which provides an asthmatic population
most representative of the United States, and several single-city studies indicate a robust
association of O3 concentrations with respiratory symptoms and increased medication use in
asthmatics. However, there are a number of well-conducted, albeit relatively smaller, studies
that have not found such effects.
Panels of healthy subjects
Fewer studies examined the effect of O3 on respiratory symptoms in healthy individuals.
Neas et al. (1995) reported that, in school children, evening cough was associated with O3 levels
weighted by hours spent outdoors. The mean daytime 12-h avg O3 (8 a.m.-8p.m.) concentration
was 50.0 ppb. The study by Linn and colleagues (1996) of 269 school children in southern
California found no associations between respiratory symptoms and O3, but subjects were
exposed to fairly low O3 concentrations (mean 24-h avg O3 of 5 ppb [SD 3]) as determined by
personal monitors. Gold et al. (1999) examined symptoms in 40 healthy children in southwest
Mexico City. The mean ambient 24-h avg O3 concentration was 52.0 ppb (IQR 25). Pollutant
exposures were associated with increased production of phlegm in the morning, although the
effects of the air pollutants (PM2 5, PM10, and O3) could not be separated in multipollutant
models. Hoek and Brunekreef (1995) did not find a consistent association between ambient O3
levels and the prevalence or incidence of respiratory symptoms in children living in two rural
towns in the Netherlands. Mean 1-h max O3 concentrations were 57 ppb (range 22 to 107)
in Deurne and 59 ppb (range 14 to 114) in Enkhuizen. Collectively, these studies indicate that
there is no consistent evidence of an association between O3 and respiratory symptoms among
healthy children.
7.2.5 Acute Airway Inflammation
Acute airway inflammation has been shown to occur among adults exposed to 80 ppb O3
over 6.6 hours with exercise in controlled chamber studies (Devlin et al., 1991). Kopp and
colleagues (1999) attempted to document inflammation of the upper airways in response to
summer season O3 exposures by following a group of 170 school children in two towns in the
German Black Forest from March to October of 1994. To assess inflammation, nasal lavage
samples were collected at 11 time points spanning the follow-up period. The mean Va-h max O3
7-55
-------
concentrations were 33 ppb (5th % to 95th %: 1 to 72) in Villingen and 54 ppb (5th % to
95th %: 23 to 92) in Freudenstadt. The nasal lavage samples were analyzed for markers of
inflammation, including eosinophil cationic protein, albumin, and leukocyte counts. Subjects
who were sensitized to inhaled allergens were excluded. When analyzed across the entire
follow-up period, no association was detected between upper airway inflammation and O3
concentrations. More detailed analysis showed that the first significant O3 episode of the
summer was followed by a rise in eosinophil cationic protein levels; however, subsequent
and even higher O3 episodes had no effect. These findings suggest an adaptive response of
inflammation in the nasal airways that is consistent with controlled human exposure studies (see
Section 6.9, Effects of Inflammation and Host Defense).
Frischer and colleagues (1993) collected nasal lavage samples from 44 school children in
Umkirch, Germany the morning after "low" O3 days (<140 |ig/m3 or approximately 72 ppb) and
"high" O3 days (>180 |ig/m3 or approximately 93 ppb) to measure levels of biochemical markers
of inflammation. The researchers found that higher O3 levels were associated with increased
polymorphonuclear leukocyte counts in all children, and increases in myeloperoxydases and
eosinophilic cation proteins among children without symptoms of rhinitis (n = 30). These results
indicated that O3 was associated with inflammation in the upper airways. Frischer et al. (1997)
further investigated whether hydroxyl radical attacks played a role in mediating the O3-
associated inflammatory response of the airways. Ortho- and/>ara-tyrosine levels were
measured in the nasal lavage samples and the ortholpara radical ratio was used to determine the
generation of hydroxyl radicals. Significant increases in the ortholpara ratio were observed on
days following high ambient O3 levels. However, the ortholpara ratio was not related to
polymorphonuclear leukocyte counts, suggesting that there was no detectable relationship
between hydroxyl radical attacks and the inflammatory response seen in these children. Similar
to the study by Kopp et al. (1999), the ortholpara ratio decreased at the end of the summer,
although O3 concentrations were still high—thus providing additional evidence for a possible
adaptive response. These findings, however, do not preclude the possibility that other
unmeasured effects, including cell damage or lower airway responses, may have occurred with
ongoing summer season exposures. In fact, a study of joggers repeatedly exposed to O3 while
exercising over the summer in New York City suggested that cell damage may occur in the
absence of ongoing inflammation (Kinney et al., 1996b).
7-56
-------
In two Mexico City studies by Romieu et al. (1998, 2002), the effect of antioxidant
supplements on the association between O3 and lung function in outdoor workers and asthmatic
children was investigated. Mean 1-h max O3 concentrations were 123 ppb for the outdoor
workers and 102 ppb for the children. Romieu and colleagues (1998) observed significant
inverse associations between O3 and lung function parameters, including FVC, FEVl3 and
FEF25.75 (forced expiratory flow at 25 to 75% of FVC), among outdoor workers who received the
placebo but not among those taking the antioxidant supplement during the first phase of testing.
Likewise, O3 concentrations were associated with declines in lung function among children with
moderate-to-severe asthma who were on the placebo, but no associations were found among
those who were taking the vitamin C and E supplement (Romieu et al., 2002). These results
indicate that supplementation with antioxidants may modulate the impact of O3 exposure on the
small airways of two potentially at-risk populations, i.e., outdoor workers and children with
moderate-to-severe asthma. In a further analysis of the study conducted in asthmatic children,
genetic factors were found to contribute to the variability between individuals in the effects of O3
on lung function (Romieu et al., 2004). Individuals with polymorphism of the glutathione
S-transferase gene (GSTM1 null genotype) lack glutathione transferase enzyme activity, which
plays an important role in protecting cells against oxidative damage. Results from this analysis
indicate that asthmatic children with GSTM1 null genotype were found to be more susceptible to
the impact of O3 exposure on small airways function. Romieu et al. (2004) noted that
supplementation with the antioxidant vitamins C and E above the minimum daily requirement
might compensate for the genetic susceptibility.
7.2.6 Acute Ozone Exposure and School Absences
The association between school absenteeism and ambient air pollution was assessed in a
limited number of studies (Chen et al., 2000; Gilliland et al., 2001). In the study by Chen and
colleagues (2000), daily school absenteeism was examined in 27,793 students (kindergarten to
sixth grade) from 57 elementary school students in Washoe County, NV over a 2-year period.
The mean 1-h max O3 was 37.5 ppb (SD 13.4) during the study period. One major limitation of
this study was that the percent of total daily absences was the outcome of interest, not illness-
related absences, because reasons for absences were not noted in all schools. In models
adjusting for PM10 and CO, ambient O3 levels were associated with school absenteeism. With a
7-57
-------
distributed lag of 1 to 14 days, O3 concentrations were associated with a 10.4% (95% CI: 2.7,
18.1) excess rate of school absences per 40 ppb increase in 1-h max O3. PM10 and CO
concentrations also were associated with school absenteeism; but, the effect estimate for PM10
was negative. The inverse relationship between O3 and PM10 may have partially attributed to the
negative association observed between PM10 and school absenteeism.
Ozone-related school absences also were examined in a study of 1,933 fourth grade
students from 12 southern California communities participating in the Children's Health Study
(Gilliland et al., 2001). Due to its comprehensive characterization of health outcomes, this study
is valuable in assessing the effect of O3 on illness-related school absenteeism in children. The
study spanned a period, January through June 1996, that captured a wide range of exposures
while staying mostly below the highest levels observed in the summer season. Mean 8-h avg O3
(10 a.m.-6 p.m.) concentrations ranged from 35 to 55 ppb across the 12 communities. All school
absences that occurred during this period were followed up with phone calls to parents to
determine whether they were illness-related. For illness-related absences, further questions
assessed whether the illness was respiratory or gastrointestinal, with respiratory symptoms
including runny nose/sneeze, sore throat, cough, earache, wheezing, or asthma attacks. Multiple
pollutants were measured at a central site in each of the 12 communities. A two-stage GAM was
used to examine the effects of O3 on school absences. The analysis controlled for temporal
cycles, day of week, and temperature, and expressed exposure as a distributed lag out to 30 days.
The 30-day distributed lag was found to best fit the data. Associations were found between the
30-day distributed lag of 8-h avg O3 (10 a.m.-6 p.m.) and all illness-related absence categories.
A 108% (95% CI: 29, 235) increase in illness-related absences was observed with a 30 ppb
increase in 8-h avg O3. Larger O3 effects were seen for respiratory causes (147% [95% CI: 6,
478] than for nonrespiratory causes (61% [95% CI: 9, 138]). Among the respiratory absences,
larger effects were seen for lower respiratory diseases than for upper respiratory diseases.
Multipollutant analyses were not performed; however, in single-pollutant models neither PM10
or NO2 were associated with any respiratory or nonrespiratory illness-related absences. Some
concern exists regarding the possibility of residual seasonal confounding given the 6-month time
span of the monitoring period and the long lag periods of exposure, which are likely to capture
seasonally changing factors such as pollen episodes. Further, the biological relevance of O3
concentrations lagged 30 days present an interpretive challenge.
7-58
-------
The school absenteeism data from the Children's Health Study was used to illustrate
alternative new modeling approaches in two subsequent papers. Berhane and Thomas (2002)
noted that systematic comparison of modeling approaches could potentially provide more insight
into the modeling process. They used a two-stage iteratively weighted filtered least squares
model which allowed explicit modeling of the autoregressive structure and overdispersion; this
model provided information pertaining to how air pollution operates in various communities.
Results indicated that the overall model might not be adequate for some communities, but the
community-specific correction for overdisperion and autocorrelation compensated for this
shortcoming. For a 30-day distributed lag, the effect estimate for illness-related absences was
similar, 105% (95% CI: -10, 368) increase per 30 ppb increase in 8-h avg O3. The largest O3
effect was observed in the communities with low long-term O3, PM10, and NO2 concentrations.
Once again, PM10 was not associated with illness-related absences. Analyses for specific illness
categories were not conducted. Berhane and Thomas (2002) also examined effects from long-
term exposure to air pollutants. Long-term average concentrations of PM10, but not O3, were
associated with increases in illness-related absenteeism.
A further analysis by these authors (Rondeau et al., 2005) used another new approach,
a three-stage logistic transition model. Unlike the previous two analyses, this analysis allowed
the simultaneous examination of the effects of daily exposure to air pollution and individual risk
factors using binary time-series data structures, without aggregating over subjects or time.
In contrast to the results from Berhane and Thomas (2002), the Rondeau et al. results suggested
a chronic effect of O3 on school absenteeism (41% [95% CI: 2, 95] increase per 30 ppb increase
in 8-h avg O3 over a 5-day lag period) but no acute effect (-0.2% [95% CI: -24, 31] with a
30-day distributed lag) after adjustment for individual factors. The acute O3 effect on respiratory
absences was positive but nonsignificant (12% [95% CI: -22, 62]). Rondeau et al. did not
compare their results to those from Gilliland et al. Statistical tests conducted to compare the
effect estimates indicated that the acute O3 estimates for total illness-related absences were
significantly different (p = 0.03) in the two studies; however, the estimates for respiratory
absences were not found to be different (p = 0.15). Both acute and chronic exposure to PM10
was not associated with illness-related school absences. The authors noted that the different
results compared to the analyses by Gilliland et al. (2001) and Berhane and Thomas (2002)
might be due to different modeling approaches or the use of binary instead of Poisson time-series
7-59
-------
data structures. The two methods papers by Berhane and Thomas (2002) and by Rondeau et al.
(2005) focused on introducing new methods and used the Children's Health Study data for
illustrative purposes, as opposed to comprehensive analysis of the substantive findings.
Hence, greatest emphasis should be placed on the results from Gilliland et al (2001).
Limited evidence from Chen et al. (2000) and Gilliland et al. (2001) suggest that
ambient O3 concentrations, accumulated over 2 to 4 weeks, may be associated with school
absenteeism, particularly illness-related absences. Further replication is needed before firm
conclusions can be reached regarding the effect of O3 on school absences.
7.2.7 Cardiovascular Endpoints
Several air pollution studies have examined various cardiovascular endpoints (see
Table AX7-2 in Annex 7, Section AX7.1). The earlier studies focused on PM effects. For a
more thorough discussion of these PM studies and their health endpoints, refer to Section 8.3.1
of the 2004 PM AQCD (U.S. Environmental Protection Agency, 2004). More recent studies
have examined associations of O3 and other pollutants with various measures of heart beat
rhythms in panels of elderly subjects, as discussed below. Other studies examined the increased
risk of MI related to air pollutant exposures.
7.2.7.1 Cardiac Autonomic Control
Alterations in heart rate and/or rhythm are thought to reflect pathophysiologic changes that
may represent possible mechanisms by which ambient air pollutants such as O3 may exert acute
effects on human health. Decreased HRV has been identified as a predictor of increased
cardiovascular morbidity and mortality. Brook et al. (2004) state that HRV, resting heart rate,
and blood pressure are modulated by a balance between the two determinants of autonomic tone
(the sympathetic and parasympathetic nervous systems). They noted that decreased HRV
predicts an increased risk of cardiovascular morbidity and mortality in the elderly and those with
significant heart disease, which is generally determined by analyses of time (e.g., standard
deviation of normal R-R intervals) and frequency domains (e.g., low frequency/high frequency
ratio by power spectral analysis, reflecting autonomic balance) measured during 24 hours of
electrocardiography. Decreased parasympathetic input to the heart may provide an important
7-60
-------
mechanistic link between air pollution and cardiovascular mortality by promoting fatal
tachy arrhy thmi as.
The potentially adverse effects of air pollutants on cardiac autonomic control were
examined in a large population-based study, among the first in this field. Liao et al. (2004)
investigated short-term associations between ambient pollutants and cardiac autonomic control
from the fourth cohort examination (1996-1998) of the population-based Atherosclerosis Risk in
Communities Study (ARIC). PM10, O3, and other gaseous air pollutants were examined in this
study. PM10 (24-h avg) and O3 exposures (8-h avg, 10 a.m. to 6 p.m.) one day prior to the
randomly allocated examination date were used. The mean 8-h avg O3 level was 41 ppb
(SD 16). They calculated 5-minute HRV indices between 8:30 a.m. and 12:30 p.m. and used
logarithmically-transformed data on high-frequency (0.15 to 0.40 Hz) and low-frequency
(0.04 to 0.15 Hz) power, standard deviation of normal R-R intervals, and mean heart rate. The
effective sample sizes for O3 and PM10 were 5,431 and 4,899, respectively, from three U.S. study
centers in North Carolina, Minnesota, and Mississippi. PM10 concentrations measured one
day prior to the HRV measurements were inversely associated with both frequency- and
time-domain HRV indices. Ambient O3 concentrations were inversely associated with
high-frequency power among whites. Consistently more pronounced associations were
suggested between PM10 and HRV among persons with a history of hypertension. Liao et al.
(2004) noted that these findings may represent potentially important arrhythmogenic
mechanisms of ambient air pollution. The acute adverse effect of air pollution on cardiac
autonomic control is based on the hypothesis that increased air pollution levels may stimulate the
autonomic nervous system and lead to an imbalance of cardiac autonomic control characterized
by sympathetic activation unopposed by parasympathetic control. Such an imbalance of cardiac
autonomic control may predispose susceptible people to greater risk of life-threatening
arrhythmias and acute cardiac events. The Liao et al. (2004) findings were cross-sectionally
derived from a population-based sample and reflect the short-term effects of air pollution on
HRV. When the regression coefficients for each individual pollutant model were compared, the
effects for PM10 were considerably larger than the effects for gaseous pollutants such as O3.
Because of the population-based sample, this study does have better generalizability than other
smaller panel studies. The findings are suggestive of short-term effects of air pollutants,
including O3, on HRV at the population level.
7-61
-------
Another population-based study of air pollutants and HRV was conducted in Boston, MA
on 497 men from the VA Normative Aging Study (NAS) (Park et al., 2005). The mean 24-h
avg O3 concentration was 23.0 ppb (SD 13.0). Several associations with HRV outcomes were
observed with a 4-h moving average of O3 concentrations; effects generally diminished with
longer averaging times of 24 and 48 hours. Stronger associations were reported with PM25.
In two-pollutant models, the magnitude of the percent changes for both PM2 5 and O3 diminished
slightly. In analyses by ischemic heart disease, hypertension, and diabetes status, stronger
associations of HRV with O3 and PM25 were observed for individuals with ischemic heart
disease and hypertension. These results are consistent with a Mexico City study (n = 34) by
Holguin et al. (2003) which reported an HRV effect for O3 in subjects with hypertension. The
association of O3 exposure with reduced low-frequency power in the full cohort seemed to be
driven by subjects not taking calcium-channel blockers (Park et al., 2005). This suggests that
this drug is blocking effects of O3 on the sympathetic pathway. This study cohort consists of all
males and almost all whites. This population-based study suggests that short-term exposure
to O3 is a predictor of alterations in cardiac autonomic function as measured by HRV among
older male adults. A potential limitation of this study is that electrocardiograms were only taken
once for each subject, so subject-specific variation of HRV measures may not be ruled out as a
potential confounder.
Two separate analyses of the same cohort of patients examined the association between air
pollution and the incidence of ventricular arrhythmias (Dockery et al., 2005; Rich et al., 2005).
A total of 203 patients with implanted cardioverter defibrillators who lived within 25 miles of
the ambient monitoring site at the Harvard School of Public Health, Boston, MA, were
monitored. They had a total of 635 person-years of follow-up or an average of 3.1 years per
subject. The median 48-h avg O3 concentration was 22.9 ppb (IQR 15.4). In the analysis by
Dockery et al. (2005), positive associations were observed between ventricular arrhythmias
within 3 days of a prior event and a 2-day mean of several air pollutants, including PM2 5, black
carbon, NO2, CO, and SO2. No associations were observed with O3. There was, however, a
suggestion of increasing risk with increasing quintiles of O3 (p <0.05). The analysis by Rich
et al. (2005) observed stronger O3 effects on ventricular arrhythmias using a case-crossover
study design. Case periods were defined by the time each arrhythmic event began; for each case,
three to four control periods were selected by matching on weekday and hour of the day within
7-62
-------
the same calender month. The median 24-h avg O3 level was 22.6 ppb (IQR 15.7). For a 20 ppb
increase in 24-h moving average O3, a 27% (95% CI: 0, 60) increased risk of ventricular
arrhythmias was estimated. Significant effects also were found for PM2 5, NO2, and SO2. In two-
pollutant models, the O3 effect was found to be generally robust. Stratified analysis by the
presence of a recent ventricular arrhythmia within the previous three days indicated that O3 was
associated with increased risk among subjects without a recent event (37% [95% CI: 6, 79]), but
not among those with recent events (5% [95% CI: -27, 49]). Rich et al. (2005) explained that
the use of the case-crossover study design and conditional analysis might have contributed to the
stronger associations observed in their study compared to Dockery et al. (2005). The
case-crossover design and conditional analysis controlled for season, time trends, weekday, as
well as other non-time-varying confounders such as underlying medical conditions and smoking
status by design, thereby eliminating any residual confounding. In the analysis by Dockery et al.
(2005), these factors were controlled through modeling. In addition, the use of a 24-h moving
average instead of a calendar-day air pollution concentration might have reduced exposure
misclassification, resulting in larger effect estimates.
Other studies do not provide evidence for an O3 effect on HRV and cardiac arrhythmias
(Peters et al., 2000a; Rich et al., 2004; Vedal et al., 2004). These studies, however, may have
had limited power to examine subtle effects. Gold et al. (2000; reanalysis Gold et al., 2003)
reported results that suggest that O3 exposure may decrease vagal tone, leading to reduced HRV.
In this Boston, MA study, the mean 1-h max O3 level was 25.7 ppb (IQR 23.0). In another
Boston study, Schwartz et al. (2005) reported a weak association of O3 with the root mean
squared differences between adjacent R-R intervals in a study of 28 elderly subjects. The
median 1-h max O3 level was 34 ppb (IQR 26). The authors noted that lack of personal exposure
measurements might render such studies less able to assess autonomic functions. This study by
Schwartz et al. (2005) reported the strongest effects for black carbon.
7.2.7.2 Acute Myocardial Infarction
The effect of O3 on the incidence of MI was examined in a limited number of studies.
Acute MI was studied in relation to air pollution in Toulouse, France based on the existence of
an acute MI registry (Monitoring Trends and Determinants in Cardiovascular Disease
[MONICA]) and an air quality network covering the same population (Ruidavets et al., 2005).
7-63
-------
The mean 8-h max O3 level was 38.6 ppb (range 2.0-82.6). After adjustment for temperature,
relative humidity, and influenza epidemics, the relative risk of acute MI occurrence was 1.76
(95% CI: 1.12, 2.45) for current-day O3 concentrations. The increased risk of MI was more
evident in the oldest group, 55 to 64 years of age. Further, the oldest subjects without a personal
history of ischemic heart disease were more susceptible to an acute event when O3 levels
increased. No PM data was reported in this study.
In a case-crossover study (n = 772) in Boston, MA, Peters et al. (2001) reported an odds
ratio of 1.27 (95% CI: 0.87, 1.88) per 40 ppb increase in 2-h avg O3 (1 hour before onset of
event). The mean 24-h avg O3 level was 19.9 ppb (5th % to 95th %: 6 to 36). Stronger effects
on the incidence of MI were observed for PM25 and PM10. It should be noted that Peters et al.
(2001) used unidirectional sampling (i.e., control periods were selected only before the case
period) in their case-crossover design. Unidirectional sampling can lead to time trend bias and
overlap bias (i.e., biased conditional logistic regression estimating equations), which can result
in overestimated effects of exposure (Janes et al., 2005).
7.2.7.3 Cardiovascular Endpoints in Human Clinical Studies
In a controlled human exposure study discussed in Chapter 6, Sections 6.3.4 and 6.10,
Gong et al. (1998a) studied 10 nonmedicated hypertensive and 6 healthy male adults exposed to
0.3 ppm O3 with intermittent exercise in relation to various cardiovascular effects. The overall
results did not indicate acute cardiovascular effects of O3 in either the hypertensive or control
subjects. The authors observed an increase in rate-pressure product and heart rate, a decrement
for FEVj, and a >10 mm Hg increase in the alveolar/arterial pressure difference for O2
following O3 exposure. These findings suggest that O3 may exert cardiovascular effects
indirectly by impairing alveolar-arterial O2 transfer and potentially reducing O2 supply to the
myocardium. Ozone exposure may increase myocardial work and impair pulmonary gas
exchange to a degree that could perhaps be clinically important in persons with significant pre-
existing cardiovascular impairment.
1.2.1 A Summary of Field Studies with Cardiovascular Outcomes
A limited epidemiologic database examining cardiovascular outcomes in relation to O3
exposures is available. Among these studies, three were population-based and involved cohorts
7-64
-------
such as the ARIC (Liao et al., 2004), MONICA (Ruidavets et al., 2005), and NAS (Park et al.,
2005). Such studies may offer more informative results based on their large subject-pool and
design. Results from these three studies were suggestive of an association between O3 exposure
and the cardiovascular endpoints studied. As in the case of respiratory disease outcomes, Brook
et al. (2004) stated that the increase in relative risk for cardiovascular disease due to air pollution
is small compared with the impact of the established cardiovascular risk factors. However,
because of the enormous number of people affected, even conservative risk estimates can
translate into a substantial increase in mortality due to cardiovascular disease within the
population and, therefore, could potentially imply a notable public health problem.
7.2.8 Summary of Field Studies Assessing Acute Ozone Effects
• Results from recent field/panel studies support the evidence from controlled human
exposure studies that acute O3 exposure is associated with a significant effect on lung
function, as indicated by decrements in FEVl3 FVC, and PEF. The declines in lung
function were noted particularly in children and asthmatics.
• Limited evidence suggests that more time spent outdoors, higher levels of exertion,
and the related increase in O3 exposure may potentiate the risk of respiratory effects.
In addition to children and asthmatics, adults who work or exercise outdoors may be
particularly vulnerable to O3-associated health effects.
• Many new studies have examined the association between O3 concentrations and a
wide variety of respiratory symptoms (e.g., cough, wheeze, production of phlegm,
and shortness of breath). Collectively, the results suggest that acute exposure to O3
is associated with increased respiratory symptoms and increased as-needed medication
use in asthmatic children.
• Other panel studies evaluated O3 effects on other health outcomes, including school
absences and markers of inflammation and oxidative damage. Ozone exposure was
associated with increased inflammation and generation of hydroxyl radicals in the
upper airways. Use of antioxidant supplements was found to diminish the O3 effect on
lung function. Some studies suggest that O3 exposure, accumulated over 2-4 weeks, is
associated with increases in respiratory-related school absences, but further replication
is needed before firm conclusions can be drawn about O3 effects on school absences.
• Some field studies have examined the association between O3 and cardiac physiologic
outcomes. The current evidence is rather limited but suggestive of a potential effect on
HRV, ventricular arrhythmias, and MI incidence.
7-65
-------
7.3 ACUTE EFFECTS OF OZONE ON DAILY EMERGENCY
DEPARTMENT VISITS AND HOSPITAL ADMISSIONS
7.3.1 Summary of Key Findings on Studies of Emergency Department Visits
and Hospital Admissions from the 1996 Ozone AQCD
In the 1996 O3 AQCD, aggregate population time-series studies of O3-related health effects
provided relevant evidence of acute responses, even below a 1-h max O3 of 0.12 ppm.
Emergency room visits and hospital admissions were examined as possible outcomes following
exposure to O3. In the case of emergency room visits, the evidence was limited (Bates et al.,
1990; Cody et al., 1992; Weisel et al., 1995; White et al., 1994), but results generally indicated
an O3 effect on morbidity. The strongest and most consistent evidence of O3 effects, at levels
both above and below 1-h max O3 levels of 0.12 ppm, was provided by the multiple studies that
had been conducted on summertime daily hospital admissions for respiratory causes in various
locales in eastern North America (Bates and Sizto, 1983, 1987, 1989; Burnett et al., 1994;
Lipfert and Hammerstrom, 1992; Thurston et al., 1992, 1994). These studies consistently
demonstrated that O3 air pollution was associated with increased hospital admissions, accounting
for roughly one to three excess respiratory hospital admissions per million persons with each
100 ppb increase in 1-h max O3. This association had been shown to remain even after
statistically controlling for the possible confounding effects of temperature and copollutants
(e.g., H+, SO4 2, PM10), as well as when considering only days with 1-h max O3 concentrations
below 0.12 ppm. Overall, the aggregate population time-series studies considered in the 1996 O3
AQCD provided strong evidence that ambient exposures to O3 can cause significant
exacerbations of preexisting respiratory disease in the general public.
7.3.2 Review of Recent Studies of Emergency Department Visits for
Respiratory Diseases
Emergency department visits represent an important acute outcome that may be affected
by O3 exposures. Morbidities that result in emergency department visits are closely related to,
but are generally less severe than, those that result in unscheduled hospital admissions. In many
cases, acute health problems are successfully treated in the emergency department; however,
a subset of more severe cases that present initially to the emergency department may require
hospital admission.
7-66
-------
Several studies have been published in the past decade that examined temporal associations
between O3 exposures and emergency department visits for respiratory diseases (Table AX7-3 in
Annex 7, Section AX7.1). Total respiratory causes for emergency room visits typically include
asthma, pneumonia, bronchitis, emphysema, upper and lower respiratory infections such as
influenza, and a few other minor categories. Asthma visits typically dominate the daily
incidence counts. Chronic bronchitis and emphysema often are combined to define COPD,
which is a prominent diagnosis among older adults with lung disease. Figure 7-8 presents
percent changes in emergency department visits for asthma from single-pollutant models,
with results expressed in standardized increments. The lags presented in the figure vary
depending on reported results. Most studies reported effect estimates from a short lag period
(0 to 2 days). Results from Weisel et al. (2002) are not included as comparable risks estimates
for O3 were not presented. Several of the studies conducted in the United States and Canada
examined the effects of O3 during the warm season only. In general, ambient O3 concentrations
were associated with emergency department visits for asthma in warm-season only analyses.
The warm-season effect estimates tended to be positive and larger than results for cool-season or
all-year analyses.
Among studies with adequate controls for seasonal patterns, many reported at least one
positive association with O3. These studies examined emergency department visits for total
respiratory complaints (e.g., Delfmo et al., 1997b, 1998b; Herfiandez-Garduno et al., 1997;
Ilabaca et al., 1999; Jones et al., 1995; Lin et al., 1999), asthma (e.g., Friedman et al., 2001;
Jaffe et al., 2003; Stieb et al., 1996; Tenias et al., 1998; Tobias et al., 1999; Tolbert et al., 2000;
Weisel et al., 2002), and COPD (Tenias et al., 2002).
One recent study examined emergency department visits for total and cause-specific
respiratory diseases in Atlanta, GA over an 8-year period (Peel et al., 2005). A distributed lag of
0 to 2 days was specified a priori. The mean 8-h max O3 concentration was 55.6 ppb (SD 23.8).
Ozone concentrations were associated with emergency department visits for total respiratory
diseases and upper respiratory infections in all ages. A marginally significant association was
observed with asthma visits (2.6% [95% CI: -0.5, 5.9] excess risk per 30 ppb increase in 8-h
max O3), which became stronger when the analysis was restricted to the warm months (3.1%
[95% CI: 0.2, 6.2] excess risk). In multipollutant models adjusting for PM10, NO2 and CO, O3
was the only pollutant that remained significantly associated with upper respiratory infections.
7-67
-------
% Change in Emergency Department Visits for Asthma
14)
Teniasetal. (1998):
Valencia, Spain (age > 14)
-20 0 20 40 60 80 100 120 140 160 180 200
I I I 1 I 1 I 1 I 1 I 1 I 1 I 1 I 1 1 1 1 I I
U.S. and Canada A , 0
inn o X Allypar
|Jy *- J
rag -5 •Warm
• hn ° or ?
O Cool
A lag 02
-K- lag 0-2
-•- lag 0-2
-•- Iag1
• lin ?
iug *.
Europe |
( Best of lag 0 to 3 by city
v- )
X — lagO
X lagO
• Ian ^
C\ Ian 0
— K lagO
Figure 7-8. Ozone-associated percent change (95% CI) in emergency department visits for asthma per
standardized increment (see Section 7.1.3.2).
7-68
-------
Another large asthma emergency department study was carried out during the months of May
through September from 1984 to 1992 in St. John, New Brunswick, Canada (Stieb et al., 1996).
The mean 1-h max O3 level was 41.6 ppb (range 0-160). Effects were examined separately
among children aged less than 15 years and in persons aged 15 years and older. A significant
effect of O3 on emergency department visits was reported among persons 15 years and older.
There was suggestion of a threshold somewhere in the range below a 1-h max O3 of 75 ppb.
Another study in Valencia, Spain from 1994 to 1995 found that emergency room visits for
asthma among persons over 14 years old were robustly associated with relatively low O3 levels
(mean 1-h max O3 of 32.4 ppb [range 6.9-81.2]) (Tenias et al., 1998). The excess risk of asthma
emergency room visits was larger in the warm season (May to October; mean 1-h max O3 of
38.2 ppb), with an 85% (95% CI: 20, 188) excess risk per 40 ppb increase in 1-h max O3.
During the cool season (November-April; mean 1-h max O3 of 26.5 ppb), the excess risk was
31%(95%CI: -24, 125).
Among the studies that reported a positive association between O3 and emergency
department visits for respiratory outcomes, O3 effects were found to be robust to adjustment
for PM10, NO2, SO2, and BS (Lin et al., 1999; Peel et al., 2005; Tenias et al., 1998). One study
by Tolbert and colleagues (2000) observed that the significant univariate effects of both O3
and PM10 on pediatric asthma emergency department visits in Atlanta, GA became
nonsignificant in two-pollutant regressions, reflecting the high correlation between the two
pollutants (r = 0.75).
For several other studies with total respiratory and asthma outcomes, inconsistencies
confound the interpretation of likely effects. For example, in a Montreal, Canada study by
Delfino et al. (1997b), O3 effects on total respiratory emergency department visits were seen in a
short data series from the summer of 1993 (mean 8-h max O3 of 30.7 ppb [SD 11.5]) but not in a
similar data series from the summer of 1992 (mean 8-h max O3 of 28.8 ppb [SD 11.3]). The
significant 1993 results were seen only for persons older than 64 years. A very similar analysis
of two additional summers (1989 [mean 8-h max O3 of 44.1 ppb] and 1990 [mean 8-h max O3
of 35.4 ppb]) revealed an O3 association only for 1989, but again only in persons over 64 years
old (Delfino et al., 1998b). An analysis of data on respiratory emergency department visits
from June to August of 1990 (mean 24-h avg O3 of 28.2 ppb [SD 11.7]) in Baton Rouge, LA
reported O3 effects in adults, but not in children or among the elderly (Jones et al., 1995).
7-69
-------
Tobias and colleagues (1999) showed that regression results for asthma emergency
department visits could be quite sensitive to methods used to control for asthma epidemics.
Ozone was associated with the outcome variable in only one of eight models tested. An Atlanta,
GA study by Zhu et al. (2003) examined asthma emergency department visits of children during
three summers using Bayesian hierarchical modeling to address model variability. Data were
analyzed at the ZIP code level to account for spatially misaligned longitudinal data. Results
indicated a positive, but nonsignificant relationship between O3 and emergency room visits
for asthma. Ozone levels were not reported in this study.
Other studies also reported no association between O3 and emergency department visits for
respiratory causes (Atkinson et al., 1999a; Castellsague et al., 1995; Chew et al., 1999; Hwang
and Chan, 2002; Sunyer et al., 1997). Using Bayesian hierarchical modeling, Hwang and Chan
(2002) examined the effect of air pollutants on daily clinic visits for lower respiratory illnesses
across 50 communities in Taiwan. The mean 1-h max O3 for all 50 communities was 54.2 ppb,
with individual-community means ranging between 38.9 to 78.3 ppb. All pollutants except O3
were associated with daily clinic visits. In a pooled analysis of emergency admissions for
asthma in four European cities as part of the Air Pollution on Health: European Approach
(APHEA) study, no overall effect of O3 was observed (Sunyer et al., 1997). Median 1-h max O3
levels were relatively low, ranging from 14 to 37 ppb across the four cities. Atkinson et al.
(1999a) also did not find an association in London, England between O3 and emergency
department visits at a mean 8-h max O3 of 17.5 ppb (SD 11.5). One study by Thompson et al.
(2001) in Belfast, Northern Ireland found no O3 effect in the warm season, but a decreased risk
of childhood asthma admissions (-21% [95% CI: -33, -6] per 20 ppb increase in 24-h avg O3)
in the cold season. The O3 levels were similar in both seasons, with mean 24-h avg O3
concentrations of 18.7 ppb in the warm season and 17.1 ppb in the cold season. After adjusting
for benzene levels, O3 was no longer associated with asthma emergency department visits during
the cold season. The inverse relationship of O3 with benzene concentrations (r = -0.65), and
perhaps with other pollutants, might have contributed to the apparent protective effect of O3.
A study by Hajat et al. (1999, 2002) of physician consultations for asthma, lower respiratory
diseases, and upper respiratory diseases in London reported negative associations with O3, which
was also suggestive of residual confounding by copollutants or weather factors (note that data
were analyzed using Poisson GAM with default convergence criteria). Several other emergency
7-70
-------
department studies looking at O3 are more difficult to interpret because of inadequate control for
seasonal patterns, very low O3 levels, or because no quantitative results were shown for O3
(Buchdahl et al., 1996, 2000; Garty et al., 1998; Holmen et al., 1997; Lierl and Hornung, 2003;
Lipsett et al., 1997; Nutman et al., 1998).
Although several studies found a significant association between O3 concentrations and
emergency department visits for respiratory causes, some inconsistencies were observed. The
inconsistencies may be attributable, at least partially, to differences in model specifications and
analysis approach among the various studies. For example, ambient O3 concentrations, length of
the study period, and statistical methods used to control confounding by seasonal patterns and
copollutants appear to affect the observed O3 effect on emergency department visits. Studies
that stratified analyses by season generally reported a positive association between O3
concentrations and emergency department visits for asthma in the warm season.
7.3.3 Studies of Hospital Admissions for Respiratory Diseases
Hospital admissions represent a medical response to a serious degree of morbidity for a
particular disease. Scheduled hospitalizations are planned in advance when a particular clinical
treatment is needed. However, unscheduled admissions are ones that occur in response to
unanticipated disease exacerbations and are more likely to be affected by environmental factors,
such as air pollution. As such, the hospital admissions studies reviewed here focused
specifically on unscheduled admissions. Study details and results from hospital admissions
studies published during the past decade are summarized in Table AX7-4 (in Annex 7,
Section AX7.1). As a group, these hospitalization studies tend to be larger, in terms of
geographic and temporal coverage, and indicate results that are generally more consistent than
those reviewed above for emergency department visits. As in the case for all studies that
examine changes in aggregate measures of acute disease outcomes over time, the following
should be considered in comparing results: (1) difference in types of respiratory diseases for
hospital admission; (2) age of study population; (3) mean level of O3 during study; (4) single-
city versus multicity studies; (5) length of study (e.g., <5 years versus >5 years); (6) analysis by
season versus all year; (7) O3-only versus multipollutant models; (8) number of exposure lag
days; and (9) type of study (e.g., case-crossover versus time-series). These factors are
7-71
-------
considered in the sections below with further discussion on potential confounding of the O3
effect estimate by seasonal factors and copollutants.
7.3.3.1 All-year and Seasonal Effects of Ozone on Respiratory Hospitalizations
The effect of O3 on respiratory hospitalizations was examined in various studies conducted
in the United States and abroad. Figures 7-9 and 7-10 present risk estimates from all total
respiratory hospital admission studies. Burnett et al. (1995), which did not present quantitative
results for O3, and Yang et al. (2003), which only presented odds ratios, were not included in the
figures. In cases where multiple lags were presented, the multiday lag was selected to represent
the cumulative effect from all days examined. If only single-day lags were analyzed, the effect
estimate of the shortest lag time, usually a lag of 0 or 1 day, was presented. Figure 7-9 plots the
effect estimates and 95% CIs from 15 studies that analyzed all-year data. The risk estimates are
arranged by age groups. The preponderance of positive risk estimates, with some that are
statistically significant, is readily apparent. Figure 7-10 presents the season-stratified effect
estimates by region. For studies that reported risk estimates from all four seasons, only the
summer and winter estimates are presented. It appears that the warm-season estimates,
collectively, tend to be larger, positive values compared to all-year and cool-season estimates.
All of the negative estimates were from analyses using cool-season data only, which might
reflect the inverse correlation between O3 and one or another copollutant, especially PM, during
that season. None of the studies presented results for the O3 effect during the cool season after
adjusting for potential confounding by PM.
Among the respiratory hospitalization studies, the most robust and informative results were
observed when a broad geographic area was examined using a consistent analytical methodology
(Anderson et al., 1997; Burnett et al., 1995, 1997a). These studies have all reported an O3 effect
on respiratory hospital admissions. The largest such study to date was carried out using data on
all-age respiratory hospital admissions from 16 Canadian cities with populations exceeding
100,000 during the period 1981 to 1991 (Burnett et al., 1997a). In addition to O3, the authors
evaluated health effects of SO2, NO2, CO, and coefficient of haze (a surrogate for black carbon
particle concentrations). Pooling the 16 cities, a positive association was observed between
respiratory hospital admissions and the 1-day lag O3 concentration in the spring (5.6% [95% CI:
1.6, 9.9] excess risk per 40 ppb increase in 1-h max O3) and summer (6.7% [95% CI: 3.5, 10.0]).
7-72
-------
Burnett et al. (2001 ):Toronto,Canada (age <2) -
Atkinson etal,(1999b): London, England —
Ponce de Leon et al. (1996): London, England -
Braga et al. (1999): Sao Paulo, Brazil (age 0-12) -
Gouveia and Fletcher (2000a): Sao Paulo, Brazil (age <5) -
Petroeschevsky et al. (2001): Brisbane, Australia (age 0-4) -
(age 5-14)
Wongetai,(1999a): Hong Kong (age0-4) -
Linn et al. (2000): Los Angeles, CA (age 30+)
Atkinson etal.(1999b): London, England —
Ponce de Leon et al. (1996): London, England
Prescottetal.(1998):Edinburgh,Scotland(age<65) -
Schouten et al. (1996): Amsterdam, the Netherlands —
Schouten et al. (1996): Rotterdam, the Netherlands -
Petroeschevsky et al, (2001): Brisbane, Australia -
Wong et al. (1999a): Hong Kong (age 5-64) -
Atkinson etal.(1999b): London, England
Ponce de Leon et al. (1996): London, England -
Prescott et al. (1998): Edinburgh, Scotland -
Schouten et al, (1996): Amsterdam, the Netherlands -
Petroesche¥sky et al. (2Q01):Brisbane, Australia -
Wong et al. (1999a): Hong Kong -
Gwynn and Thurston (2001): New York City (whites only) —
(non-whites only) -
Gwynn et al. (2000): Buffalo, NY -
Luginaah et al. (2005): Windsor, Canada (males only) -
(females only)
Atkinson etal.(1999b): London, England -
Ponce de Leon et al. (1996): London, England —
Hagen et al. (2000): Drammen, Norway -
Oftedal et al. (2003):Drammen, Norway
Petroeschevsky et al. (2001):8risbane, Australia -
Wong et al. (1999a): Hong Kong -
% Change in Respiratory Hospitalization
-20
I
-10
I
10
I
20
30
I
lagO
Age 15-64 Years
Age 65+ Years
All Ages
lag 0-4
lagO
- lagO
lag 0-3
lagO
ag3
— •
lag 0-2
lag 0-3
lag 0-2
lag 0-5
lag 2
lag 0-3
• lagO
- • - lag 0-2
lag 0-3
lag 0-1
Iag3-
lag 0-3
Iag1
lagt
Iag1
Iag1
Iag1
lag 0-3
Figure 7-9. Ozone-associated percent change (95% CI) in total respiratory hospitalizations
for all-year analyses per standardized increment (see Section 7.1.3.2). Effect
estimates are arranged by age groups.
7-73
-------
% Change in Respiratory Hospitalizations
-20
Linn etal. (2000):
Los Angeles, CA (age 30+)
Schwartz etal. (1996):
Cleveland, OH (age 65+)
Burnett etal.(1997a):
16 Canadian Cities (all ages)
Burnett etal. (1997b):
Toronto, Canada (all ages)
Burnett etal. (2001):
Toronto, Canada (age < 2)
Poncede Leon etal, (1996):
London, England (all ages)
Schouten etal. (1996):
Amsterdam and Rotterdam,
the Netherlands (all ages)
Wongetal.{1999a):
Hong Kong (all ages)
0
i
20
I
40
i
60
U.S. and Canada |
-X-
-e-
Europe
-e
-e-
lagO
lag 1-2
Iag1
lag 1-3
lag 0-4
lag 0-2
lag 2
-e-
lag 0-3
Figure 7-10. Ozone-associated percent change (95% CI) in total respiratory
hospitalizations by season per standardized increment (see Section 7.1.3.2).
The results for fall were also positive, though of smaller magnitude (3.8% [95% CI: -0.2, 7.9]).
There was no evidence for an O3 effect in the winter season (-0.8% [95% CI: -4.8, 3.3]). The
mean 1-h max O3 concentrations for the spring, summer, fall, and winter were 40 ppb, 38 ppb,
21 ppb, and 26 ppb, respectively. Control outcomes related to blood, nervous system, digestive
system, and genitourinary system disorders were not associated with O3.
7-74
-------
In a previous study focused mainly on evaluating health impacts of sulfate particles,
Burnett et al. (1995) reported results from a time-series analysis of all-age respiratory hospital
admissions to 168 hospitals in Ontario, Canada over a 6-year period (1983 to 1988). The mean
1-h max O3 was 36.3 ppb for the entire year, ranging from 21.9 ppb in December to 52.9 ppb in
July. The outcome data were prefiltered to remove seasonal variations using a weighted 19-day
moving average. The authors reported that O3 was associated with respiratory hospital
admissions; however, no quantitative results for O3 were presented.
Results from an analysis of five European cities indicated strong and consistent O3 effects
on unscheduled hospital admissions for COPD (Anderson et al., 1997). The five cities examined
(London, Paris, Amsterdam, Rotterdam, and Barcelona) were among those included in the
multicity APHEA study. The number of years of available data varied from 5 to 13 years among
the cities. City-specific effect estimates were pooled across cities using weighted means.
An association with O3 was observed in full-year analyses. All-year median 1-h max O3 levels
ranged from 19 to 40 ppb across the five cities. Season-stratified analyses indicated that the O3
effect was larger in the warm season (median 1-h max O3 range 25-47 ppb), 4.7% (95% CI: 1.6,
7.9) excess risk per 40 ppb increase in 1-h max O3, compared to the cool season (median 1-h
max O3 range 10-33 ppb), 1.6% (95% CI: -3.1, 7.9) excess risk. There was no significant
heterogeneity in O3 effects among the cities.
Several additional studies carried out in one or two cities over a span of five or more years
provided substantial additional evidence regarding O3 effects on respiratory hospital admissions
(Anderson et al., 1998; Burnett et al., 1999, 2001; Moolgavkar et al., 1997; Petroeschevsky et al.,
2001; Ponce de Leon et al., 1996; Sheppard et al., 1999 [reanalysis Sheppard, 2003]; Yang et al.,
2003). Moolgavkar and colleagues (1997) reported significant and robust O3 effects on
respiratory hospital admissions in adults 65 years and older in Minneapolis and St. Paul, MN
(mean 24-h avg O3 of 26.2 ppb), but not in Birmingham, AL (mean 24-h avg O3 of 25.1 ppb).
The absence of effects in the southern city may reflect less penetration of O3 into the indoor
environment due to greater use of air conditioning, and thus less correlation between central
site O3 monitoring and actual exposures of the urban populace. Ozone effects on all-age and
age-stratified asthma and total respiratory hospital admissions were observed in Brisbane,
Australia (Petroeschevsky et al., 2001). Effect sizes were found to be consistent in the warm and
cool seasons (data not provided). Petroeschevsky et al. commented that the year-round effect
7-75
-------
of O3 might reflect the relatively small degree of seasonal variation in O3 levels observed in
Brisbane. Mean 8-h avg O3 (10 a.m.-6 p.m.) levels for the winter, spring, summer, and fall were
16.1 ppb, 23.3 ppb, 19.9 ppb, and 16.7 ppb, respectively. The authors also noted that given the
subtropical climate in Brisbane, characterized by warm, dry winters, perhaps the proportion of
the population exposed to winter O3 concentrations was higher than in cities where inclement
winter weather might force populations indoors.
Another set of studies examined associations between O3 and respiratory-related
hospitalizations in single cities over shorter (<5 year) time spans. Positive and significant O3
effects were reported in Cleveland, OH (Schwartz et al., 1996); New York City (Gwynn and
Thurston, 2001); Northern New Jersey (Weisel et al., 2002); Toronto, Canada (Burnett et al.,
1997b); Helsinki, Finland (Ponka and Virtanen, 1996); Sao Paulo, Brazil (Braga et al., 1999;
Gouveia and Fletcher, 2000a); and Hong Kong (Wong et al., 1999a). The Helsinki study by
Ponka and Virtanen (1996) reported significant effects of O3 on both asthma and on digestive
disorders in a setting of very low O3 concentrations (mean 8-h max O3 of 11 ppb), which raises
questions of plausibility.
Less consistent effects of O3 were seen in other respiratory hospitalization studies
(Schouten et al., 1996; Lin et al., 2003, 2004; Linn et al., 2000; Morgan et al., 1998a; Oftedal
et al., 2003). In a study conducted in Amsterdam and Rotterdam, the Netherlands, associations
between O3 and respiratory admissions were observed; however, results were difficult to
interpret due to the large number of statistical tests performed (Schouten et al., 1996). In a
California study by Neidell (2004), a negative association was observed between hospitalizations
for asthma and naturally occurring seasonal variations in O3 within ZIP codes for children aged
0 to 18 years. However, the O3 effect was found to be influenced by socioeconomic status.
Among children of low socioeconomic status, O3 generally was associated with increased
hospitalizations, with statistical significance reached in certain age groups. Neidell further stated
that avoidance behavior on high O3 days (i.e., 1-h max O3 >200 ppb) may have attributed to the
negative relationship observed in children of higher socioeconomic status.
No associations between respiratory hospital admissions and O3 were seen in studies from
Los Angeles, CA (Nauenberg and Basu, 1999); Detroit, MI (Lippman et al., 2000 [reanalysis Ito,
2004]); Vancouver, Canada (Lin et al., 2004); London, England (Atkinson et al., 1999b);
Edinburgh, Scotland (Prescott et al., 1998); and Drammen, Norway (Hagen et al., 2000).
7-76
-------
Several of these studies were carried out in locations with low O3 levels, suggestive of a
nonlinear concentration-response relationship. For example, the mean 1-h max O3 was
28.02 ppb in the Vancouver, Canada study by Lin et al. (2004). The Edinburgh, Scotland study
by Prescott et al. (1998) had a mean 24-h avg O3 of 14.5 ppb. Inadequate control of seasonal
confounding may underlie some of the nonsignificant and negative findings. An additional
factor likely contributing to lack of associations observed is the relatively small sample sizes
included in some of these studies.
For respiratory hospitalization outcomes, the largest, most significant associations with O3
concentrations were observed when using short lag periods, in particular a 0-day lag (exposure
on same day) and a 1-day lag (exposure on previous day). In the study of 16 Canadian cities by
Burnett et al. (1997a), the strongest association between O3 and respiratory hospitalizations was
found at a 1-day lag. A decline in the magnitude and significance of the effect was seen with
increasing days lagged for O3. Anderson et al. (1997) investigated the association between O3
and daily hospital admissions for COPD in five European cities. Lags up to 5 days were
examined, and the largest risk estimates were found using 0- and 1-day lags. These results
suggest that O3 has a short-term effect on respiratory hospitalizations.
Burnett et al. (2001) investigated the association between respiratory hospitalizations
and O3 in children less than 2 years of age (note that analyses were performed using default
convergence criteria for Poisson GAM with a nonparametric LOESS prefilter applied to
pollution and hospitalization data). Lags up to 5 days were examined after stratifying by season
(Figure 7-11). The mean 1-h max O3 during the summer season was 45.2 ppb (IQR 25). In the
summer, significant associations between O3 and daily admissions were found in several of the
single-day lags, with the largest risk estimate of 12.5% (95% CI: 5.7, 19.7) excess risk per
40 ppb increase in 1-h max O3 at a 1-day lag. The positive effect estimates observed at multiple
single-day lags indicated that O3 exposure likely had an immediate effect that persisted over
several days. Using a cumulative lag period of 0- to 4-days, the O3-related risk estimate was
30.2% (95% CI: 18.0,42.4).
Weisel et al. (2002) stated that a lag period of 1 to 3 days between exposure to O3 and
hospital admissions or emergency department visits for asthma was plausible, because it might
take time for the disease to progress to the most serious responses following exposure. Also,
taking medication could further delay the progression of the adverse effect. Thus, although
7-77
-------
-------
7.3.3.2 Potential Confounding of the Ozone Effect on Respiratory Hospitalizations
by Copollutants
As in the case for most air pollution studies, potential confounding of the association
between O3 and respiratory hospitalizations by copollutants generally was examined using
multipollutant regression models. The changing relationship between O3 and copollutants by
season complicates assessment of potential confounding of the O3 effect by copollutants.
Figure 7-12 compares the risk estimates from models with and without adjustment for PM
indices. This figure indicates that O3 risk estimates are fairly robust to PM adjustment in all-year
and warm-season only data. Given the inverse relationship typically observed between O3 and
PM during the cool season, the influence of PM adjustment on O3 risk estimates during that
season is of particular interest. However, none of the hospitalization studies examined O3 risk
estimates after adjusting for PM in cool-season only data.
Several analyses of a large data set from Toronto, Canada spanning the years 1980 to 1994
reported O3 effects on respiratory hospitalizations for all ages (Burnett et al., 1997b, 1999) and
for persons less than 2 years old (Burnett et al., 2001). In the 1999 and 2001 studies, analyses
were performed using Poisson GAM (default convergence criteria) with a nonparametric LOESS
prefilter applied to the pollution and hospitalization data. All studies found that O3 effects were
robust when adjusting for PM indices, whereas PM effects from single-pollutant models were
markedly attenuated when O3 was added to the regression. These results imply more robust
associations with respiratory hospitalizations for O3 than PM.
Results from the APHEA study indicated strong and consistent O3 effects on unscheduled
hospital admissions for COPD (Anderson et al., 1997). Significant effects also were seen for
BS, TSP, and NO2. The authors reported that among all pollutants examined, the most consistent
and significant findings were for O3. No two-pollutant model results were reported. Several
additional studies also observed that there was no substantial difference in the O3 effect after
adjusting for PM in the regression model (Gouveia and Fletcher, 2000a; Petroeschevsky et al.,
2001; Ponce de Leon et al., 1996).
Collectively, these results suggest that copollutants generally do not confound the
association between O3 and respiratory hospitalizations. Ozone risk estimates were robust to PM
adjustment in all-year and warm-season only data.
7-79
-------
% Change in Respiratory Hospitalization
Burnett etal.(1997b):
Toronto, Canada
(all ages, warm season only)—
with PM2 5
with PM10
with PM10_25
Burnett etal. (2001):
Toronto, Canada
(age <2,warm season only) —
with PM25 -
with PM10_25
Ponce de Leon etal. (1996):
London, England (all ages) -
with BS-
Hagen etal.(2000):
Drammen, Norway (all ages) -
withPMln-
Gouveia and Fletcher (2000a):
Sao Paulo, Brazil (age <5) -
withPM10-
Petroeschevsky etal. (2001):
Brisbane, Australia (all ages) -
with nephelometer
Wongetal.(1999a):
Hong Kong (all ages) -
20 -10 0 10 20 30 40 5
i i i i i i
Canada |
Europe
O03only
Q
• , • Oi with PM
laq 1-3 w 3 vvlul rm
f\
- lag 1
^>
^ A lag 0
Latin America
£ lagO
Australia
| Asia |
PI
° lag 2
V
Pi
^ ^ lag 0-3
•
Figure 7-12. Ozone-associated percent change (95% CI) in total respiratory
hospitalizations with adjustment for PM indices per standardized
increment (see Section 7.1.3.2). Analyses performed using all-year
data unless noted otherwise.
7.3.4 Association of Ozone with Hospital Admissions for
Cardiovascular Disease
Some hospital admissions studies have examined the association of O3 with cardiovascular
outcomes (see Figure 7-13). Many reported negative or inconsistent associations (Ballester
et al., 2001; Burnett et al., 1999; Fung et al., 2005; Koken et al., 2003; Linn et al., 2000;
7-80
-------
-30
% Change in Cardiovascular Hospitalizations
-20 -10 0 10 20 30 40 50
60
Linn et al. (2000):
Los Angeles, CA (age 30+)
Burnett etal. (1997b):
Toronto, Canada
Fung et al. (2005):
Windsor, Canada (age < 65)
(age > 65)
Atkinson etal. (1999b):
London, England
Poloniecki etal, (1997):
London, England
Prescott etal. (1998):
Edinburgh, Scotland (age 0-65)
(age 65+)
Ballester etal. (2001):
Valencia, Spain
Morgan etal. (1998a):
Sydney, Australia
Petroeschevskyetal. (2001):
Brisbane, Australia
Chang etal. (2005):
Taipei, Taiwan
Yangetal.(2004a):
Kaohsiung, Taiwan
Wongetal.(1999a):
Hong Kong
Wong etal. (1999b):
Hong Kong
i
i
i
i
i
U.S. and Canada |
Australia
-X-! lagO
-•- Ia9°
)— : lag 0
lag 2-4
lag 0-2
lag 0-2
lag 2
lag 1-3
lag 1-3
lag 2
lagO
lag 3
-e-
lag 0-2
i
lag 0-2
lag 0-2
-e-
lag 0-2
lag 0-5
-0 lag 0-5
- lag 0-1
- lag 0-1
-9 lag 0-1
Figure 7-13. Ozone-associated percent change (95% CI) in total cardiovascular
hospitalizations per standardized increment (see Section 7.1.3.2).
Analyses include all ages unless otherwise noted.
7-81
-------
Lippmann et al., 2000 [reanalysis Ito, 2004]; Mann et al., 2002; Morgan et al., 1998a;
Petroeschevsky et al., 2001; Poloniecki et al., 1997; Prescott et al., 1998). In the Ito (2004)
reanalysis of the Lippmann et al. (2000) study in Detroit, MI, no associations were observed
between cause-specific cardiovascular admissions (i.e., ischemic heart disease, dysrhythmias,
heart failure, and stroke) and O3 concentrations (mean 24-h avg O3 of 25 ppb) using all-year
data. The study by Linn et al. (2000) examined associations between O3 concentrations and
cardiovascular admissions during all four seasons as well as using all-year data in Los Angeles,
CA. Mean 24-h avg O3 levels for the spring, summer, fall, and winter were 32 ppb, 36 ppb,
13 ppb, and 14 ppb, respectively. Positive, but nonsignificant effects were observed during the
spring and summer (0.6% [95% CI: -1.4, 2.6] and 0.2% [95% CI: -1.7, 2.2] respectively, per
20 ppb increase in 24-h avg O3), while negative associations were observed during the fall and
winter (-0.6% [95% CI: -3.3, 2.2] and -4.1% [95% CI: -7.1, -1.1] respectively).
Some other studies, especially those that examined the relationship when O3 exposures
were higher, have found robust positive associations between O3 and cardiovascular hospital
admissions (Atkinson et al., 1999b; Burnett et al., 1997b; Chang et al., 2005; Tsai et al., 2003a;
Wong et al., 1999a,b; Yang et al., 2004a). For example, Burnett et al. (1997b) reported a
positive association between O3 and cardiovascular hospital admissions in Toronto, Canada in
a summer-only analysis (mean 1-h max O3 of 41.2 ppb). The results were robust to adjustment
for various PM indices, whereas the PM effects diminished when adjusting for gaseous
pollutants. Other studies stratified their analysis by temperature, i.e., by warms days (>20 °C)
versus cool days (<20 °C). Several analyses using warms days consistently produced positive
associations (Chang et al., 2005; Tsai et al., 2003a; Yang et al., 2004a). On the other hand, in
two studies conducted in Hong Kong, total cardiovascular (as well as circulatory, ischemic heart
disease, and heart failure) were all reported to be significantly associated with O3 in the cool but
not the warm season (Wong et al., 1999a,b). In Wong et al. (1999b), O3 concentrations were
similar in both seasons, with warm-season O3 levels being slightly lower (mean 8-h avg O3 =
16.1 ppb) than the cool season levels (mean O3 = 18.0 ppb). The authors speculated that
differing activity patterns and home ventilation factors may have contributed to the seasonal
differences in O3 effects. Weather in Hong Kong is mild throughout the year, but less humid and
cloudy in the cool season. Thus, during the cool season people are more likely to open windows
7-82
-------
or stay outdoors, resulting in higher personal exposures even with similar ambient O3
concentrations across seasons.
An increasing number of hospitalization studies have examined the association between O3
and cardiovascular admissions. Some studies, especially those that performed stratified analyses
by seasonal or meteorological factors, have observed positive associations. However, the overall
evidence remains inconclusive regarding the effect of O3 on cardiovascular hospitalizations.
7.3.5 Summary of Acute Ozone Effects on Daily Emergency Department
Visits and Hospital Admissions
• The vast majority of emergency room visits and hospitalization studies conducted
over the past decade have looked at effects of O3 on either total respiratory diseases
and/or asthma. Many of these studies analyzed O3 risk estimates using year-round
data. Given the strong seasonal variations in O3 concentrations and the changing
relationship between O3 and other copollutants by season, inadequate adjustment for
seasonal effects might have masked or underestimated the association between O3
and the respiratory disease outcomes. Ozone was generally found to be associated
with respiratory hospitalizations and asthma emergency department visits during the
warm season but not during the cool season.
• Several studies have examined the association between O3 and respiratory
hospitalizations while controlling for other pollutants in the analytical model.
In most cases, O3 effects have been reported to be robust to adjustment for
copollutants, particularly PM. Therefore, the evidence is supportive of independent
O3 effects on respiratory hospital admissions.
• Some hospital admission studies examined the effect of O3 on cardiovascular
outcomes. A few studies observed positive O3 associations, largely in the warm
season. Overall, however, the currently available evidence is inconclusive regarding
any association between ambient O3 exposure and cardiovascular hospitalizations.
7.4 ACUTE EFFECTS OF OZONE ON MORTALITY
7.4.1 Summary of Key Findings on Acute Effects of Ozone on Mortality
from the 1996 Ozone AQCD
A limited number of studies, most of which were from the 1950s and 1960s, had
examined O3-mortality associations at the time of the 1996 O3 AQCD. The 1996 O3 AQCD
considered these historical studies to be flawed because of inadequate adjustment for seasonal
trends or temperature and the use of questionable exposure indices. There were only a few other
7-83
-------
time-series studies that had examined O3-mortality associations between the 1980s and mid-
1990s. These latter studies used more sophisticated approaches in addressing seasonal
confounding and weather models. One of them (Shumway et al., 1988) focused on possible
associations with long-term O3 fluctuations in Los Angeles, CA but did not examine short-term
associations. A study that reanalyzed Los Angeles, CA data with a focus on short-term
associations (Kinney and Ozkaynak, 1991) did find that, of the PM and gaseous criteria
pollutants, O3 (reported as total oxidants) was most strongly associated with total nonaccidental
mortality. Then two more studies, one conducted in Detroit, MI (O3 concentrations not
provided) (Schwartz, 1991) and the other in St. Louis, MO (mean 24-h avg O3 of 22.5 ppb) and
Kingston-Harriman, TN (mean 24-h avg O3 of 23.0 ppb) (Dockery et al., 1992), reported that
PM but not O3 was significantly associated with mortality. However, the 1996 O3 AQCD noted
that, without sufficient presentation of model specifications, it was difficult to evaluate whether
the lack of O3-mortality associations was possibly due to mis-specification of the weather model.
In summary, because of the insufficient number of studies that examined O3-mortality
associations and the uncertainties regarding weather model specifications, the 1996 O3 AQCD
was unable to quantitatively assess O3-mortality excess risk estimates or, even, to provide
qualitative assessment of the likelihood of O3-mortality associations.
7.4.2 Introduction to Assessment of Current Ozone-Mortality Studies
Introductory discussions of PM-mortality effects often cite historical air pollution incidents
such as the 1952 London, England smog episode, in which thousands of deaths were attributed
to the air pollution from coal burning. There is no counterpart "historical episode" for O3-
mortality effects. Instead, the early recognition of the adverse health effects of summer oxidant
air pollution, mainly from Los Angeles and other major cities with a high density of
automobiles, were based on symptoms such as eye and throat irritations. Thus, the focus of PM
epidemiology and that of O3 epidemiology have been historically different.
As shown in Table AX7-5 in Annex 7, Section AX7.1, the number of short-term mortality
studies that analyzed O3 has increased markedly since the last publication of the 1996 O3 AQCD.
The increased attention to PM-mortality associations in the early 1990s led to an increase in
studies that also examined O3, most often as a potential confounder for PM. Although many of
these PM studies also reported O3 estimates, they often lacked specific hypotheses regarding
7-84
-------
mortality effects of O3, given that the main focus of these studies was to examine PM-mortality
associations. This is in contrast to the O3-morbidity studies, most of which were specifically
designed to examine effects of "summer haze" and O3 (or oxidants) on respiratory and other
symptoms, lung function, emergency department visits, etc. However, new studies with
hypotheses developed specifically for O3 effects on mortality have become available, such as the
large U.S. 95 communities study by Bell et al. (2004), the U.S. 14 cities study by Schwartz
(2005), and the 23 European cities study by Gryparis et al. (2004) discussed in the next section.
7.4.3 Single-Pollutant Model Ozone-Mortality Risk Estimates
To facilitate a quantitative overview of the O3-mortality effect estimates and their
corresponding uncertainties, the percent excess risks of total nonaccidental mortality calculated
using all-year data are plotted in Figures 7-14 and 7-15. Studies that only conducted seasonal
analyses are presented in the next section. These figures do not include studies that only
examined cause-specific mortality. Figure 7-14 presents only those results derived from single-
day lag models. Results from multiday lag models are shown in Figure 7-15. All effect
estimates are from single-pollutant models and include all age groups unless otherwise noted.
The majority of the estimates are positive, with a few exceptions. Of particular note, five
multicity studies, three from the United States (Bell et al., 2004; Samet et al., 2000 [reanalysis
Dominici et al., 2003]; Schwartz, 2005) and two from Europe (Gryparis et al., 2004; Touloumi
et al., 1997), showed generally positive associations.
The initial primary objective of the original NMMAPS (Samet et al., 2000; reanalysis
Dominici et al., 2003) was to investigate the effects of PM, but the study also examined
mortality risk estimates from gaseous pollutants in 90 U.S. cities over the period of 1987 to
1994. Among the 90 cities, 80 monitored O3 either year-round or during the warm season.
The study illustrated that the mortality risk estimates for O3 varied by season. The estimate
using all available data was about half of that for summer-only data at a lag of 1-day (see
Section 7.6.3.2 for further discussion). Bell et al. (2004) extended the original NMMAPS
by adding six more years (from 1987 to 2000) and 15 more communities (a total of
95 communities), and examined the effects of O3 on mortality. Due to its extensive coverage
and its specific focus on O3- mortality effects, Bell et al. (2004) offer a more comprehensive
analysis than the original Samet et al. (2000) analysis.
7-85
-------
Bell etal.(2004): U.S.95 communities-
Samet et al.(2000; reanalysis Dominici et al.,2003): U.S.90 cities-
Schwartz (2005): U.S.14dties-
Kinney and Ozkaynak (1991): Los Angeles County, CA-
Kinney et al. (1995): Los Angeles County, CA-
Fairley (2003): Santa Clara County, CA-
Dockeryetal.(1992): St. Louis, MO-
Lippmann et al. (2000; reanalysis Ito, 2004): Detroit, Mi-
Chock et al. (2000): Pittsburgh, PA (age 0-74)-
Dockery et al. (1992): Eastern Tennessee-
Villeneuveetal.(2003): Vancouver,Canada (age 65+)-
Touloumi et al. (1997): 4 European cities (Athens, Barcelona, London, Paris) -
Anderson et al. (1996): London, England-
Bremner et al. (1999): London, England-
Prescottetal.(1998): Edinburgh, Scotland-
Zmirouetal.(1996): Lyon.France-
Hoek et al. (2000;reanalysis Hoek, 2003): The Netherlands-
Roemer and van Wijnen (2001): Amsterdam,the Netherlands-
Verhoeffetal.(1996): Amsterdam,the Netherlands-
Peters et al. (2000b): Coal basin in Czech Republic-
Peters et al. (2000b): NE Bavaria, Germany-
Garcia-Aymerich et al.(2000): Barcelona, Spain-
Sunyeretal.(1996): Barcelona,Spain-
Borja-Aburto et al. (1997): Mexico City—
Borja-Aburto et al. (1998): SW Mexico City-
Gouveia and Fletcher (2000b): Sao Paulo, Brazil-
Ostro et al. (1996): Santiago, Chile-
Morgan etal.(1998b): Sydney, Australia-
Simpson et al.(1997): Brisbane, Australia-
Kim etal. (2004): Seoul, Korea-
Lee etal. (1999): Seoul, Korea-
Lee and Schwa rtz (1999): Seou I, Korea —
Lee etal. (1999): Ulsan.Korea-
% Change in Mortality
-10
i
0
I
10
I
20
I
U.S. and Canada
•Iag1
Australia
(Q) Multicity combined
• Single city
ID lagO
© lagO
0 lagO
«• Iag1
«- Iag1
lagO
— Iag1
lagO
- lagO
Iag1
Best of lag 0 to 3 by city
• lagO
- lag 2
lagO
lagO
Iag1
Iag1
lagO
- lagO
-• lagO
—• Iag5
—• lagO
lagO
- lagO
lagO
lagO
lagO
lagO
lagO
Figure 7-14. All cause (nonaccidental) O3 excess mortality risk estimates (95% CI)
for all year analyses per standardized increment (see Section 7.1.3.2).
Analyses include all ages unless otherwise noted. Only results from
single-day lag models are presented.
7-86
-------
Bel! etal. (2004): _
U.S, 95 communities
Gamble (1998): Dallas,TX-
Ito and Thurston (1996):.
Cook County, IL
Lippmann etal. (2000;
reanalysis Ito, 2004): Detroit, MI"
Lipfertetal.(2000a):
7 counties in PA and NJ
Klemm etal. (2004):
Atlanta, GA (age 65+)"
Vllleneuve etal. (2003):
Vancouver, Canada (age 65+)
Gryparisetal,(2004): _
23 European cities
Touloumi et al, (1997): 4 European cities
(Athens, Barcelona, London, Paris)"
Hoek et al. (2000; reanalysis Hoek,
2003): The Netherlands"
Roemer and van Wijnen (2001):
Amsterdam,the Netherlands
Borja-Aburto et al. (1998):
SW Mexico City
O'Neill et al. (2004): Mexico City.
% Change In Mortality
-10
i
Europe
0
l
10
I
20
lag 0-1 • SE not given; significant at p = 0,055
© lag 0-6
(Q> Multicity combined
• Single city
lag 1-2
lag 0-1
lag 0-3
lag 0-1
lag 0-2
lag 0-1
Best of cumulative lags
up to 5 days by city
lag 0-8
• lag 0-6
• lag 1-2
Figure 7-15. All cause (nonaccidental) O3 excess mortality risk estimates (95% CI)
for all year analyses per standardized increment (see Section 7.1.3.2).
Analyses include all ages unless otherwise noted. Only results from
multiday lag models are presented.
The results of the study by Bell et al. (2004) are discussed in detail in this document
because of the study's emphasis on U.S. data and the inclusion of 95 large communities across
the country, making this mortality study the most representative one of the U.S. population.
In addition, this study is one of the few that have focused specifically on O3 hypotheses testing
and investigated several important issues. Among the 95 communities examined in this study,
55 monitored O3 throughout the year and 32 monitored O3 only during the warm season
(generally April to October). Eight additional cities switched from warm-season only to year-
7-87
-------
round monitoring or year-round to warm-season only monitoring at some point during the study
period. Figure 7-16 presents the median 24-h avg O3 concentrations for the 95 communities
from 1987 to 2000. In the 55 communities that had all-year data available, the median 24-h avg
O3 concentrations ranged from 14.4 ppb in Newark, NJ to 37.3 ppb in Bakersfield, CA. In the
40 communities with warm-season only data, the median 24-h avg O3 levels ranged from
20.4 ppb in Portland, OR to 36.2 ppb in Memphis, TN. The range of median concentrations
from communities monitored all year overlap considerably with the range from communities
that only monitored during the warm season. This is expected, given that communities that have
higher O3 concentrations year-round are generally monitored throughout the year. The mean
24-h avg O3 concentration from all available data was approximately 26 ppb for the
95 communities.
Within-community results were first calculated, using single-day lags of 0, 1, 2, and
3 days, and a 7-day distributed lag in O3 exposure. The individual-community maximum
likelihood effect estimates for all cause mortality per 20 ppb increase in 24-h avg O3 from a
constrained 7-day distributed lag model are presented in Figure 7-17. The heterogeneity of the
effect estimates from the individual communities is partially attributable to differences in
pollution characteristics, the use of air conditioning, time-activity patterns, and socioeconomic
factors. A two-stage Bayesian hierarchical model was also used to determine a national average
effect estimate, taking into consideration community-to-community variation. Figure 7-18
presents the Bayesian community-specific and national average O3 risk estimates for all cause
mortality per 20 ppb increase in 24-h avg O3 with a constrained 7-day distributed lag. The
Bayesian community-specific estimates were shrunk to the national average estimate by a factor
that was inversely proportional to the heterogeneity of the community-specific relative rates.
Because of the random variation, as well as the smaller sample sizes within each city, particular
consideration is given to the Bayesian national average effect estimate.
In the U.S. 95 communities study, the largest risk estimate for O3-mortality was obtained
with a 0-day lag, followed by diminishing risk estimates with 1-, 2-, and 3-day lags (Figure 7-19,
upper panel). Ozone exposure at a 0-day lag was associated with a 0.50% (95% PI: 0.24, 0.78)
excess risk in mortality per 20 ppb increase in 24-h avg O3. The 7-day distributed lag model,
which examined the cumulative effect from the same day and six previous days, also is shown in
Figure 7-19 (lower panel). A cumulative excess mortality risk of 1.04% (95% PI: 0.54, 1.55)
-------
Newark _
Honolulu -
Des Molnes -
Oakland -
New York City -
Tacoma -
San Jose -
Boston -
Washington -
Chicago -
Jersey City -
Denver -
Philadelphia -
New Orleans -
Nashville -
Pittsburgh -
Houston -
Cedar Rapids -
Syracuse -
Los Angeles -
Kansas City, KS -
Rochester -
Baton Rouge -
San Antonio -
Buffalo -
Lafayette -
Colorado Springs -
Stockton -
Corpus Christ) -
Orlando -
St Petersburg -
Santa Ana /Anaheim -
Jacksonville -
Tampa -
Phoenix -
Wichita -
Sacramento -
Austin -
Miami -
El Paso -
Lake Charles -
Dallas / Fort Worth -
Modesto -
Tucson -
Little Rock -
Albuquerque -
Shreveport -
San Diego -
Oklahoma City -
Fresno -
Tulsa •
Las Vegas -
Riverside -
San Bernardino -
Bakersfield -
Portland -
Seattle -
Johnstown -
Omaha -
Detroit -
Louisville -
Birmingham -
Atlanta -
St. Louis -
Jackson -
Cleveland -
Cincinnati -
Lincoln -
Columbus, OH -
Mobile -
Providence -
Toledo -
Biddeford -
Grand Rapids -
Dayton -
Columbus, GA -
Arlington -
Milwaukee -
Coventry -
Kansas City, MO -
Huntsville -
Worcester -
Akron -
Madison -
Indianapolis -
Knoxville -
Baltimore -
Lexington -
Salt Lake City -
Spokane -
Fort Wayne -
Charlotte -
Muskegon -
Raleigh -
Memphis -
O3 Concentration (ppb)
10 20 30 40 50 60
I I I I I I
All Year
Figure 7-16. Median 24-h avg O3 concentrations (10th percentile to 90th percentile range)
for 95 U.S. communities (NMMAPS) from 1987 to 2000, arranged by O3
concentration. Results from all available data are presented. Fifty-five of
the 95 communities had year-round data. The remaining communities
mostly had data during the warm season only.
Source: Derived from iHAPSS (2005).
7-89
-------
% Change in Mortality
-2
Salt Lake City —
Lexington —
Denver —
Orlando —
Little Rock —
Birmingham —
Wichita —
Las Vegas —
Coventry —
Fort Wayne —
Omaha —
El Paso —
San Antonio —
Cedar Rapids —
Lafayette —
Spokane —
Atlanta —
Oklahoma City —
Grand Rapids —
Indianapolis —
Boston —
Sacramento —
Dayton —
Tucson —
Raleigh —
Muskeg on —
San Diego —
St. Petersburg —
Tacoma —
D3ion rtougc
t A /A T^-
an ia O1"-' ' p.v . j.
Los Angeles —
Miami —
et 7.5
Baltimore
rroano
Ik
c p jacKBon
Albuquerque —
A *•
Cleveland —
Portland —
Providence —
Houston —
Pittsburgh —
Tulsa —
Chicago —
Mobile —
Phoenix —
Detroit —
Washington —
Biddaford —
Modesto —
Dallas / Fort Worth —
Columbus, OH —
Rochester —
Columbus, GA —
Huntsville —
Memphis —
Worcester —
Nashville —
Philadelphia —
Louisville —
Toledo —
New York City —
Syracuse —
Lincoln —
Honolulu —
Shreveport —
Colorado Springs —
Knoxville —
Arlington —
Newark —
Cincinnati —
Madison —
0 -15 -10 -5 0 5 10 15
I I I I I I I
i
*, :
_ ;
*J :
^5 ; warm seas
• i
0 '
*— :
*_
* :
o
O :
o
O :
2
. o
: '
: g
: ~
: 0
; 2
; S
i *
: 0
'• fl
' n
; JL
: i
; ^
; *
i O
; ™
: _
I 0
: *_
• ~^
' _
20
I
All Year O
on Only •
Figure 7-17. Community-specific maximum likelihood estimates for the percent change
(95% CI) in daily mortality per 20 ppb increase in 24-h avg O3 in the
previous week using a constrained distributed lag model for 95 U.S.
communities (NMMAPS), arranged by size of the effect estimate. Results
from all available data are presented. Fifty-five of the 95 communities had
year-round data, marked by open circles (o). The cities with mostly
warm-season data are marked by closed circles (•).
Source: Derived from Bell (2006).
7-90
-------
% Change in Mortality
Denver
3an Antonio
i ?*i p^"*k
Little Rock
Birmingham —
ct D *n uC^°
Lafayette —
Wichita —
Salt Lake City —
Los Angeles —
Coventry —
Atlanta —
Oklahoma City —
Cedar Rapids —
Sacramento —
umana
t A /A h
a ma nna ; Ananoim
f if .P.
n M
inaiamipojis
urana i capias
Baton ItOUgG
Pu~_
?/
p AKron
Charlotte —
Spokane —
Muskegon —
Buffalo —
Seattle —
Milwaukee —
Bakersfield —
Stockton —
Jersey City —
Kansas City, KS —
Kansas City, MO —
Oakland —
Albuquerque —
Baltimore —
Fresno —
New Orleans —
Jackson —
Portland —
San Jose —
Biddeford —
Riverside —
Johnstown —
Austin —
San Bernardino —
Providence —
Cleveland —
Arlington —
Modesto —
Tulsa —
Mobile —
Humtsviile —
Columbus, OH —
Lincoln —
Washington —
Pittsburgh —
Toledo —
Nashville —
Detroit —
Madison —
Columbus, QA —
Houston —
Phoenix —
Knoxville —
Louisville —
Honolulu —
Rochester —
Memphis —
Worcester —
Colorado Springs —
Syracuse —
Chicago —
Dallas / Fort Worth —
Cincinnati —
Philadelphia —
Newark —
New York City —
3-2-101234
I I I I I I I
O »«»___ e»_
: *
! X
: X
: n
! 2
: "ri
: *
: J*
: X
: X
: y
: 5
: ~_
: ^
: *
; 2
n
; 9,
; ^
: 0
: •
: i
: —
: i
1 0
' 0
': n
1 0
! 9,
I "
I *^
I *^
5
l
All Year O
on Only •
Figure 7-18. Community-specific Bayesian estimates and national average for the percent
change (95% PI) in daily mortality per 20 ppb increase in 24-h avg O3 in the
previous week using a constrained distributed lag model for 95 U.S.
communities (NMMAPS), arranged by size of the effect estimate. Results
from all available data are presented. Fifty-five of the 95 communities had
year-round data, marked by open circles (o). The cities with mostly warm-
season data are marked by closed circles (•).
Source: Derived from Bell et al. (2004).
7-91
-------
5
4
* 3
<0
c
9
Q 2
1
0
3-day 2-day
0-day
Single-day Lags
-0.5
0.5 1
% Change in Mortality
1.5
S 3
(A
0> o
-0.5
0- to 6-day
0.5 1
% Change in Mortality
Multiday Lag
1.5
Figure 7-19. Percent changes in all cause mortality per 20 ppb increase in 24-h avg O3 in
all ages Single-day lags (0-, 1-, 2-, and 3-day) are shown in the upper panel.
The cumulative multiday lag (0- to 6-day) is shown in the lower panel.
Source: Derived from Bell et al. (2004).
per 20 ppb increase in 24-h avg O3 during the previous week was observed. In a related U.S.
study of the 19 largest cities by Huang et al. (2005), the O3 estimate for the summer season was
1.47% (95% PI: 0.54, 2.39) excess risk of cardiopulmonary mortality with current-day
exposure. Mean 24-h avg O3 levels ranged from 18 ppb in Oakland, CA to 56 ppb in
San Bernadino, CA. Smaller effects also were observed with 1- and 2-day lags of exposure.
The effect estimate for the 7-day distributed lag was 2.52% (95% PI: 0.94, 4.10) excess risk of
cardiopulmonary mortality. These findings suggest that the effect of O3 on mortality is
immediate, but also may persist over multiple days.
7-92
-------
The influence of higher O3 levels on the risk estimate also was evaluated in the U.S.
95 communities study. When the data were restricted to days with 24-h avg O3 levels less
than 60 ppb for the 1-day lag analysis, the national estimate did not substantially change
(0.30% [95% PI: 0.06, 0.54] per 20 ppb increase for days with levels below 60 ppb versus
0.36% [95% PI: 0.12, 0.61] for all days). These results suggest that the O3-mortality
associations may occur at 24-h avg O3 levels below 60 ppb.
Schwartz (2005) examined O3-mortality associations using data from 14 U.S. cities.
Median 1-h max O3 levels ranged from 35.1 ppb in Chicago, IL to 60.0 ppb in Provo, UT.
A case-crossover study design was used to compare the influence of adjustment methods for
temperature (regression splines of temperature versus matching case and control periods by
temperature). The risk estimate obtained by matching (0.92% [95% CI: 0.06, 1.80] per 40 ppb
increase in 1-h max O3 at a 0-day lag) was similar to that obtained with regression splines
(0.76% [95% CI: 0.13, 1.40]), suggesting that the O3-mortality risk estimates were not sensitive
to these adjustment methods for temperature.
The APHEA 1 project (Touloumi et al., 1997) reported a pooled random effects estimate of
4.5% (95% CI: 1.6, 7.7) per 40 ppb increase in 1-h max O3 using the best single-day lag model
results from four European cities (Athens, Barcelona, London, and Paris). Mean 1-h max O3
levels ranged from 21.2 ppb in London to 48.4 ppb in Athens. As an extension of the four
European cities study, researchers of the APHEA 2 project investigated the effect of O3 on total,
cardiovascular, and respiratory mortality in 23 cities throughout Europe (Gryparis et al., 2004).
Ozone data were available year-round in all 23 cities. A cumulative lag of 0 to 1 days was
hypothesized a priori. A two-stage hierarchical model, which accounted for statistical variance
and heterogeneity among cities, was used to estimate the pooled regression coefficients.
Because of substantial heterogeneity among cities, random effects regression models were
applied. The pooled effect estimate for the 23 European cities (0.23% [95% CI: -0.85, 1.95]
per 40 ppb increase in 1-h max O3 for all seasons) was positive but considerably smaller
compared to that obtained in the APHEA 1 study. Mean O3 concentrations for all-year data were
not presented. Gryparis et al. (2004) noted that there was a considerable seasonal difference in
the O3 effect on mortality; thus, the small effect for the all-year data might be attributable to
inadequate adjustment for confounding by season. This seasonal effect is discussed further in
the next section.
7-93
-------
Collectively, the single-pollutant model estimates from the single- and multiple-city
studies shown in Figures 7-14 and 7-15 suggest that an excess risk of total nonaccidental
mortality is associated with acute O3 concentrations. Despite the different analytical approaches
and alternative model specifications used in the various studies, overall, most of the positive
estimates range from 0.5 to 5% excess risk in mortality per standardized increment for all single-
pollutant, all-year analyses.
7.4.4 Meta-Analyses of Ozone-Mortality Risk Estimates
Several studies in recent years conducted meta-analyses of O3-mortality associations (Levy
et al., 2001; Stieb et al., 2002, 2003; Thurston and Ito, 2001; World Health Organization, 2004).
Figure 7-20 presents the combined O3 risk estimates from the various meta-analyses. Most of
these analyses included GAM studies using default convergence criteria except for the one by
Stieb et al. (2003), which compared effect estimates from GAM-affected studies to non-GAM
studies. All of these meta-analyses reported fairly consistent and positive combined estimates,
approximately 2% excess total nonaccidental mortality per standardized increment (see Section
7.1.3.2). However, most of these studies were not analytical in design, in that they did not
attempt to examine the source of heterogeneity, although one suggested an influence of weather
model specification (Thurston and Ito, 2001) and another reported evidence of publication bias
(World Health Organization, 2004) in the past literature. None of these studies address the issue
of season-specific estimates, therefore, interpreting these combined estimates requires caution.
Most recently, three research groups conducted independent meta-analyses of O3-mortality
associations (Bell et al., 2005; Ito et al., 2005; Levy et al., 2005) which attempted to evaluate the
source of heterogeneity using the most up-to-date literature database. These analyses have also
been systematically compared and discussed (Bates, 2005; Goodman, 2005). The all-season
combined point estimates per standardized increment from these three meta-analyses were
remarkably consistent: 1.75% (95% PI: 1.10, 2.37), 1.6% (95% CI: 1.1, 2.0), and 1.64% (95%
CI: 1.25, 2.03), for the Bell et al., Ito et al., and Levy et al. studies, respectively. All three
studies also indicated that the estimates were higher in warm seasons. Each of these studies is
briefly summarized below. Their findings related to specific issues are discussed later in the
corresponding sections.
7-94
-------
% Change in Mortality
Bell et al. (2005):
32 U.S. and non-U.S. studies
(41 estimates)
Ito et al. (2005):
38 U.S. and non-U.S. studies
(43 estimates)
Levyetal. (2001):
6 U.S. and European studies
(6 estimates, PM-adjusted)
Levy et al. (2005):
28 U.S., Canadian, and European studies
(48 estimates)
Stieb et al. (2003):
10 U.S. and non-U.S. studies
(10 estimates, non-GAM studies only)
Thurston and Ito (2001):
7 U.S. and European studies
(7 estimates, PM-adjusted)
World Health Organization (2004):
8 European studies
(15 estimates)
3
I
4
I
Figure 7-20. Combined all cause (nonaccidental) O3 excess mortality risk estimates
(95% CI) from recent meta-analyses per standardized increment
(see Section 7.1.3.2). Note that all meta-analyses, except Stieb et al. (2003),
included studies which used Poisson GAM with default convergence criteria.
Bell et al. (2005) conducted a meta-analysis of 144 effect estimates from 39 U.S. and
non-U.S. studies and estimated pooled effects by lags, age groups, specific causes, and exposure
metrics. The results were also compared with their NMMAPS results (Bell et al., 2004).
A two-stage Bayesian hierarchical model was used to estimate the combined estimate by taking
into account the within-city variance (the statistical uncertainty) and between-study variance
(the heterogeneity across cities). Bell et al. (2005) concluded that the results provided strong
evidence of a short-term association between O3 and mortality that was not sensitive to
adjustment for PM or for model specifications (discussed in Section 7.4.6). However, they
7-95
-------
suggested that, based on comparisons between the meta-analysis results and NMMAPS results,
there was evidence of publication bias (1.75% [95% CI: 1.10, 2.37] per 20 ppb increase in 24-h
avg O3 for meta-analysis versus 0.50% [95% CI: 0.24, 0.78] for NMMAPS 0-day lag results).
Ito et al. (2005) both conducted a meta-analysis of 43 U.S. and non-U.S. studies and also
analyzed data from 7 U.S. cities to further examine issues identified by their meta-analysis.
Adjusting for PM did not substantially influence the O3-mortality effect estimates in either the
meta-analysis or 7 U.S. cities analysis. The multicity analysis further indicated that the
difference in the weather adjustment model could result in a 2-fold difference in risk estimates
(e.g., 1.96% versus 0.96% in multicity combined estimates across alternative weather models for
the O3-only, all-year case). In the meta-analysis, they found suggestive evidence of publication
bias (a significant asymmetry in the funnel plot), but adjusting for the asymmetry reduced the
combined estimate only slightly (from 1.6% [95% CI: 1.1, 2.0] to 1.4% [95% CI: 0.9, 1.9] per
20 ppb increase in 24-h avg O3). The extent of potential bias implicated in this study differed
compared to that reported by Bell et al. (2005). The source of this difference is not clear, but Ito
et al. stated that sensitivity analyses comparing estimates from commonly used weather model
specifications suggest that the stringent weather model used in NMMAPS may tend to yield
smaller risk estimates than those used in other studies.
Levy et al. (2005) analyzed 48 estimates from 28 studies from the United States, Canada,
and Europe using an empiric Bayesian meta-regression with covariates including the relationship
between O3 and other pollutants, proxies for the relationship between personal exposure and
ambient concentration such as air conditioning prevalence, and statistical methods used. They
found that the air conditioning prevalence (a greater effect in cities with less air conditioning)
and lag time (same-day effects larger than lagged effects) were the strongest predictors of
between-study variability. The warm-season estimates were larger than the cool-season
estimates. The influences of copollutants were inconsistent, but a potential influence of
summertime PM2 5 was found.
As stated earlier, the combined O3 excess mortality risk estimates from the meta-analyses
by Bell et al. (2005), Ito et al. (2005), and Levy et al. (2005) were all very consistent. Although
the analyses were conducted independently, there was considerable overlap among the risk
estimates used in the three meta-analyses; thus, the agreement in the combined risk estimates
was not unexpected. The common findings among these three meta-analyses, aside from the
7-96
-------
consistency in their combined estimates, include: (1) no difference in estimates between GAM
studies using default versus stringent convergence criteria; (2) larger estimates in warm seasons;
and (3) no strong indication of PM confounding. Both the Bell et al. and Levy et al. analyses
found that the estimates at lag 0-day were larger than for longer lags. Both the Bell et al. and Ito
et al. studies also suggested evidence of publication bias. The positive O3 effect estimates, along
with the sensitivity analyses in these three meta-analyses provide evidence of a robust
association between ambient O3 and mortality. The combined effect estimates from the various
meta-analyses ranged from 1.5 to 2.5% excess risk in all-cause mortality.
7.4.5 Seasonal Variation in Ozone-Mortality Risk Estimates
Since the O3 seasonal cycle follows the seasonal cycle of temperature (which is inversely
related to the mortality seasonal cycle), inadequate adjustment of temporal trends in the
regression model could lead to negative O3-mortality risk estimates. In addition (as discussed in
Section 7.1.3.5) low-level O3 during the winter in some cities may be negatively correlated with
PM and other primary pollutants, resulting in negative correlations between O3 and mortality
even in short-term relationships. Such a confounding effect by season could be substantially
reduced by conducting season-stratified analyses.
A fewer number of O3-mortality studies performed seasonal analyses. Figure 7-21 presents
the studies that reported O3 risk estimates for all-cause mortality by season. For those studies
that obtained O3 risk estimates for each of the four seasons, only summer and winter results are
shown. The estimates for year-round data analyses, when available, also are shown for
comparisons. In all the studies, the O3 risk estimates are larger during the warm season than the
cool season, with the all-year estimates generally falling between the two seasonal estimates.
In three U.S. or European multicity studies (Gryparis et al., 2004; Samet et al., 2000
[reanalysis Dominici et al., 2003]; Schwartz, 2005), season-stratified analyses indicated that
the O3-mortality effect estimates were significant and positive in the warm season, with larger
effects observed compared to the year-round analyses. The effect estimates from the cool season
were notably smaller and less significant. In 14 U.S. cities, Schwartz (2005) observed an excess
mortality risk of 1.0% (95% CI: 0.3, 1.8) per 40 ppb increase in 1-h max O3 during the warm
season and 0% (95% CI: -1.1,1.1) during the cold season with a 0-day lag of exposure. Ozone
7-97
-------
% Change in Mortality
Bell et al. (2004): U.S. 95 communities -
Samet et al. (2000; reanalysis
Dominici et al. 2003): U.S. 90 cities"
Schwartz (2005): U.S. 14 cities -
Ostro (1995): San Bernardino County
and Riverside County,CA
Gamble (1998}:Dallas,TX-
Moolgavkar et al. (1995):Philadelphia, PA-
Chock et al. (2000): Pittsburgh, PA (age 0-74) -
Vedaletal. (2003): Vancouver, Canada-
Gryparisetal.(2004):21 European cities -
Anderson et al. (1996): London, England -
Hoek et al. (2000; reanalysis
Hoek,2003):The Netherlands"
Sunyer et al. (1996): Barcelona, Spain -
Borja-Aburto et al.(1997): Mexico City-
Cifuentesetal.(2000):Santiago,Chile-
Ostro et al. (1996): Santiago, Chile -
Simpson et al. (1997): Brisbane, Australia -
Kim et al. (2004): Seoul, Korea -
-10
I
10
I
20
I
30
I
U.S. and Canada
-e-
-X-
-e-
lag 1-2
lagO
Latin America
Australia
-e-
-X-
*
lag 0-6
Iag1
lagO
lagO
• SE not given; significant at p = 0.05 level
O SE not given; not significant at p = 0.05 level
lagO
lag 0-1
lagO
lagO
X
lagO
lag 1-2
lagO
-6-
Iag1
Figure 7-21. All-cause (nonaccidental) O3 excess mortality risk estimates (95% CI)
by season per standardized increment (see Section 7.1.3.2). Analyses include
all ages unless otherwise noted.
7-98
-------
concentrations by season were not presented. Gryparis et al. (2004) estimated an excess risk of
1.8% (95% CI: 0.99, 3.06) and 0.70% (95% CI: -0.70, 2.17) per 30 ppb increase in 8-h max O3
at a 0- and 1-day lag during the warm season and cold season, respectively, in 21 European
cities. The median 8-h max O3 concentrations ranged from 30 to 99 ppb during the warm season
and 8 to 49 ppb during the cold season. In the U.S. 90 cities study (of which 80 cities had O3
data available), the winter (December, January, and February) mortality estimate was negative
(Samet et al., 2000 [reanalysis Dominici et al., 2003]) and was most likely attributable to the
inverse relationship between O3 and PM in the winter.
In the U.S. 95 communities study by Bell et al. (2004), no significant difference was found
between the estimates from all available data and warm-season-only data (for April- October);
however, no cool-season-only analyses were performed. The warm-season effect estimate using
the 7-day constrained distributed lag model was 0.78% (95% PI: 0.26, 1.30) excess risk per
20 ppb increase in 24-h avg O3, compared to 1.04% (95% PI: 0.54, 1.55) calculated using all
available data. In the 55 communities with year-round O3 data, the all-year effect estimate was
0.96% (95% PI: 0.32, 1.57). As stated in the previous section, the range of median
concentrations from the 55 communities monitored all year overlapped considerably with the
range from the 40 communities with warm-season-only data (see Figure 7-16).
All three recent meta-analyses (Bell et al., 2005; Ito et al., 2005; Levy et al. 2005), found
that the warm-season estimates were larger than all-year estimates. In Bell et al. (2005), the
warm-season estimate was 3.02% (95% PI: 1.45, 4.63), compared to the all-year estimate of
1.75% (95% PI: 1.10, 2.37). In the subset of 10 cities examined by Ito et al. (2005), the warm-
season and all-year estimates were 3.5% (95% CI: 2.1, 4.9) and 2.2% (95% CI: 0.8, 3.6),
respectively. Similarly, Levy et al. (2005) observed a 3.38% (95% CI: 2.27, 4.42) excess risk in
the warm season compared to a 1.64% (95% CI: 1.25, 2.03) excess risk using all-year data.
All results presented are percent excess risk in mortality per standardized increment.
Studies that conducted analysis by season indicate that O3-mortality risk estimates are
often larger in the warm season compared to the colder season. The seasonal dependence
of O3-mortality effects complicates interpretation of O3 risk estimates calculated from year-round
data without adequate adjustment of temporal trends.
7-99
-------
7.4.6 Ozone-Mortality Risk Estimates Adjusting for PM Exposure
The confounding between "winter type" pollution (e.g., CO, SO2, and NO2) and O3 is not
of great concern, because the peaks of these pollutants do not strongly coincide. The main
confounders of interest for O3, especially for the northeast United States, are "summer haze-
type" pollutants such as acid aerosols and sulfates. Since very few studies included these
chemical measurements, PM (especially PM2 5) data, may serve as surrogates. However, due to
the expected high correlation among the constituents of the "summer haze mix," multipollutant
models including these pollutants may result in unstable coefficients; and, therefore,
interpretation of such results requires some caution.
Figure 7-22 shows the O3 risk estimates with and without adjustment for PM indices using
all-year data in studies that conducted two-pollutant analyses. Approximately half of the O3 risk
estimates increased slightly, whereas the other half decreased slightly with the inclusion of PM
in the models. In general, the O3-mortality risk estimates were robust to adjustment for PM in
the models, with the exception of Los Angeles, CA data with PM10 (Kinney et al., 1995) and
Mexico City data with TSP (Borja-Aburto et al., 1997).
The U.S. 95 communities study by Bell et al. (2004) examined the sensitivity of acute
O3-mortality effects to potential confounding by PM10. Restricting analysis to days when
both O3 and PM10 data were available, the community-specific O3-mortality effect estimates as
well as the national average results indicated that O3 was robust to adjustment for PM10 (Bell
et al., 2004). There were insufficient data available to examine potential confounding by PM2 5.
One study (Lipfert et al., 2000a) reported O3 risk estimates with and without sulfate adjustment.
Lipfert et al. (2000a) calculated O3 risk estimates based on mean (45 ppb) less background (not
stated) levels of 1-h max O3 in seven counties in Pennsylvania and New Jersey. The O3 risk
estimate was not substantially affected by the addition of sulfate in the model (3.2% versus
3.0% with sulfate) and remained statistically significant.
Several O3-mortality studies examined the effect of confounding by PM indices in different
seasons (Figure 7-23). In analyses using all-year data and warm-season only data, O3 risk
estimates were once again fairly robust to adjustment for PM indices, with values showing both
slight increases and decreases with the inclusion of PM in the model. In the analyses using cool-
season data only, the O3 risk estimates all increased slightly with the adjustment of PM indices,
although none reached statistical significance.
7-100
-------
-10
Samet et al. (2000, reanalysis
Dominici et al., 2003): U.S. 90 cities -
Schwartz (2005): U.S. 14 cities -
with PM10 _
Kinney and Ozkaynak (1991):
Los Angeles County, CA —
with KM -
Kinney et al. (1995): Los Angeles County, CA -
with PM10 -
Fairley (2003): Santa Clara County, CA -
Gamble (1998): Dallas, TX -
with PM10 -
Ito and Thurston (1996): Cook County, IL -
with PM10 -
Lipfert et al. (2000a): 7 counties in PAand NJ -
withPM25 -
with PM10 -
Chock et al. (2000): Pittsburgh, PA (age 0-74) -
Touloumi et al. (1997): 4 European cities
(Athens, Barcelona, London, Paris) -
with BS -
Anderson et al. (1996): London, England -
with BS -
Hoek (2000, reanalysis Hoek, 2003):
The Netherlands —
with PM10 -
Verhoeffetal. (1996):
Amsterdam, the Netherlands -
with PM10 -
Borja-Aburto et al. (1997): Mexico City -
with TSP -
Borja-Aburto et al. (1998): SW Mexico City -
with PM25 -
Ostro et al. (1996): Santiago, Chile -
Kim et al. (2004): Seoul, Korea -
with PM-|o -
-5
I
% Change in Mortality
0 5 10
4-
15
I
-e-
-e-
lagO
Latin America
lag 0
lagO
lag 1
Iag1
-e-
lag 1-2
-•-
o'
lag 1-2
i=.n n 1 A f SE not given; significant at p = 0.055
lay U- I ^ I
• -^ SE not given; not
significant at p = 0.055
lagO
Best of lag Oto 3
by city
lagO
-e-
Iag1
-e—
lag 2
-e-
lagO
lag 1 -2
lag 1
Iag1
Figure 7-22. All-cause (nonaccidental) O3 excess mortality risk estimates (95% CI)
with adjustment for PM indices for all-year analyses per standardized
increment (see Section 7.1.3.2). Analyses include all ages unless
otherwise noted.
7-101
-------
% Change in Mortality
-15
Ostro (1995): San Bernardino County
and Riverside County, CA —
with PM2 5 -
Moolgavkar et al. (1995): Philadelphia, PA -
with TSP -
03 only -
with TSP •
Chock et al. (2000): Pittsburgh, PA (age 0-74) -
with PM10 -
O3 only —
with PM10 -
03 only —
with PM10 -
Gryparisetal.(2004):21 European cities —
with PM10 -
03 only —
with PM10 -
Anderson et al. (1996): London, England -
with BS -
03 only •
with BS •
O3 only •
with BS -
Ostro et al. (1996): Santiago, Chile -
with PM10 -
03 only —
Kim et al. (2004): Seoul, Korea -
with PM10 -
03 only —
with PM10 -
-10
I
-5
I
5
I
10
-e-
-e-
-x-
-e-
-e-
| Europe
Latin America
-X-
Asia
| X All year + Warm Q Cool |
lagO
Iag1
lagO
lag 0-1
lagO
-e-
-e-
-X-
Iag1
Figure 7-23. All-cause (nonaccidental) O3 excess mortality risk estimates (95% CI)
with adjustment for PM indices by season per standardized increment
(see Section 7.1.3.2). Analyses include all ages unless otherwise noted.
7-102
-------
The three recent meta-analyses (Bell et al., 2005; Ito et al., 2005; Levy et al. 2005) all
examined the influence of PM on O3 risk estimates. No substantial influence was observed in
any of these studies. In the analysis by Bell et al. (2005), the combined estimate without PM
adjustment was 1.75% (95% PI: 1.10, 2.37) from 41 estimates, and the combined estimate with
PM adjustment was 1.95% (95% PI: -0.06, 4.00) from 11 estimates per 20 ppb increase in 24-h
avg O3. In the meta-analysis of 15 cities by Ito et al. (2005), the combined estimate was 1.6%
(95% CI: 1.1, 2.2) and 1.5% (95% CI: 0.8, 2.2) per 20 ppb in 24-h avg O3 without and with PM
adjustment, respectively. The additional time-series analysis of six cities by Ito et al. found that
the influence of PM by season varied across alternative weather models but was never
substantial. Levy et al. (2005) examined the regression relationships between O3 and PM
indices (PM10 and PM2 5) with O3-mortality effect estimates for all year and by season. Positive
slopes, which might indicate potential confounding, were observed for PM2 5 on O3 risk estimates
in the summer and all-year periods, but the relationships were weak. The effect of one causal
variable (i.e., O3) is expected to be overestimated when a second causal variable (e.g., PM) is
excluded from the analysis, if the two variables are positively correlated and act in the same
direction. However, the results from these meta-analyses, as well as several single- and
multiple-city studies, indicate that copollutants generally do not appear to substantially
confound the association between O3 and mortality.
7.4.7 Ozone Risk Estimates for Specific Causes of Mortality
In addition to all-cause mortality, several studies examined broad underlying causes of
mortality, such as cardiovascular and respiratory causes. The U.S. 95 communities study (1987-
2000) analyzed O3 effect estimates from cardiovascular and respiratory mortality (Bell et al.,
2004). The analysis by Bell et al. (2005) used all available data, which included all-year data
from 55 communities and warm-season only data from 40 communities. The national average
estimate from the constrained distributed lag model was slightly greater for cardiopulmonary
deaths than deaths from all causes, with an excess risk of 1.28% (95% PI: 0.62, 1.97) compared
to 1.04% (95% PI: 0.54, 1.55) per 20 ppb increase in 24-h avg O3 in the preceding week.
In a related study, Huang et al. (2005) examined O3 effects on cardiopulmonary mortality
during the summers (June to September) of 1987 to 1994 in 19 large U.S. cities from the
NMMAPS database. Figure 7-24 presents the Bayesian city-specific and overall average O3 risk
7-103
-------
% Change in Mortality
-2
Santa Ana / Anaheim —
San Bernardino —
Chicago —
Pittsbyrgh —
Phoenix —
San Antonio —
Miami —
Atlanta —
Los Angeles —
Houston —
San Diego —
New York City —
Dallas / fort Worth —
Philadelphia —
Cleveland —
Oakland —
Seattle —
San Jose —
OVERALL -
0 -10 0 10 20 30 4(
i i i i i i i i i i i
•'[;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
Figure 7-24. City-specific Bayesian estimates and overall average for the percent change
(95% PI) in cardiopulmonary mortality per 20 ppb increase in 24-h avg O3 in
the previous week using a constrained distributed lag model for 19 U.S. cities
(NMMAPS), arranged by size of the effect estimate. Analyses were
conducted using summer (June to September) data only.
Source: Derived from Huang et al. (2005).
estimates for cardiopulmonary mortality per 20 ppb increase in 24-h avg O3 from a constrained
7-day distributed lag model. The O3 effect estimate was 2.52% (95% PI: 0.94, 4.10) excess risk
in cardiopulmonary mortality per 20 ppb increase in 24-h avg O3 in the preceding week for the
combined analysis of all cities. For analyses of summer data, confounding of the O3 effect by
PM is of concern as daily variations in O3 may be correlated to PM during the summer months.
Huang et al. observed that when PM10 was included in the model, the O3 effect estimate, on
average, remained positive and significant. As PM10 measurements were available only every
7-104
-------
1 to 6 days, only single-day lags were examined. At a 0-day lag, O3 was associated with a 1.47%
(95% PI: 0.54, 2.39) excess risk versus a 1.49% (95% PI: -0.66, 3.47) excess risk in
cardiopulmonary mortality in the O3-only model and after adjustment for PM10, respectively.
The slight sensitivity of the O3 health effects to the inclusion of PM10 in the model may indicate
a true confounding effect. However, as only the days with PM10 data available were included in
the analysis, the lack of significance is likely attributable to higher statistical uncertainty due to
the lack of daily PM10 measurements.
Figure 7-25 presents effect estimates for associations between O3 and cardiovascular
mortality for all-year and warm-season analyses. All studies, with the exception of Ponka et al.
(1998), showed positive associations between O3 and cardiovascular mortality. The median
24-h avg O3 was 9 ppb (5th % to 95th %: 2 to 26) in the Helsinki study by Ponka et al. (1998).
In addition to Cook County, IL (median 24-h avg O3 of 18 ppb), Moolgavkar (2000) also
examined the effect of O3 on cardiovascular mortality in Los Angeles County, CA (median 24-h
avg O3 of 24 ppb) and Maricopa County, AZ (median 24-h avg O3 of 25 ppb). Ozone was not
found to be associated with cardiovascular or cerebrovascular mortality in these two counties,
even when data was restricted to the warm season (quantitative risk estimates not presented).
In a follow-up study, Moolgavkar (2003) analyzed all circulatory system deaths in one broad
group in order to increase the power of the season-specific analyses using the Cook County and
Los Angeles County data. In this analysis, O3 concentrations were marginally associated with
increased circulatory mortality during the summer in Los Angeles County (1.6% [95% CI: -0.2,
3.5] excess risk per 20 ppb increase in 24-h avg O3). Note that both analyses were performed
using Poisson GAM with default convergence criteria.
As with all-cause mortality, there appears to be heterogeneity in the effect estimates across
studies. The cardiovascular mortality estimate from the meta-analysis by Bell et al. (2005)
appears to be close to the mode of the effect estimates from the various studies, as shown in
Figure 7-25. This is expected, given that many of these studies were also included in the
meta-analysis. Bell et al. (2005) observed that the posterior mean estimate for cardiovascular
causes (2.23% [95% PI: 1.36, 3.08] excess risk per 20 ppb increase in 24-h avg O3 from 25
estimates) was slightly larger than that for total mortality (1.75% [95% PI: 1.10, 2.37] excess
risk from 41 estimates). However, since cardiovascular deaths account for the largest fraction
(over 40%) of total deaths, it is not surprising that the risk estimates for cardiovascular mortality
7-105
-------
% Change in Cardiovascular Mortality
-20 -15 -10 -5 0 5 10 15 20 25 30 35 40
-------
are somewhat similar to those from all-cause mortality. Overall, the cardiovascular mortality
risk estimates in the current literature show consistently positive associations with some
heterogeneity (most estimates fall within the range of 1 to 8% per 40 ppb increase in 1-h
avg O3).
Several studies observed that the risk estimates for the respiratory category were larger
than the cardiovascular and total nonaccidental categories (e.g., Anderson et al., 1996; Gouveia
and Fletcher, 2000b; Gryparis et al., 2004; Zmirou et al., 1998). In the European 21 multicities
study (Gryparis et al., 2004), the warm-season effect estimate for respiratory mortality was
6.75% (95% CI: 4.38, 9.10) excess risk per 30 ppb increase in 8-h max O3, compared to 2.70%
(95% CI: 1.29, 4.32) for cardiovascular mortality and 1.82% (95% CI: 0.99, 3.06) for total
mortality. In contrast, other studies have found that the risk estimates for the respiratory
category were smaller or even negative, whereas the risk estimates for total or cardiovascular
categories were positive (e.g., Borja-Aburto et al., 1998; Bremner et al., 1999; Lipfert et al.,
2000a; Morgan et al., 1998b). The apparent inconsistencies across studies may be due in part to
the differences in model specifications, but they may also reflect the lower statistical power
associated with the smaller daily counts of the respiratory category (usually accounting for less
than 10% of total deaths) compared to the larger daily counts for the cardiovascular category
(approximately 40 to 50% of total deaths). Thus, an examination of the differences in risk
estimates across specific causes requires a large population and/or a long period of data
collection. In the meta-analysis by Bell et al. (2005), which combined 23 estimates from
17 studies for respiratory mortality, the effect estimate for respiratory causes was smaller (0.94%
[95% PI: -1.02, 2.96] excess risk per 20 ppb increase in 24-h avg O3) compared to the estimates
for total mortality (1.75% excess risk) and cardiovascular mortality (2.23% excess risk).
The analyses of a 9-year data set for the whole population of the Netherlands
(population = 14.8 million) provided risk estimates for more specific causes of mortality,
including COPD, pneumonia, and subcategories of cardiovascular causes (Hoek et al., 2000,
2001; reanalysis Hoek, 2003). The median 8-h avg O3 (12 p.m.-8 p.m.) was 24 ppb (range 1 to
117). The effect estimate for total nonaccidental mortality was 1.6% (95% CI: 0.9, 2.4) excess
risk per 30 ppb increase in 8-h avg O3. In comparison, the excess risk estimates for pneumonia
and COPD were 5.6% (95% CI: 1.8, 9.5) and 0.8% (95% CI: -2.4, 4.2), respectively. The
effect estimates for some of the cardiovascular subcategories, including heart failure (3.8% [95%
7-107
-------
CI: 0.5, 7.3]) and thrombosis-related disease (6.0% [95% CI: 1.1, 10.8]), showed greater risk
estimates than that for total mortality. However, these associations were not specific to O3.
For example, most of the pollutants examined, including PM10, BS, SO2, NO2, CO and NO3 ,
were associated with pneumonia. Therefore, it is difficult to make a causal inference specific
to O3 on the basis of these results.
De Leon et al. (2003) examined the role of contributing respiratory causes in the
association between air pollution and nonrespiratory mortality (circulatory and cancer) in
New York City during the period of 1985 to 1994. The mean 24-h avg O3 concentration was
21.6 ppb (5th % to 95th %: 7.0 to 45.0). The main finding of this study was that the estimated
excess mortality risks for PM10 were higher for nonrespiratory deaths that had contributing
respiratory causes compared to deaths without contributing respiratory causes in the older age
(75+ years) group. This pattern was also seen for CO and SO2, but not for O3. Therefore, this
study did not suggest a role of contributing respiratory causes in the association between O3 and
nonrespiratory causes of deaths.
In summary, several single-city studies observed positive associations between ambient O3
concentrations and cardiovascular mortality. In addition, a meta-analysis that examined specific
causes of mortality found that the cardiovascular mortality risk estimates were higher than those
for total mortality. The findings regarding the effect size for respiratory mortality have been less
consistent, possibly because of lower statistical power in this subcategory of mortality.
7.4.8 Ozone-Mortality Risk Estimates for Specific Subpopulations
Some studies have examined O3-mortality risk estimates in potentially susceptible
subpopulations, such as those persons with underlying cardiopulmonary disease. Sunyer et al.
(2002) examined the association between air pollution and mortality in a cohort of patients (467
men and 611 women) with severe asthma in Barcelona, Spain during the period of 1986 to 1995.
A case-crossover study design was used to estimate excess odds of mortality adjusting for
weather and epidemics in three groups: (1) those who had only one asthma emergency
department visit; (2) those who had more than one asthma emergency department visit; and
(3) those who had more than one asthma and COPD emergency department visit. Those with
more than one asthma emergency department visit showed the strongest associations with the
examined air pollutants, with NO2 being the most significant predictor, followed by O3. Sunyer
7-108
-------
et al. (2002) reported a significant association between O3 and all-cause deaths for this group
during the warm season, with an odds ratio of 2.82 (95% CI: 1.15, 6.87) per 40 ppb increase in
1-h max O3, compared to an odds ratio of 1.03 (95% CI: 0.60, 1.78) for those with only one
asthma emergency department visit and 1.08 (95% CI: 0.60, 1.92) for the group with a
concurrent diagnosis of COPD. The median 1-h max O3 concentration for all-year data was
36 ppb, but seasonal O3 concentrations were not presented. In another Barcelona study, Saez
et al. (1999) examined asthma mortality death among persons aged 2 to 45 years. Once again,
O3 and NO2 were the only air pollutants found to be significantly associated with asthma
mortality death. While the similarity of the patterns of associations between O3 and NO2 makes
it difficult to speculate on the specific causal role of O3, the results of these studies suggest that
individuals with severe asthma may make up a subpopulation that is sensitive to these pollutants.
Sunyer and Basagna (2001) also performed an analysis of emergency department visits for
individuals with COPD. The results from this study suggested that PM10, but not gases were
associated with mortality risks for the COPD cohort. The mean O3 concentration was not
provided (IQR was 11 ppb for 1-h max O3). However, a Mexico City study by Tellez-Rojo et al.
(2000) observed a significant association between COPD mortality and O3, as well as PM10,
among patients living outside a medical unit. For a cumulative 5-day lag, an excess risk of
15.6% (95% CI: 4.0, 28.4) per 40 ppb increase in 1-h max O3 was observed for COPD mortality.
The mean 1-h max O3 was 69.4 ppb (SD 17.2).
Goldberg et al. (2003) investigated associations between air pollution and daily mortality
with congestive heart failure as the underlying cause of death in patients aged 65 years or more
in Montreal, Quebec during the period of 1984 to 1993. The mean 24-h avg O3 was 15.0 ppb
(SD 8.8). Analysis was stratified into two groups, those whose underlying cause of death was
congestive heart failure and those with a diagnosis of congestive heart failure one year before
their death. No association was found between daily mortality for congestive heart failure and
any pollutants. However, associations were found between daily mortality and coefficient of
haze, SO2, and NO2 among patients who were classified as having congestive heart failure before
death. In the case of O3, positive risk estimates were observed for year-round and warm-season
data; however, the results were not statistically significant. Although the 10-year study period
for this data was long, the daily mean death counts for the specific subcategory chosen was
relatively small (0.7/day for mortality with congestive heart failure as underlying cause of death
7-109
-------
and 4.0/day for total mortality in patients previously diagnosed with congestive heart failure),
thus limiting the power of the study.
In the meta-analysis by Bell et al. (2005), a combined estimate was obtained for the elderly
population (age 64 years and older or 65 years and older) using 10 estimates from 9 studies. The
posterior mean estimate for the elderly category (2.92% [95% PI: 1.34, 4.51] per 20 ppb
increase in 24-h avg O3) was larger than that from all ages (1.75% [95% PI: 1.10, 2.37] from
41 estimates). The results from the Bell et al. (2005) meta-analysis suggest that the elderly
population may be particularly susceptible to O3-related mortality.
Few studies have examined O3-mortality effects for specific subpopulations. Among those
that investigated the effect of air pollution in populations with underlying cardiopulmonary
diseases, associations were not unique to O3 but, rather, were shared with other pollutants. There
is suggestive evidence that severe asthmatics may be susceptible to the mortality effects
associated with NO2 and O3. In addition, the meta-analysis by Bell et al. (2005) suggests that the
elderly population may be more affected by O3.
7.4.9 Summary of Acute Ozone Effects on Mortality
• A substantial body of new data on acute mortality effects of O3 has emerged since the
1996 O3 AQCD. While uncertainties remain in some areas, it can be concluded that
robust associations have been identified between various measures of daily O3
concentrations and increased risk of mortality. Most of the single-pollutant model
estimates from single-city studies fall in the range between 0.5 to 5% excess deaths per
standardized increment. The corresponding summary estimates in large U.S. multicity
studies ranged between 0.5 to 1%, with some studies noting heterogeneity across cities
and studies. These associations could not be readily explained by confounding due to
time, weather, nor copollutants. Differences between populations as well as model
specifications likely contributed to some of the observed heterogeneity in risk
estimates across studies.
• Most studies calculated O3-mortality risk estimates using all-year data. The results
from studies that conducted stratified analyses by season suggested that the O3 risk
estimates were larger in the warm season. Some of the risk estimates in the cool
season were negative, possibly reflecting the negative correlation between low-level
O3 and PM (and/or other primary pollutants) during that season. Thus, even with
adjustment for temporal trends, the O3 risk estimates obtained for year-round data may
be misleading. In locations with considerable seasonal variation, season-specific
analyses may better elucidate the effect of O3 on mortality.
7-110
-------
• The majority of the available O3-mortality risk estimates were computed for a single-
day lag. Choosing the optimal lag out of several lags examined may bias the single-
day risk estimate upward. However, recent findings from the largest U.S. 95
communities study indicated that a strong association between O3 and mortality was
observed with a 7-day distributed lag model. Thus, it is possible that the effect of
acute O3 exposure on mortality persists over several days. Since the parameters
estimated from single-day lag versus multiday lag models are not the same,
comparison of these results are difficult.
• Some studies examined specific subcategories of mortality, but most of these studies
had limited statistical power to detect associations due to the small daily mortality
counts. A recent meta-analysis indicated that there was a slightly greater risk of
cardiovascular mortality compared to total mortality.
• Few studies examined the effect of O3 on mortality in subpopulations with underlying
cardiopulmonary diseases. Similar to cause-specific mortality, these population-
specific studies had limited statistical power to detect associations.
7.5 EFFECTS OF CHRONIC OZONE EXPOSURE
7.5.1 Summary of Key Findings on Studies of Health Effects and Chronic
Ozone Exposure from the 1996 Ozone AQCD
The 1996 O3 AQCD concluded that there was insufficient evidence from the limited
number of studies to determine whether long-term ambient O3 exposures resulted in chronic
health effects. However, the aggregate evidence suggested that chronic O3 exposure, along with
other environmental factors, could be responsible for health effects in exposed populations.
7.5.2 Introduction to Morbidity Effects of Chronic Ozone Exposure
Several new longitudinal epidemiologic investigations have yielded information on health
effects of long-term O3 exposures. Epidemiologic interest in investigating long-term effects has
been motivated by several considerations. Animal toxicology studies carried out from the late
1980s onward demonstrated that long-term exposures can result in permanent changes in the
small airways of the lung, including remodeling of the airway architecture (specifically the distal
airways and centriacinar region) and deposition of collagen, as discussed earlier in Chapter 5.
These changes result from the damage and repair processes that occur with repeated exposure.
Indices of fibrosis also were found to persist after exposure in some of the studies. Collectively,
7-111
-------
these findings provide a potential pathophysiologic basis for the changes in airway function
observed in children in longitudinal studies. Seasonal ambient patterns of exposure may be of
greater concern than continuous daily exposure. In the classical study by Tyler et al. (1988),
young monkeys with seasonal exposure to O3, but not those with daily exposure, experienced
increases in total lung collagen content, chest wall compliance, and inspiratory capacity,
suggesting a delay in lung maturation in seasonally-exposed animals.
Controlled human exposure studies clearly demonstrated acute inflammation in the lung at
ambient exposure levels. Epidemiologic studies could examine whether repeated exposures over
multiple episode periods and/or multiple years would lead to persistent inflammation and result
in damage to the human lung, especially in the small, terminal bronchiolar regions where
vulnerability is greatest. However, the challenges to addressing these issues in epidemiologic
studies are formidable, and as a result there exists relatively limited literature in this area. Long-
term O3 concentrations tend to be correlated with long-term concentrations of other pollutants,
making specific attribution difficult. Subtle pulmonary effects require health outcome measures
that are sensitive, and must usually be directly collected from individual human subjects, rather
than from administrative data bases. Although these factors make chronic studies difficult and
expensive to conduct, efforts must be made to design studies with adequate power to examine
the hypothesis being tested. Epidemiologic studies have the potential to provide important new
insights on the links between chronic exposure to O3 and the occurrence of human health effects.
This section reviews studies published from 1996 onward in which health effects were
tested in relation to extended O3 exposures ranging from several weeks to many years
(Table AX7-6 in the Annex 7, Section AX7.1). The available literature falls into four general
categories: (1) studies examining seasonal changes in lung function as related to O3 exposures
in peak season; (2) studies addressing smaller increases in lung function during childhood or
decline of lung function beyond childhood in relation to long-term O3 exposures; (3) studies
addressing respiratory inflammation in high versus low exposure groups or time periods; and
(4) studies addressing longitudinal and cross-sectional associations between long-term O3
exposures and asthma development and prevalence.
7-112
-------
7.5.3 Seasonal Ozone Effects on Lung Function
While it has been well-documented in both chamber and field studies that daily, multihour
exposures to O3 result in transient declines in lung function, much less is known about the effects
of repeated exposures to O3 over extended periods on lung function. Several new studies
reported during the past decade have examined lung function changes over seasonal time periods
with differing levels of O3 exposures (Frischer et al., 1999; Horak et al., 2002a,b; Ihorst et al.,
2004; Kinney and Lippmann, 2000; Kopp et al., 2000). The seasonal effects of O3 are examined
first in this section. The next section then discusses effects of O3 exposures over years, as
opposed to over seasons, in addition to multiyear analyses of seasonal studies.
In a large Austrian study, Frischer et al. (1999) carried out repeated lung function
measurements in 1,150 school children (mean age 7.8 years) from nine towns that differed in
mean O3 levels. Lung function was measured in the spring and fall over a 3-year period from
1994 to 1996, yielding six measurements per child. Mean summertime 24-h avg O3
concentrations ranged from 32.4 to 37.3 ppb during the three summers. Growth-related
increases in lung function over the summer season were reduced in relation to seasonal mean O3
levels. Ozone was associated with a change of -156.6 mL (95% CI: -209.5, -103.7) (central
estimate: -0.029 mL/day/ppb x 90 days/year x 3 years x 20 ppb) in FEVj increase for each
20 ppb increase in mean 24-h avg O3 concentrations over the three summers and -129.6 mL
(95% CI: -193.1, -66.1) over the three winters. When analyses were restricted to children who
had spent the whole summer period in their community, the changes were greater, with an O3-
related -183.6 mL (95% CI: -278.9, -88.3) change in FEVj increase over three summers.
Other pollutants (PM10, SO2, and NO2) had less consistent associations with changes in lung
function. Horak et al. (2002a,b) extended the study of Frischer et al. (1999) with an additional
year of data and found that seasonal mean O3 was associated with a negative effect on increases
in lung function, confirming results from the previous 3-year study. In an editorial, Tager (1999)
stated that the Frischer et al. (1999) data provided the first prospective evidence of an association
between exposure to ambient air pollution and alterations in lung function in children. Tager
further noted that the prospective study design represented a substantial improvement over data
derived from cross-sectional studies and should be emulated. However, Tager also cautioned
that it was difficult to attribute the reported effects to O3 alone independently of copollutants.
7-113
-------
Kopp et al. (2000) reported smaller increases in lung function among a cohort of 797
children from 10 communities in Austria and southwestern Germany who were exposed to high
levels of ambient O3 (mean O3 concentration of 44 to 52 ppb from April to October 1994).
Children residing in lower O3 (24 to 33 ppb) areas experienced a 43 mL increase in FEVl3
whereas those in high O3 areas only manifested a 16 mL increase during the summer of 1994.
Similar results were found in data for the summer of 1995. In another study of 15 communities
in Austria and southwestern Germany, Ihorst et al. (2004) examined 2,153 children with a
median age of 7.6 years and reported summer pulmonary function results which indicated that
significantly lower FVC and FEVj increases were associated with higher O3 exposures in the
summer, but not in the winter. Semi-annual mean O3 concentrations ranged from 22 to 54 ppb
during the summer and 4 to 36 ppb during the winter.
In a pilot study (Kinney and Lippmann, 2000), 72 nonsmoking adults (mean age 20 years)
from the second year class of students at the U.S. Military Academy in West Point, NY provided
two lung function measurements, one before and one after a 5-week long summer training
program at four locations. There was a greater decline in FEVj among students at the Fort Dix
location (78 mL) as compared to students at the other locations (31 mL). Ozone levels at Fort
Dix averaged 71 ppb (mean of daily 1-h max O3) over the summer training period versus mean
values of 55 to 62 ppb at the other three locations. In addition to the higher mean O3 level, Fort
Dix had greater peak O3 values (23 hours >100 ppb) compared to the other locations (1 hour
>100 ppb). Ambient levels of other pollutants, PM10 and SO2, were relatively low during the
study and did not vary across the four sites. Although the conclusions that can be drawn are
limited by the small study size, the results appear to be consistent with a seasonal decline in lung
function that may be due, in part, to O3 exposures. An exploratory observation from this study
was that there appeared to be a larger decline for those subjects who completed their post-
summer lung function measurements within two weeks after returning from training compared to
those measured 3 to 4 weeks after training, results consistent with some degree of rebound of
function following the summer exposure period.
Collectively, the above studies indicate that seasonal O3 exposure is associated with
smaller increases in lung function in children. The study by Kinney and Lippman (2000) also
provides some limited evidence suggesting that seasonal O3 also may affect lung function in
young adults.
7-114
-------
7.5.4 Chronic Ozone Exposure Effects on Lung Function and
Respiratory Symptoms
Lung capacity grows during childhood and adolescence as body size increases, reaches
a maximum during the 20s, and then begins to decline steadily and progressively with age.
There has long been concern that long-term exposure to air pollution might lead to slower
growth in lung capacity, diminished maximally attained capacity, and/or more rapid decline
in capacity with age. The concern arises by analogy with cigarette smoking, where it is well-
documented that lung function declines more rapidly with age in a dose-dependent manner
among adults who smoke cigarettes. Adults who stop smoking return to a normal rate of decline
in capacity, although there is no evidence that they regain the capacity previously lost due to
smoking (Burchfiel et al., 1995). Because O3 is a strong respiratory irritant and is associated
with acute lung function declines as well as inflammation and re-structuring of the respiratory
airways, it seems plausible that there might be a negative impact of long-term O3 exposures on
lung function. Exposures that negatively affect increases in lung function during childhood, in
particular, might have greater long-term risks. Thus, studies of effects on the rates of increases
in lung function in children are especially important.
Several studies published over the past decade have examined the relationship between
long-term O3 exposure and lung function. The most extensive and robust study of respiratory
effects in relation to long-term air pollution exposures among children in the United States is the
Children's Health Study carried out in 12 communities of southern California starting in 1993
(Avol et al., 2001; Gauderman et al., 2000, 2002, 2004a,b; Peters et al., 1999a,b). The first
cohort included children from the fourth, seventh, and tenth grades. A total of 3,676 students
completed questionnaires regarding their lifetime residential histories, historic and current health
status, residential characteristics, and physical activity. Among those students, 3,293 also
performed pulmonary function tests at the time of enrollment. Peters et al. (1999a) examined the
relationship between long-term (1986 to 1990) O3 exposures and self-reports of respiratory
symptoms and asthma in a cross-sectional analysis. Mean 1-h max O3 levels from 1986 to 1990
ranged from 30.2 ppb to 109.2 ppb across the 12 communities. For outcomes of current asthma,
bronchitis, cough, and wheeze, the reported odds ratios were 0.95 (95% CI: 0.70,1.29), 1.14
(95% CI: 0.84, 1.55), 0.98 (95% CI: 0.82, 1.17), and 1.08 (95% CI: 0.87, 1.35), respectively,
per 40 ppb increase in 1-h max O3. In another cross-sectional analysis of the relationship
7-115
-------
between lung function at baseline and levels of air pollution in the community, there was
evidence that annual mean O3 levels were associated with decreased FVC, FEVl3 PEF, and
FEF25.75 (the latter two being statistically significant) among females but not males (Peters et al.,
1999b).
Avol et al. (2001) examined 110 children from the first cohort who had moved from the
participating communities in southern California to other communities to determine whether
changes in air quality caused by relocation were associated with changes in annual increases in
lung function. Negative effects of O3 were observed for all spirometric parameters except FEVl3
but the associations were not as strong as those observed for PM10.
To examine the association between long-term exposure to air pollution and changes in
lung function, fourth graders (n = 1,759) from the first cohort were followed over eight years
(Gauderman et al., 2000, 2004a,b). During the follow-up period from 1993 to 2001, there was
an attrition of approximately 10% of subjects per year, resulting in 747 subjects being tested in
2001. The longitudinal pulmonary function data for each child was related to average air
pollution levels in each study community using a multistage regression approach. During the
monitoring period, there was substantial variation in the average levels of various air pollutants
across the 12 communities, but relatively little year-to-year variation within each community.
However, O3 levels did not vary widely across the communities in comparison to the other
pollutants. The mean annual average of 8-h avg O3 concentrations ranged from approximately
30 ppb in Long Beach to 65 ppb in Lake Arrowhead from 1994 to 2000. Average levels of O3
were not significantly correlated across communities with any other study pollutant. Analyses
indicated that there was no evidence that either 8-h avg O3 (10 a.m. to 6 p.m.) or 24-h avg O3
was associated with any measure of lung function growth over a 4-year (age 10 to 14 years;
Gauderman et al., 2000) or 8-year (age 10 to 18 years; Gauderman et al., 2004) period.
However, most of the other pollutants examined (including PM25, NO2, acid vapor, and
elemental carbon) were found to be significantly associated with reduced growth in lung
function. Also, a clinically low FEVj (i.e., FEVj <80% of predicted value) at age 18 years was
reported to be correlated with all pollutants examined except for O3 (Figure 7-26).
The Children's Health Study enrolled a second cohort of fourth graders (n = 1,678) in 1996
(Gauderman et al., 2002). While the strongest associations with negative lung function growth
were observed with acid vapors in this cohort, children from communities with higher 4-year
7-116
-------
Si
3
"D
1
0)
Q.
t
3?
o
00
V
•I"
8
6
4.
2
n
R = 0.04 4. up
P = 0.89
^ LB ^ RV
SD
* AT * AL
4SM 4LM ^LE A
A. 1 M
25 35 45 55 65
O3 from 10 a.m. to 6 p.m. (ppb)
0)
5 8-
T3
I 6^
b
CD
O
5?
CO
V
75
2
R = 0.66
P = 0.02
LB
20 30 40 50
PM10 (Mg/m3)
60
70
£ 10
0)
_3
5 8-
1 6-
± 4
o
V
LU
R = 0.75
P = 0.005
10
20
N02 (ppb)
LB
4SD
30
£ 10
o>
T3
o
09
V
UJ
40 "-
6
4-
2
0
R = 0.79
P = 0.002
AT. 4-'
AL
10
ML
15 20
PM2 5 (Mg/m3)
25
30
£
-------
average O3 levels also experienced smaller increases in various lung function parameters. The
strongest effect of O3 was on PEF, with children from the least-polluted community having a
1.21% (95% CI: 0.36, 2.06) greater increase in PEF compared to those from the most-polluted
communities. In two-pollutant models, only 8-h avg O3 and NO2 were significant joint
predictors of FEVj and maximal midexpiratory flow (MMEF). The range of mean annual
average 8-h avg O3 levels across the 12 communities was similar to that for the first cohort, 27 to
67 ppb.
In both cohorts of fourth graders, stratified analyses by time spent outdoors indicated a
stronger association between long-term O3 exposure and smaller increases in lung function
during the 4-year follow-up in children who spent more time outdoors, as shown in Table 7-3.
Jedrychowski et al. (2001) had reported a link between repeated respiratory symptoms and
smaller lung function increases. Gauderman et al. (2002), therefore, suggested that the
observation of reduced increases in lung function with increasing annual average air pollution
might be a consequence of repeated acute respiratory events after short-term increases in
pollutant levels. The findings of larger deficits in children who spend more time outdoors in the
afternoon adds some support for this possibility.
Although results from the second cohort of children are supportive of an association, the
definitive 8-year follow-up analysis of the first cohort (Gauderman et al., 2004a) provides little
evidence that long-term exposure to ambient O3 at current levels is associated with significant
deficits in the growth rate of lung function in children. It should be noted, however, that O3
exposure prior to the study period is unknown. In addition, there was only about a 2- to 2.5-fold
difference in O3 concentrations from the least to most polluted communities (mean annual
average of 8-h avg O3 ranged from 30 to 65 ppb), versus the ranges observed for the other
pollutants (which had 4- to 8-fold differences in concentrations). Finally, the results from the
stratified analyses by time spent outdoors indicate the potential for substantial misclassification
of long-term O3 exposure. The Children's Health Study, like most long-term epidemiologic
studies, estimated O3 exposure using centrally-located ambient monitors. Gonzales et al. (2003)
and Kiinzli et al. (1997) evaluated the use of retrospective questionnaires to reconstruct past
time-activity and location pattern information. Both studies found that questionnaires or activity
diaries might improve the assessment of chronic exposure in epidemiologic studies.
7-118
-------
Table 7-3. Difference in Annual Percent Increases in Lung Function from the Least to
the Most Polluted Community in the Children's Health Study by Time Spent Outdoors'
Lung Function
Parameter Cohort
FVC Cohort
Cohort
FEV; Cohort
Cohort
MMEF Cohort
Cohort
PEF Cohort
Cohort
More Time Outdoors °
b
1
2
1
2
1
2
1
2
% Change
Less Time Outdoors °
(95%CI)d %
-0.02% (-0.57,
-0
-0
.57% (-
.25% (-
-0.68%(-
-0
.55% (-
-0.48%(-
-0
-1.
.77% (-
.33% (-
1
1
1
2
1
2
2
.03,
.18,
.36,
.08,
.71,
.03,
.43,
0.54)
-0.09)
0.68)
0.00)
1.01)
0.78)
0.52)
-0.24)
Change (95% CI)d
-0.04% (-0.45,
-0
.06% (-0
.76,
-0.05% (-0.58,
-0
0.
-0
0.
-0
.29% (- 1.02,
23% (0.89, 1
.80% (-2.07,
.25% (-0.
.71% (-1
65,
.71,
0.37)
0
.66)
0.49)
0.46)
.36)
0.50)
1.
0
16)
.30)
a Results are derived from Gauderman et al. (2002).
b Cohort 1 includes children enrolled in 1993 as 4th graders and followed through 1997 (n = 1,457).
Cohort 2 includes children enrolled in 1996 as 4th graders and followed through 2000 (n = 1,678).
°More or less time outdoors is based on reported time spent outdoors during weekday afternoons.
Subjects were split into the two groups on the basis of the median reported time outdoors within each cohort.
dPercent change in lung function is per 30 ppb increase in 8-h avg O3 (10 a.m.-6 p.m.).
In a study of 15 communities in Austria and Germany, Ihorst et al. (2004) found no
associations between increases in lung function and mean summer O3 levels for FVC and FEVj
over a 3.5-year period, in contrast to the significant seasonal effects discussed in the earlier
section above. Unlike the reduced increases in lung function parameters over the first two
summers among children in high O3 areas noted above, a greater increase was observed by Ihorst
et al. during the third summer and no difference was observed during the fourth summer. The
mean O3 concentration during the first two summers across the 15 communities was similar to
that during the last two summers, 36.6 ppb and 35.1 ppb, respectively. The authors concluded
that medium-term effects on school children lung function were possibly present but were not
detected over a 3- to 5-year period due to partial reversibility. A study by Frischer et al. (1999)
showed results similar to the Ihorst et al. (2004) study. Although O3 was related to smaller
7-119
-------
increases in lung function when three years of data were analyzed collectively, the magnitude
and direction of the effect changed throughout the years. Ozone was associated with a change of
-34.0 mL (95% CI: -58.7, -9.3) in FEVj increase in the first year, compared to +7.3 mL (95%
CI: -20.8, 35.6) in the third year for each 20 ppb increase in mean 24-h avg O3. Mean
summertime O3 concentrations decreased slightly over the three years; the mean levels being
37.3 ppb, 35.4 ppb, and 32.4 ppb during the first, second, and third years, respectively.
A study by Gong et al. (1998b) examined lung function changes in 164 nonsmoking adults
(mean age 45 years) from a high O3 community in southern California, recruited from a cohort
of 208 who had been tested on two previous occasions. In the earlier analysis by Detels et al.
(1987), a significant decline in lung function was observed from 1977/1978 to 1982/1983.
In contrast, Gong et al. (1998b) observed a slight increase in FVC and FEVj from 1982/1983 to
1986/1987. The mean annual average 1-h max O3 level was 110 ppb from 1972 to 1982. For
1983 to 1989, the annual average 1-h max O3 levels ranged from 103 to 134 ppb. A consistent
decline in FEVj/FVC ratio was observed at all three time points (p < 0.0001). Among the
45 subjects who further participated in the controlled exposure study (0.40 ppm O3 over 2 hours
with intermittent exercise), declines in FEVj and FVC were associated with acute O3 exposure.
However, the acute changes in lung function were not associated with long-term changes in lung
function over a decade.
Evidence for a relationship between long-term O3 exposures and decrements in maximally
attained lung function was observed in a nationwide cohort of 520 first year students at Yale
College in New Haven, CT (Galizia and Kinney 1999; Kinney et al., 1998). Each student
performed one lung function test in the spring of their first year at college. Ozone exposures
were estimated by linking 10-year average summer season (June to August) 1-h max O3 levels at
the nearest monitoring station to the residential locations reported each year from birth to the
time of measurement. The mean 10-year average 1-h max O3 was 61.2 ppb (range 13-185).
Students who had lived four or more years in areas with long-term mean O3 levels above 80 ppb
had significantly lower FEVj (-3.07% [95% CI: -0.22, -5.92]) andFEF25.75 (-8.11% [95% CI:
-2.32, -13.90]) compared to their classmates with lower long-term O3 exposures. Stratification
by gender indicated much larger effect estimates for males than for females, which might reflect
higher outdoor activity levels and corresponding higher O3 exposure during childhood. Ozone
was the only air pollutant examined in this study. The authors noted, therefore, that it was not
7-120
-------
possible to determine whether, and to what extent, the observed effects are due to O3 alone or O3
in combination with unmeasured copollutants.
A similar study of 130 first year college freshmen at the University of California at
Berkeley also reported significant effects of O3 on lung function (Kiinzli et al., 1997; Tager
et al., 1998). Enrollment was limited to students from either the San Francisco or Los Angeles,
CA metropolitan areas. Lifetime monthly average 8-h avg O3 (10 a.m.-6 p.m.) concentrations
ranged from 16 to 33 ppb for San Francisco residents and 25 to 74 ppb for Los Angeles
residents. After controlling for city of origin, long-term O3 exposures were found to be
associated with declines in FEF25_75 and FEF75 (forced expiratory flow after 75% of FVC has
been exhaled). No effects were seen for PM10 and NO2. Kiinzli et al. (1997) noted that
significant changes in these mid- and end-expiratory flow measures could be considered early
indicators for pathologic changes that might ultimately progress to COPD, as evidenced by
animal studies that showed that the primary site of O3 injury in the lung was the centriacinar
region (see Chapter 5). This study was repeated in 2000 to 2002, with 255 freshmen who were
life-long residents of the Los Angeles or San Francisco areas (Tager et al., 2005). Lifetime
monthly average 8-h avg O3 levels ranged from 18 to 65 ppb for the entire cohort, with medians
of 36 ppb for the men and 33 ppb for the women. In contrast to results from the first cohort,
associations between long-term O3 exposure and declines in FEF25.75 and FEF75 were not
observed in the second cohort of freshmen. However, when the analysis was stratified by gender
and an interaction term for intrinsic airway size (FEF25.75/FVC ratio) was included in the model,
lifetime exposure to O3, as well as PM10 and NO2, was found to be associated with decreased
FEF25.75 and FEF75 for both men and women. The adverse impact of O3 exposure decreased with
increasing intrinsic airway size. Also, in multipollutant models including all three pollutants,
meaningful changes were not observed in the O3 effect estimates, but the PM10 and NO2
estimates were reduced substantially.
The cross-sectional studies in college freshmen provide suggestive evidence that young
adults who have grown up in high O3 communities may have reduced lung function compared to
those from low O3 communities. However, attributing the effects specifically to O3 is difficult,
given that the effects of coexposures to other ambient air pollutants have not necessarily been
adequately addressed in these cross-sectional long-term exposure studies. Thus, results of the
longitudinal southern California Children's Health Study, as well as those from the European
7-121
-------
studies, provide little evidence for impacts of long-term, relatively low-level O3 exposures on
lung function development in children. On the other hand, these studies appear to indicate likely
strong effects on lung function growth for long-term exposures to PM2 5 and/or certain other
copollutants.
7.5.5 Chronic Ozone Exposure and Respiratory Inflammation
As noted in Chapter 6, controlled human exposure chamber studies have demonstrated that
brief (2 to 6.6 hours) exposures to O3 while exercising result in inflammation in the lung,
including the alveolar region where gas exchange takes place. This acute exposure effect is
potentially important for effects of chronic exposure, because repeated inflammation can result
in the release of substances from inflammatory cells that can damage the sensitive cells lining
the lung. Over extended periods, repeated insults of this kind can lead to permanent damage to
and restructuring of the small airways and alveoli. In addition, since inflammation is a
fundamental feature of asthma, there is concern that O3-induced inflammation can exacerbate
existing asthma or perhaps promote the development of asthma among genetically pre-disposed
individuals. Several studies are discussed next, examining different outcomes related to
inflammation.
In a study by Kinney et al. (1996b), bronchoalveolar lavage fluids were collected in the
summer and winter from a group of 19 adult joggers living and working on an island in
New York City harbor. The mean 1-h max O3 concentrations for a 3-month period were 58 ppb
(maximum 110) in the summer and 32 ppb (maximum 64) in the winter. PM10 and NO2
concentrations were similar across the two seasons. There was little evidence for acute
inflammation in bronchoalveolar lavage fluids collected during the summer as compared to that
collected from the same subjects in the winter. However, there was evidence of enhanced cell
damage, as measured by lactate dehydrogenase, in the summer lavage fluids. These results
indicate that acute inflammatory responses may diminish with repeated exposures over the
course of a summer (as also demonstrated by multiday chamber exposures; see Chapter 6,
Section 6.9), but cell damage may be ongoing.
In a cross-sectional study by Frischer et al. (2001), urinary eosinophil protein was analyzed
as a marker of eosinophil activation in 877 school children living in nine Austrian communities
with varying ranges of O3 exposure levels. The results indicated that O3 exposure was
7-122
-------
significantly associated with eosinophil inflammation after adjusting for gender, site, and atopy.
The mean 30-day average O3 concentration before sample collection was 31.6 ppb (5th % to
95th%: 11.8 to 51.5).
Pollution effects in the nose can be viewed as a potential surrogate measure for effects that
may occur in the lungs, though doses to nasal tissues are usually higher for a given pollutant
concentration. As discussed in Chapter 5, morphological effects of O3 on the upper respiratory
tract can indicate quantitative changes in the nasal transitional respiratory epithelium. The
persistent nature of the O3-induced mucous cell metaplasia in rats, as discussed in Chapter 5,
suggests that O3 exposure may have the potential to induce similar long-lasting alterations in the
airways of humans. In a cross-sectional cohort study by Calderon-Garciduefias et al. (1995),
nasal lavage samples collected from children living in Mexico City (n = 38) were compared to
those from children living in a clean coastal town (n = 28). In Mexico City, the 1-h avg O3
concentrations exceeded 120 ppb for 4.4 h/day, while O3 levels were not detectable in the coastal
town. Mexico City children were examined four times within a 1-month period. Nasal
cytologies revealed that all Mexico City children had abnormal nasal mucosae, including
mucosal atrophy, marked decreases in the numbers of ciliated-type cells and goblet cells,
and squamous metaplasia. Significantly higher nasal polymorphonuclear leukocyte counts
(p < 0.001) and expression of human complement receptor type 3 (CD1 Ib) (p < 0.001) were also
observed in Mexico City children compared to the control children. Analyses using repeated
measurements indicated that the inflammatory response did not appear to correlate with the
previous day's O3 concentration, suggesting that there might be a competing inflammatory
mechanism at the bronchoalveolar level with structural injury following acute exposure.
Limited epidemiologic research has been conducted relating long-term O3 exposure to
inflammation. In the Mexico City study (Calderon-Garciduenas et al., 1995), specific attribution
of the nasal abnormalities to long-term O3 exposure is difficult, given the complex pollutant
mixture present in the ambient air. However, the inflammatory changes such as increased
eosinophil levels observed in the Austrian study (Frischer et al., 2001) would be consistent with
known effects of O3.
7-123
-------
7.5.6 Risk of Asthma Development
Recent longitudinal cohort studies have reported associations between long-term O3
exposures and the onset of asthma (McConnell et al., 2002; McDonnell et al., 1999). Significant
associations between long-term O3 exposure and new cases of asthma among adult males were
observed in a cohort of nonsmoking adults in California (Greer et al., 1993; McDonnell et al.,
1999). The Adventist Health and Smog (AHSMOG) study cohort of 3,914 (age 27 to 87 years,
36% male) was drawn from nonsmoking, non-Hispanic white California Seventh Day
Adventists, who were surveyed in 1977, 1987, and 1992. To be eligible, subjects had to have
lived 10 or more years within 5 miles of their current residence in 1977. Residences from 1977
onward were followed and linked in time and space to interpolate concentrations of O3, PM10,
SO42 , SO2, and NO2. New asthma cases were defined as self-reported doctor-diagnosed asthma
at either the 1987 or 1992 follow-up questionnaire among those who had not reported having
asthma upon enrollment in 1977. During the 10-year follow-up (1977 to 1987), the incidence of
new asthma was 2.1% for males and 2.2% for females (Greer et al., 1993). Ozone concentration
data were not provided. A relative risk of 3.12 (95% CI: 1.16, 5.85) per 10 ppb increase in
annual mean O3 (exposure metric not stated) was observed in males, compared to a relative risk
of 0.94 (95% CI: 0.65, 1.34) in females. In the 15-year follow-up study (1977-1992), 3.2% of
the eligible males and a slightly greater 4.3% of the eligible females developed adult asthma
(McDonnell et al., 1999). The mean 20-year average for 8-h avg O3 (9 a.m. to 5 p.m.) was
46.5 ppb (SD 15.3) from 1973 to 1992. For males, the relative risk of developing asthma was
2.27 (95% CI: 1.03, 4.87) per 30 ppb increase in 8-h avg O3. Once again, there was no evidence
of an association between O3 and new-onset asthma in females (relative risk of 0.85 [95% CI:
0.55, 1.29]). The lack of an association does not necessarily indicate no effect of O3 on the
development of asthma among females. For example, differences between females and males in
time-activity patterns may influence relative exposures to O3, leading to greater misclassification
of exposure in females. None of the other pollutants (PM10, SO42", SO2, and NO2) were
associated with development of asthma in either males or females. Adjusting for copollutants
did not diminish the association between O3 and asthma incidence for males. The consistency of
the results in the two studies with different follow-up times, as well as the independent and
robust association between annual mean O3 concentrations and asthma incidence, provide
supportive evidence that long-term O3 exposure may be associated with the development of
7-124
-------
asthma in adult males. However, because the AHSMOG cohort was drawn from a narrow
subject definition, the representativeness of this cohort to the general U.S. population may
be limited.
A similar study of incident asthma cases in relation to air pollutants (O3, PM, NO2, and
acid vapors) among children was carried out in the Children's Health Study (McConnell et al.,
2002). A total of 3,535 initially nonasthmatic children (ages 9 to 16 years at enrollment) were
followed for up to 5 years to identify new-onset asthma cases. Communities were stratified by
pollution levels, with six high-O3 communities (mean 1-h max O3 of 75.4 ppb [SD 6.8] over four
years) and six low-O3 communities (mean 50.1 ppb [SD 11.0]). Ozone concentrations were not
strongly correlated with the other pollutants. New diagnoses of asthma were reported for
265 children during the follow-up period. Asthma risk was not higher for residents of the six
high-O3 communities versus residents of the six low-O3 communities. However, within the
high-O3 communities, asthma risk was 3.3 (95% CI: 1.9, 5.8) times greater for children who
played three or more sports as compared with children who played no sports. None of the
children who lived in high-O3 communities and played three or more sports had a family history
of asthma. This association was absent in the low-O3 communities (relative risk of 0.8 [95% CI:
0.4, 1.6]). No associations with asthma were seen for PM10, PM25, NO2, or inorganic acid
vapors. These results suggest effect modification by physical activity of the impacts of O3 on
asthma risk. The overall observed pattern of effects of sports participation on asthma risk was
robust to adjustment for socioeconomic status, history of allergy, family history of asthma,
insurance, maternal smoking, and body mass index. Playing sports may index greater outdoor
activity when O3 levels are higher and an increased ventilation rate, which may lead to increased
O3 exposure. It should be noted, however, that these findings were based on a small number of
new asthma cases (n = 29 among children who played three or more sports) and were not based
on a priori hypotheses. Future replication of these findings in other cohorts would lend greater
weight to a causal interpretation.
Recent cross-sectional surveys have detected no associations between long-term O3
exposures and asthma prevalence, asthma-related symptoms, or allergy to common aeroallergens
in children after controlling for covariates (Charpin et al., 1999; Kuo et al., 2002; Ramadour
et al., 2000). It should be noted that O3 levels were quite low in all cases, with a range of 16 to
27 ppb for 8-h max O3. The longitudinal study design, which observes new onset of asthma
7-125
-------
prospectively, provides stronger evidence in relation to the question of O3 effects on asthma
development.
7.5.7 Respiratory Effects of Chronic Ozone Exposure on Susceptible
Populations
Studies on the effect of long-term O3 exposure on respiratory health have focused mostly
on children, a potentially susceptible population. Seasonal O3 exposure was associated with
smaller increases in lung function and respiratory inflammation in children. Other studies have
investigated additional groups of potentially susceptible individuals. In the Children's Health
Study, McConnell et al. (1999) examined the association between O3 levels and the prevalence
of chronic lower respiratory tract symptoms in 3,676 southern California children with asthma.
In this cross-sectional study, bronchitis, phlegm, and cough were not associated with annual
mean O3 concentrations in children with asthma or wheeze. All other pollutants examined
(PM10, PM2 5, NO2, and gaseous acid) were associated with an increase in phlegm but not cough.
The mean annual average 1-h max O3 concentration was 65.6 ppb (range 35.5 to 97.5) across the
12 communities.
In another Children's Health Study analysis, McConnell et al. (2003) evaluated
relationships between air pollutants and bronchitic symptoms among 475 children with asthma.
The mean 4-year average 8-h avg O3 (10 a.m.-6 p.m.) concentration was 47.2 ppb (range 28.3 to
65.8) across the 12 communities. For a 1 ppb increase in 8-h avg O3 averaged over 4 years, the
between-community odds ratio was 0.99 (95% CI: 0.98, 1.01) compared to the within-
community odds ratio of 1.06 (95% CI: 1.00, 1.12). The authors commented that if the larger
within-community effect estimates were correct, then other cross-sectional (between-
community) studies might have underestimated the true effect of air pollution on bronchitic
symptoms in children. These differences might be attributable to confounding by poorly
measured or unmeasured risk factors that vary between communities. PM2 5, NO2, and organic
carbon also were associated with bronchitic symptoms. In two-pollutant models, the within-
community effect estimates for O3 were markedly reduced and no longer significant in some
cases.
One recent study examined a susceptible group not examined before. Goss et al. (2004)
investigated the effect of O3 on pulmonary exacerbations and lung function in individuals over
7-126
-------
the age of 6 years with cystic fibrosis (n = 11,484). The study included patients enrolled in the
Cystic Fibrosis Foundation National Patient Registry, which contains demographic and clinical
data collected annually at accredited centers for cystic fibrosis. For 1999 through 2000, the
annual mean O3 concentration, calculated from 1-h averages from 616 monitors in the U.S. EPA
Aerometric Information Retrieval System (AIRS), was 51.0 ppb (SD 7.3). Exposure was
assessed by linking air pollution values from AIRS with the patient's home ZIP code. No clear
association was found between annual mean O3 and lung function parameters. However, a
10 ppb increase in annual mean O3 was associated with a 10% (95% CI: 3, 17) increase in the
odds of two or more pulmonary exacerbations. Significant excess odds of pulmonary
exacerbations also were observed with increased annual mean PM10 and PM2 5 concentrations.
The O3 effect was robust to adjustment for PM10 and PM25, 8% (95% CI: 1,15) increase in
odds of two or more pulmonary exacerbations per 10 ppb increase in annual mean O3.
In summary, few studies have identified and investigated potentially susceptible
populations. Results from the Children's Health Study do not provide strong evidence that
asthmatic children are particularly susceptible to long-term exposure to O3. Ozone exposure
was, however, shown to be associated with adverse respiratory health responses in individuals
with cystic fibrosis.
7.5.8 Effects of Chronic Ozone Exposure on Mortality and Cancer Incidence
There is inconsistent and inconclusive evidence for a relationship between long-term O3
exposure and increased mortality and cancer risk (see Table AX7-7 in Annex 7, Section AX7.1).
In the Harvard Six Cities Study (Dockery et al., 1993), adjusted mortality rate ratios were
examined in relation to long-term mean O3 concentrations in six cities: Topeka, KS; St. Louis,
MO; Portage, WI; Harriman, TN; Steubenville, OH; and Watertown, MA. Mortality rate ratios
were adjusted for age, sex, smoking, education, and body mass index. Mean O3 concentrations
from 1977 to 1985 ranged from 19.7 ppb in Watertown to 28.0 ppb in Portage. Long-term
mean O3 concentrations were not found to be associated with mortality in the six cities.
However, strong associations were observed between long-term mean concentrations of PM15,
PM2 5, and sulfate particles and mortality.
In a large prospective cohort study of approximately 500,000 U.S. adults, Pope et al.
(2002) examined the effects of long-term exposure to air pollutants on mortality (American
7-127
-------
Cancer Society, Cancer Prevention Study II). All cause, cardiopulmonary, lung cancer, and all
other cause mortality risk estimates for long-term O3 exposure are shown in Figure 7-27. No
consistent positive associations were observed between O3 and mortality. The mortality risk
estimates were larger when analyses were restricted to the summer months (July to September)
when O3 levels were generally higher. The mean summertime 1-h max O3 level from 1982 to
1998 was 59.7 ppb (SD 12.8). The O3-mortality risk estimates were positive for all-cause and
cardiopulmonary mortality, with a marginally significant estimate for cardiopulmonary mortality
in the summer months; but a nonsignificant negative O3 risk estimate was observed for lung
cancer mortality. In contrast, consistent positive and significant effects of PM2 5 were observed
for both lung cancer and cardiopulmonary mortality.
o
3?
in
o»_
Is"
'£
— i
Mortality
1
•i
4?
-:
Years of Data
Collection
1980
1982-1998
1982-1998 (Jul-Sep)
No. of Metropolitan
Areas
134
119
134
No. of Participants
in Thousands
569
525
557
1-h Max O3
Mean (SD)
47.9 ppb (11.0)
45.5 ppb (7.3)
59.7 ppb (12. 8)
Figure 7-27. Adjusted O3-mortality relative risk estimates (95% CI) by cause of mortality
and time period of analysis per subject-weighted mean O3 concentration in
the Cancer Prevention Study II by the American Cancer Society.
Source: Derived from Pope et al. (2002).
7-128
-------
Lipfert et al. (2000b, 2003) reported positive effects on all-cause mortality for peak O3
exposures (95th percentile levels) in the U.S. Veterans Cohort study of approximately 50,000
middle-aged men recruited with a diagnosis of hypertension. The actual analysis involved
smaller subcohorts based on exposure and mortality follow-up periods. Four separate exposure
periods were associated with three mortality follow-up periods. The mean 95th percentile O3
concentration ranged from 85 to 140 ppb in the four exposure periods. In a preliminary
screening of regression results, Lipfert et al. (2000b) found a negative association for mean O3
and a positive relationship for peak O3; thus, peak O3 was used in subsequent analyses.
The mean of the peak values ranged from 85 to 140 ppb over the four exposure periods.
For concurrent exposure periods, peak O3 was positively associated with all-cause mortality,
with a 9.4% (95% CI: 0.4, 18.4) excess risk per mean 95th percentile O3 less estimated
background level (not stated). When exposure periods preceding death were considered, no
association between O3 and mortality was observed (-0.2% [95% CI: -12.5, 12.1]). In a further
analysis, Lipfert et al. (2003) reported the strongest positive association for concurrent exposure
to peak O3 for the subset of subjects with low diastolic blood pressure during the 1982 to 1988
period. Once again, the O3 effect was diminished when exposure (1982 to 1988) preceded
mortality (1989 to 1996).
A long-term prospective cohort study (AHSMOG) of 6,338 nonsmoking, non-Hispanic
white individuals living in California examined the association between air pollutants and lung
cancer incidence (Beeson et al., 1998). The mean monthly average 24-h avg O3 level from 1973
to 1992 was 26.2 ppb (SD 7.7). Over the follow-up period of 1977 to 1992, 20 females (35%
smokers, n = 7) and 16 males (37.5% smokers, n = 6) developed lung cancer. An association
was observed between long-term O3 exposure and increased incidence of lung cancer in males
only. The relative risk for incident lung cancer among males was 3.56 (95% CI: 1.35, 9.42) for
an interquartile range increase in hours per year (556 hours/year) when O3 levels exceeded
100 ppb. A stronger association was seen for males who never smoked (4.48 [95% CI: 1.25,
16.04]) compared to those who smoked in the past (2.15 [95% CI: 0.42, 10.89]) (Beeson et al.,
1998).
A related study by Abbey et al. (1999) examined the effects of long-term O3 exposure on
all-cause (n = 1,575), cardiopulmonary (n = 1,029), nonmalignant respiratory (n = 410), and lung
cancer (n = 30) mortality in the same AHSMOG study population. A particular strength of this
7-129
-------
study was the extensive effort devoted to assessing long-term air pollution exposures, including
interpolation to residential and work locations from monitoring sites over time and space.
No associations with long-term O3 exposure were observed for all cause, cardiopulmonary, and
nonmalignant respiratory mortality. However, effects of O3 on lung cancer mortality paralleled
the results of the previous study by Beeson and colleagues. That is, an association between lung
cancer mortality and chronic O3 exposure was observed in males only, with a relative risk of
4.19 (95% CI: 1.81, 9.69) for an interquartile range increase in hours per year (551 hours/year)
when O3 levels exceeded 100 ppb. The gender-specific O3 effects may be partially attributable
to differences in activity and time spent outdoors. The questionnaires indicated that males spent
approximately twice as much time outdoors and performed more vigorous exercises outdoors,
especially during the summer, compared to the females. However, the very small numbers of
lung cancer deaths (n = 12 for males) raise concerns with regard to the precision of the effect
estimate, as evidenced by the wide confidence intervals. The lack of an association of chronic
O3 exposure with other mortality outcomes, which had much larger samples sizes, also is of
concern. A study by Pereira et al. (2005) provides limited evidence possibly suggestive of an
association between O3 and increase risk of cancer. Correlations between average air pollution
data from 1981 to 1990 and cases of larynx and lung cancer in 1997 were assessed in
communities of Sao Paulo, Brazil. Of all the pollutants examined (PM10, NO2, NOX, SO2, CO,
and O3), O3 was best correlated with cases of larynx (r = 0.99, p = 0.01) and lung (r = 0.72,
p = 0.28) cancer.
Only a few studies have examined the effect of chronic O3 exposure on mortality outcomes
and incidence of cancer. Consistent associations with long-term O3 exposure have not been
observed for all-cause and cardiopulmonary mortality. There is some very limited evidence
suggestive of an association between chronic O3 exposure and lung cancer incidence and
mortality; however, uncertainty remains due to the small number of available studies and the
very small numbers of lung cancer cases observed in most of the studies. Of particular note is
the fact that the very large American Cancer Society study did not find any association between
long-term O3 concentrations and lung cancer. In addition, the weight of evidence from
toxicologic studies does not support ambient O3 as a pulmonary carcinogen in laboratory animal
models (National Toxicology Program, 1994).
7-130
-------
7.5.9 Effects of Ozone on Birth-Related Health Outcomes
In recent years, air pollution epidemiologic studies have examined impacts on birth-related
endpoints, including intrauterine, perinatal, postneonatal, and infant deaths; premature births;
intrauterine growth retardation; very low birth weight (weight <1500 grams) and low birth
weight (weight <2500 grams); and birth defects. However, the majority of these studies did not
examine the effect of O3. In those limited studies that investigated O3, no associations were
found between O3 and birth outcomes, with the possible exception of birth defects. The
following is a synopsis of the literature on this topic.
Pereira et al. (1998) investigated the impacts of NO2, SO2, CO, O3, and PM10 on
intrauterine mortality in Sao Paulo, Brazil during 1991 and 1992. Mean 1-h max O3 was
34.8 ppb (SD 23.2). Intrauterine mortality was most significantly associated with NO2, and less
for SO2 and CO. No association was found for O3 or PM10. Pereira et al. (1998) also sampled
umbilical cord blood from healthy nonsmoking pregnant women soon after delivery in 1995
and analyzed the blood for carboxyhemoglobin. They found an association between
carboxyhemoglobin levels and ambient CO after adjusting for passive smoking and weight,
suggesting an impact of CO on the fetus. Also, Loomis et al. (1999) examined the association
between several air pollutants (NO2, SO2, CO, O3, PM2 5) and infant mortality in Mexico City in
the years 1993 to 1995. Mean 24-h avg O3 was 44.1 ppb (SD 15.7). They reported that the
strongest association was found for PM2 5 with a 3- to 5-day cumulative lag. They noted that
infant mortality also was associated with NO2 and O3 at a 3- to 5-day lag, but not as consistently
as PM2 5. There have been other studies that examined possible air pollution effects on
postneonatal mortality (Bobak and Leon, 1992; Bobak and Leon, 1999; Kaiser et al., 2004;
Woodruff et al., 1997), but these studies did not examine O3.
Ritz and Yu (1999) investigated the effects of ambient CO on low birth weight among
children born in southern California between 1989 and 1993. They focused on CO, stating that
"a biologic mechanism for fetal effects has been proposed for CO, but not for other air
pollutants." They found that exposure to higher levels of ambient CO during the last trimester
was associated with an increased risk for low birth weight. Using available data, they also
estimated last-trimester exposures for NO2, O3, and PM10. The last trimester average O3 level
was 20.9 ppb (range 3.0 to 49.4). NO2 and PM10 were positively associated with CO, but O3 was
negatively associated with CO (r = -0.65). Ha et al. (2001) examined CO, NO2, SO2, O3, and
7-131
-------
TSP for their associations with low birth weight in Seoul, Korea during the years 1996 and 1997
(data were analyzed using Poisson GAM with default convergence parameters). They estimated
first and third trimester exposures by averaging daily air pollution levels during the
corresponding days for the registered births. The median 8-h avg O3 levels were 22.4 ppb
(IQR 13.6) for the first trimester and 23.3 ppb (IQR 16.1) for the third trimester. Ha et al. (2001)
found that first trimester exposures of CO, NO2, SO2, and TSP were associated with increased
risk of low birth weight, whereas O3 was associated with a decreased risk. The opposite pattern
was observed for third trimester exposures, with an increased risk of low birth weight found only
for O3. When exposures from both trimesters were examined simultaneously, the associations of
first trimester exposures of CO, NO2, SO2, and TSP with increased risk of low birth weight
remained; however, the association between third trimester O3 exposure and low birth weight
was diminished. Based on these results, Ha et al. concluded that exposures to CO, NO2, SO2,
and TSP in the first trimester were risk factors for low birth weight. Note that neither of these
studies examined the air pollution effect by season. Other studies that examined the associations
between air pollution and low birth weight (Bobak, 2000; Bobak and Leon, 1999; Lin et al.,
2001; Maisonet et al., 2001; Wang et al., 1997) found associations between low birth weight and
either one or more of CO, SO2, NO2 and PM indices, but did not examine O3 data. Collectively,
these results do not provide any credible evidence of O3 contributing to low birth weight.
Two studies by Dejmek et al. (1999, 2000) examined the relationship between ambient air
pollution and risk of intrauterine growth retardation in a highly polluted area of Northern
Bohemia (Teplice District). Both studies, however, focused on PM indices and did not analyze
gaseous pollutants, such as O3.
A few studies have examined the association between air pollution and premature births
(Bobak, 2000; Ritz et al., 2000; Xu et al., 1995), but only Ritz et al. (2000) included O3 in their
analysis. Ritz et al. evaluated the effect of air pollution exposure during pregnancy on the
occurrence of preterm birth in a cohort of 97,518 neonates born in southern California.
CO, NO2, SO2, O3, and PM10 data measured at 17 air quality monitoring stations were used to
estimate the average exposures for the first month and the last 6 weeks of pregnancy. The
mean 8-h avg O3 (9 a.m. to 5 p.m.) level was 36.9 ppb during both time periods. They found
associations between PM10 levels averaged for the last 6 weeks of pregnancy as well as PM10
levels averaged over the first month of pregnancy. Similar but weaker associations were found
7-132
-------
for CO, but no association was found for O3. The reported correlation matrix indicated that O3
was negatively correlated with CO (r = -0.45) and only weakly correlated with PM10 (r = 0.2).
Other results from Beijing, China (Xu et al., 1995) and the Czech Republic (Bobak, 2000)
suggested that SO2 and TSP were associated with preterm births. Considering that O3 tends to be
negatively correlated with winter-type pollutants, O3 is unlikely to be an important risk factor for
preterm births.
Ritz et al. (2002) evaluated the effect of air pollution on the occurrence of birth defects in
neonates and fetuses delivered in southern California from 1987 to 1993 as ascertained by the
California Birth Defects Monitoring Program. They averaged air pollution (CO, O3, PM10,
and NO2) levels measured at the assigned ambient station over the first, second, and third month
of gestation. Conventional, polytomous, and hierarchical logistic regressions were used to
estimate odds ratios for subgroups of cardiac and orofacial defects. Concentration-response
relationships were observed for second month CO exposure on ventricular septal defects, and
second month O3 exposure on aortic artery and valve defects, pulmonary artery and valve
anomalies, and conotruncal defects. The odds ratios observed for these outcomes were similar
and quite large (e.g., the odds ratios comparing the highest [monthly 24-h avg mean 34.9 ppb]
to lowest [monthly mean 6.4 ppb] O3 quartiles ranged from 2.0 to 2.7), and were not sensitive in
multipollutant models. Ritz et al. (2002) reported that they did not observe consistently
increased risks and concentration-response patterns for NO2 and PM10 after controlling for the
effects of CO and O3. Results from this study contrast to those for other birth-related outcomes,
in that both CO and O3 (presumably negatively correlated pollutants) were associated with birth
defects. Further, O3 showed associations with more birth defect outcomes than did CO.
It should be noted, however, that the concentration-response relationships were quite specific to
exposures during the second month. Associations with third month exposures were often
negative (though not significantly). Since both CO and O3 show strong seasonal peaks, it is
possible that seasonal confounding could have played some role in these associations. This is
the only study to date that examined the relationship between air pollution and birth defects.
In summary, O3 was not an important predictor of several birth-related outcomes including
intrauterine and infant mortality, premature births, and low birth weight. Birth-related outcomes
generally appear to be associated with air pollutants that tend to peak in the winter and are
possibly traffic-related, most notably CO. The strong results for CO are consistent with its
7-133
-------
ability to cross the placental barrier and the high affinity that hemoglobin in fetal blood has for
binding with it. However, given that most of these studies did not analyze the data by season,
seasonal confounding may have therefore influenced the reported associations. One study
reported some results suggestive of associations between exposures to O3 in the second month of
pregnancy and birth defects, but further evaluation of such potential associations is needed.
7.5.10 Summary of Chronic Ozone Exposure Effects on Morbidity
and Mortality
• In the past decade, important new longitudinal studies have examined the effects of
chronic O3 exposure on respiratory health outcomes, including seasonal declines in
lung function, increases in inflammation, and development of asthma in children and
adults. Seasonal O3 effects on lung function have been reported in several studies.
In contrast to supportive evidence from chronic animal studies, epidemiologic studies
of inflammation, new asthma development, and longer-term lung function declines
remain inconclusive.
• Few studies have investigated the effect of long-term O3 exposure on mortality and
cancer incidence. Uncertainties regarding the exposure period of relevance,
differential effects by gender, and inconsistencies across outcomes raise concerns
regarding plausibility. There is currently little evidence for a relationship between
chronic O3 exposure and increased risk of lung cancer or of mortality.
• A limited number of studies have examined the relationship between air pollution and
birth-related health outcomes, including mortality, premature births, low birth weights,
and birth defects. The most consistent associations with various birth outcomes were
observed for CO, with very little credible evidence being found for any O3 effects.
7.6 INTERPRETIVE ASSESSMENT OF THE EVIDENCE IN
EPIDEMIOLOGIC STUDIES OF OZONE HEALTH EFFECTS
7.6.1 Introduction
In the 1996 O3 AQCD, discussion of the available epidemiologic information focused
primarily on individual-level camp and exercise studies and on studies of hospital admissions
and emergency room visits. The field studies indicated concentration-response relationships of
ambient air O3 exposure with declines in pulmonary function, increases in respiratory symptoms,
and exacerbation of asthma, especially in children. Numerous new studies provide additional
7-134
-------
evidence for evaluating associations between O3 exposure and the above respiratory health
outcomes. The 1996 O3 AQCD review of aggregate population time-series studies indicated an
association between ambient O3 concentrations and increased hospitalizations. A limited
number of studies examined O3-mortality relationships. The current O3 AQCD presents results
from time-series studies that have addressed previously unresolved issues regarding potential
linkages between ambient O3 concentrations and health outcomes, particularly mortality. Daily
time-series studies minimize confounding by population characteristics (e.g., cigarette smoking,
diet, occupation) by following the same population from day to day. However, confounders
operating over shorter time scales can affect O3 risk estimates in these studies.
In this section, issues and attendant uncertainties that affect the interpretation of available
epidemiologic evidence regarding O3 health effects are discussed. The use of various indices to
represent O3 exposure in epidemiologic studies is discussed first. Also, of interest is the issue of
confounding by temporal factors, meteorological factors, and copollutants. The shape of the
concentration-response function and heterogeneity of O3 effects are also discussed briefly.
All of these topics are important for characterizing and interpreting ambient O3-health effects
associations.
7.6.2 Ozone Exposure Indices
Three O3 indices were most often used to indicate daily O3 exposure: maximum 1-h
average (1-h max); maximum 8-h average (8-h max); and 24-h average (24-h avg). The
8-h max O3 concentration is a frequently used index in newer epidemiologic studies, as it relates
most closely to the averaging time of the current U.S. EPA NAAQS. The O3 exposure indices
are highly correlated, as indicated in several studies. For example, in the 21 European
multicities acute mortality study (Gryparis et al., 2004), 1-h max O3 was found to be highly
correlated with 8-h max O3, with a median correlation coefficient of 0.98 (range 0.91 to 0.99).
Among single-city studies, the 1-h max O3 and 8-h max O3 also were found to have correlation
coefficients ranging from 0.91 to 0.99 in various areas, such as Atlanta, GA (Tolbert et al., 2000;
White et al., 1994); southern New England (Gent et al., 2003); Ontario, Canada (Burnett et al.,
1994); and Mexico City (Loomis et al., 1996; Romieu et al., 1995). In addition, 1-h max O3 was
highly correlated with 24-h avg O3, as observed in the Mexico City study by Loomis et al. (1996)
(r = 0.77) and in the Ontario, Canada study by Burnett et al. (1994) (r = 0.87).
7-135
-------
All studies discussed in Sections 7.2 to 7.5 were examined for presentation of the three O3
exposure indices. Several presented the concentration data and correlations among 1-h max, 8-h
max, and 24-h avg O3 ambient measures. Some presented the associated risk estimates of
comparable analyses for the three exposure indices. No papers provided a statistical analysis
comparing results from the different indices. The increments used in this document to
standardize expressions of excess risks are 40 ppb for 1-h max O3, 30 ppb for 8-h max O3, and
20 ppb for 24-h avg O3, as discussed in Section 7.1.3.2. Key findings derived from the assessed
studies are discussed below, starting with two multicity mortality studies.
In the large U.S. 95 communities study by Bell et al. (2004), increases in O3-associated
daily mortality were estimated using all three O3 indices. The mean 24 h avg O3 level was
appropriately 26 ppb across the 95 communities. Ozone concentrations for 1-h max O3 and 8-h
max O3 were not provided. For the above-noted increments, the effect estimates calculated by
Bell et al. (2004) using all available data were 1.34% (95% PI: 0.84, 1.85), 1.28% (95% PI:
0.88, 1.73), and 1.04% (95% PI: 0.54, 1.55) excess risk in mortality for 1-h max O3, 8-h max O3,
and 24-avg O3, respectively. A statistical test examining differences among these risk estimates
indicated that there were no significant differences by exposure index. In the European study of
21 cities (two of the 23 cities evaluated did not have 8-h max O3 data), the O3-mortality effect
estimate for the summer season was slightly smaller for the 8-h max O3, 1.82% (95% CI: 0.99,
3.06) excess risk compared to the 1-h max O3, 2.59% (95% CI: 1.32, 4.10) excess risk; however,
the two risk estimates were not significantly different (Gryparis et al., 2004). The median 1-h
max O3 levels ranged from 44 to 117 ppb and the median 8-h max O3 levels ranged from 30 to
99 ppb across all cities during the summer.
Several single-city mortality studies examined multiple O3 exposure indices (Anderson
et al., 1996; Dab et al., 1996; Sunyer et al., 2002; Zmirou et al., 1996; Borja-Aburto et al., 1997).
These studies did not differentiate risk estimates by exposure index, because the results were
considered to be similar. Hospital admission studies also provided limited data for O3 index
comparisons. Schouten et al. (1996) found similar O3 effects on total respiratory hospitalizations
from 8-h max O3 and 1-h max O3 in the summer. Both indices resulted in a 4.0% excess risk per
standardized increment. For emergency department visits, the examples of Delfmo et al.
(1998b) and Weisel et al. (2002) indicated no difference in effect estimate when using various O3
indices. Tolbert et al. (2000) noted an increase in emergency room visits of 4.0% per standard
7-136
-------
deviation increase (approximately 20 ppb) for both 1-h max O3 and 8-h max O3 as being
expected, since the correlation between the indices was 0.99.
Peak flow asthma panel studies generally used only one index; thus, there were limited
data available for comparison. One respiratory symptom study (Gent et al., 2003) examined
both 1-h max O3 and 8-h max O3 but noted no differences in the results. Only one FEVj panel
study examined more than one O3 exposure index. Chen et al. (1999) examined 1-h max O3 and
24-h avg O3 and reported a decrement in FEVj of -25.6 mL (95% CI: -49.1, -2.1) per 40 ppb
increase in 1-h max O3 and -13.6 mL (95% CI: -33.2, 6.0) per 20 ppb increase in 24-h avg O3
in children at a 1-day lag. For 2- and 7-day lags, smaller differences were observed between the
two indices. The effect estimates calculated using 1-h max O3 and 24-h avg O3 concentrations
were not found to be significantly different for any of the lags examined.
Limited information is available by which to draw conclusions for comparison across the
three O3 indices of 1-h max O3, 8-h max O3, and 24-h avg O3. Studies conducted in various
cities have observed very high correlations among the daily O3 indices. For the same
distributional increment, the excess health risk estimates and significance of associations were
generally comparable for the three O3 indices across all outcomes. The high correlation among
the indices presents a challenge in distinguishing the most appropriate measure for
epidemiologic studies. Exploratory analyses using various O3 exposure indices are valuable in
understanding relationships. However, to address the issue of multiple hypothesis testing,
hypotheses that are confirmatory and exploratory should be decided a priori and reported
accordingly.
7.6.3 Confounding by Temporal Trends and Meteorologic Effects in
Time-Series Studies
The challenge of analyzing acute O3 effects in time-series studies is to avoid bias due to
confounding by daily to seasonal temporal factors. On a seasonal scale, the analysis must
remove the influence of the strong seasonal cycles that usually exist in both health outcomes
and O3. On a daily scale, weather factors and other air pollutants also may confound the
association of interest. This section discusses the interpretation of effect estimates after
adjusting for temporal trends and meteorologic effects.
7-137
-------
7.6.3.1 Assessment of Ozone Effects after Adjusting for Temporal Trends and
Meteorologic Effects
The relationship between O3 and health outcomes are significantly affected by temporal
trends and meteorological factors, namely temperature. Analyses of the association between
health outcomes and O3 concentrations using raw data, therefore, can be misleading. In Diaz
et al. (1999), a U-shaped relationship was observed between mortality and O3 concentrations,
and the negative portion of the slope was likely due to the opposing seasonal cycles in mortality
(high in winter) and temperature (low in winter).
Keatinge and Donaldson (2005) used two methods to identify confounding due to
meteorological factors in their study of the mortality effect of air pollutants on adults aged
65 years or older during the warm season in Greater London. One involved GAM analyses
controlling for up to seven weather covariates, and the other considered graphical comparisons
to show interrelations of mortality with weather factors and pollutants. Both methods
pre-adjusted for weather variables, which is a more stringent test of the air pollution effect as
compared to co-adjustment. Few of the individual analyses found a significant O3 effect, which
tended to be small in comparison to the weather effect. Therefore, Keatinge and Donaldson
(2005) concluded that, although air pollution may have a short-term adverse effect on health in
London during the warm season, the results suggested that measures to prevent excess mortality
in hot weather should focus on control of heat stress. Goldberg and Burnett (2003) reported a
positive slope for the temperature-mortality relationship being fitted most tightly in the mild
temperature range where mortality effects of temperature were not expected. While it is possible
that temperature has mortality effects in the mild temperature range, daily fluctuations of air
pollutants (especially O3) are strongly influenced by weather conditions and ascribing the
association between temperature and mortality entirely to effects of temperature may
underestimate the effects of air pollution.
Sensitivity analyses to examine confounding of temperature on the O3 effect were
performed in the U.S. 95 communities study by Bell et al. (2004). In analyses excluding days
with high temperature (that used a cutoff as low as 29 °C [85 °F]), the range of estimated effect
sizes (1.00 to 1.10% excess risk in mortality per 20 ppb increase in 24-h avg O3) did not differ
from the effect estimate using all the data (1.04% excess risk). Effect estimates were found to be
slightly higher at lower temperatures. In a related study of 19 U.S. cities, Huang et al. (2005)
7-138
-------
examined the potential influence of heat waves, modeled as a natural cubic spline of the
interaction between current temperature and average temperature in the three previous days, on
the O3 effect on cardiopulmonary mortality. The effect estimate was generally robust to
inclusion of this interaction term, with only a slight decrease in the estimate from 2.52% (95%
PI: 0.94, 4.10) to 2.23% (95% PI: 0.76, 3.75) excess risk per 20 ppb increase in 24-h avg O3.
Bell et al. (2004) and Huang et al. (2005) also found that O3 effects were robust to the
selection of degrees of freedom for smoothing of temporal trends. Bell et al. observed that
changing the degrees of freedom from 7 to 21 per year did not significantly affect the
O3-mortality estimates, with effect estimates ranging from 0.82 to 1.08% excess risk per 20 ppb
increase in 24-h avg O3 during the previous week. Huang et al. (2005) examined the sensitivity
of summertime O3 risk estimates to varying degrees of freedom from 4 to 16 per year. The
extent of change in the risk estimates, though varied from city to city (graphically presented),
was not substantial. Using more degrees of freedom in temporal trend fitting (i.e., controlling
shorter temporal fluctuations) means ascribing more details of daily health outcomes to
unmeasured potential confounders and possibly weakening real weather and air pollution effects.
However, results from these large multicity studies indicated that O3 effects were robust to
aggressive smoothing of temporal trends.
Ito et al. (2005) examined sensitivity of O3-mortality risk estimates to the extent of
temporal trend adjustment and to alternative weather model specifications using data from seven
U.S. cities (Cook County, IL; Detroit, MI; Houston, TX; Minneapolis, MN; New York City;
Philadelphia, PA; and St. Louis, MO). They found that varying the degrees of freedom from 4 to
26 per year did not substantially or systematically affect the O3-mortality estimates, except for
Cook County where the percent excess O3-mortality risk estimates were considerably reduced
when the temporal adjustment term with 26 degrees of freedom was applied. Ito et al. (2005)
noted that the O3 risk estimates were generally more sensitive to alternative weather models
(up to a 2-fold difference in percent excess risk) than to the degrees of freedom for temporal
adjustment. Of the four weather models examined, the four-smoother model similar to the one
used in the NMMAPS studies (Bell et al., 2004; Huang et al., 2005) generally resulted in the
smallest effect estimates.
Schwartz (2005) examined the sensitivity of the O3-mortality relationship to methods used
to control for temperature. Initially, temperature lagged 0 and 1 day was controlled using
7-139
-------
nonlinear regression splines with 3 degrees of freedom each. In a comparison analysis, control
days were restricted to a temperature-matched subset. The effect estimates for all-year data
using nonlinear regression splines (0.8% [95% CI: 0.1, 1.4] excess risk per 40 ppb increase in
1-h max O3) and temperature-matched controls (0.9% [95% CI: 0.04, 1.8] excess risk) were not
significantly different. Results were similar when restricting analysis to warm-season-only data.
Temporal cycles in daily hospital admissions or emergency department visits are often
considerably more episodic and variable than is usually the case for daily mortality. As a result,
smoothing functions that have been developed and tuned for analyses of daily mortality data
may not work as well at removing cyclic patterns from morbidity counts. Two methods are
commonly used to adjust for temporal trends. The pre-adjustment method involves applying the
adjustment to both outcome and air pollution variables prior to the regression analysis. The
co-adjustment method involves applying the adjustment as part of the regression analysis, by
fitting a function of time while simultaneously fitting the regression effect of air pollution and
weather factors. As shown in a hospital admissions study by Burnett et al. (2001), the
co-adjustment approach may lead to biased air pollution effect estimates in cases where both
outcome and pollution variables exhibit strong seasonal cycles (note that the data were analyzed
using Poisson GAM with default convergence criteria). Using year-round data, pre-adjustment
followed by regression analysis yielded a 14% (95% CI: 5, 24) increase in admissions per
40 ppb increase in 1-h max O3 with a multiday lag of 0 to 4 days. The co-adjustment method
resulted in a 7% (95% CI: 3, 11) decrease in admissions. When the authors limited the analysis
to the warm season (May-August), both methods yielded similar results (32% [95% CI: 21, 44]
versus 30% [95% CI: 17, 45] increase for co-adjustment and pre-adjustment, respectively),
implying that stratification by season can remove a significant amount of the confounding
seasonality (which also may include seasonally-varying population behavior and ventilation
conditions). This finding may be important to consider in reviewing the acute O3 mortality and
morbidity literature, because the vast majority of studies published over the past decade have
used the co-adjustment method. However, the use of pre-adjustment versus co-adjustment in
time-series studies is an unresolved issue. More empirical research in different locales is needed
to evaluate the merits of these two methods as far as O3 is concerned and to determine what
endpoints may be affected.
7-140
-------
More sensitivity analysis of O3 effect estimates to the extent of adjustment for temporal
trends and meteorological factors is needed, but perhaps it is equally as important to evaluate the
epidemiologic adequacy of a given adjustment. For example, do the fitted mortality series
sufficiently depict influenza epidemics? Or, when larger degrees of freedom (e.g., 12 degrees of
freedom per year) are used, what "unmeasured" confounders, other than weather and pollution,
are the investigators trying to adjust? Even in PM studies that conducted sensitivity analyses,
investigators rarely stated assumptions clearly or provided adequate discussions as to potential
reasons for the sensitivity of results.
Given their relationship to health outcomes and O3 exposure, adjusting for temporal trends
and meteorologic factors is critical to obtain meaningful O3 effect estimates. While the
prevailing analytical approaches fit the data flexibly, the estimated effects of meteorologic
variables and their impact on the adjusted O3 effects are not adequately discussed. More work is
needed in this area to reduce the uncertainty involved in the epidemiologic interpretation of O3
effect estimates.
7.6.3.2 Importance of Season-Specific Estimates of Ozone Health Effects
Analysis of O3 health effects is further complicated, as relationships of O3 with other
pollutants and with temperature appear to change across seasons. Moolgavkar et al. (1995)
examined the relationship between daily mortality and air pollution by season in Philadelphia,
PA for the period of 1973 to 1988. During the summer, there was a positive relationship
between O3 and TSP, as well as O3 and SO2. In contrast, the relationship of O3 with TSP
and SO2 was inverse during the winter. Ozone showed positive associations only in the summer
when the mean O3 concentration was the highest. The effect of O3 on mortality was negative
(though not significantly) in the winter when the mean O3 concentration was low. In the summer
multipollutant model, O3 was the only pollutant that remained significant. Similar results were
found in another Philadelphia study by Moolgavkar and Luebeck (1996). Neither studies
analyzed year-round data; therefore, the relationship between the excess risk estimates for all
year and each season could not be compared. The results from these studies, however, suggest
that year-round analyses may mask associations (positive or negative) that may exist in
particular seasons.
7-141
-------
Ito et al. (2005) examined O3-mortality associations in seven U.S. cities, but also described
the relationship between O3 and PM for summer months (June through August) and winter
months (December through February) in these cities (see Figure 7-28). The O3-PM relationships
were positive in the summer and negative in the winter in all of these cities, except in Houston,
where the O3-PM association was not clearly positive in the warmer months but positive in
colder months. Ito et al. (2005) found that O3-mortality associations were mostly weaker, null,
or even negative in the winter compared to the summer in most of these cities. Once again, the
exception was Houston where the cold season O3-mortality association was positive and larger
than those for year-round or warmer months. Findings from this study suggest the influence of
seasonal O3-PM relationships on O3-mortality associations. It should be noted that the contrasts
in O3 risk estimates between warm and cold months varied across the alternative weather models
examined, with the four-smoother model generally showing the least contrast.
In the analyses of the U.S. 90 cities data (of which 80 cities had O3 data available) by
Samet et al. (2000; reanalysis Dominici et al., 2003), the focus of the study was PM10, but O3 and
other gaseous pollutants also were analyzed in single- and multiple-pollutant models. In the
reanalysis (Dominici et al., 2003), O3 was associated with an excess risk of mortality in analyses
of all available data (0.4% [95% PI: 0.1, 0.7] excess risk per 20 ppb increase in 24-h avg O3 at a
1-day lag) and summer only data (1.0% [95% PI: 0.5, 1.6]; however, a negative association was
observed for the winter only analysis (-1.1% [95% PI: -2.2, 0.1]). A 2-fold greater effect was
estimated using summer data compared to all available data. The mean 24-h avg O3 levels
using all available data ranged from 12 to 36 ppb across the 80 cities. Season-specific O3
concentrations were not presented. It should be noted that the analyses by Samet et al. and
Dominici et al. used a weather model specification that is more detailed than other studies in that
it had multiple terms for temperature and dewpoint (these two variables are generally highly
correlated). Thus, it is possible that the high concurvity of O3 with these weather covariates may
have produced these conflicting results. Another possibility is that the apparent negative
relationship between O3 and mortality in the winter may have been due to confounding by PM.
In the larger U.S. 95 communities study by Bell et al. (2004), the all-available data and warm-
season-only analyses also indicated positive risk estimates (1.04% [95% PI: 0.54, 1.55] and
0.78% [95% PI: 0.26, 1.30] excess risk per 20 ppb increase in 24-h avg O3, respectively, using a
constrained distributed 7-day lag model), but the two estimates were similar in magnitude.
7-142
-------
New York City Cook County, IL Detroit, Ml Philadelphia, PA
0 1999-2002 „ 1985-1994 „ 1985-1989 0 1992-1995
tn
o
O.
Q.
"— ' 0
w cr)
O
O)
5 S
4
CM
O
*
,
+
0
<|)
0
o .
0 .
CO
o .
o .
,
+
^
^
*
*
$
* *
o
0 10 20 30 40 0 20 40 60 80 0 20 40 60 80 0 10 20 30 40
PM25(Mg/m3) PM10(Mg/m3) PM10 (Mg/m3) PM2 s (Mg/m3)
Houston, TX Minneapolis-St.Paul, MN St. Louis, MO
0 1985-1994 „ 1985-1994 0 1985-1994
If)
O ,
Q.
Q.
""• ' o .
m co
o
os
m o .
™ CM
f
4
CM
o .
+ 1 .
T
^
a\
-------
U.S. 95 communities study nearly doubled the study period by extending the analysis by six
additional years (1987 to 2000 versus 1987 to 1994) and it included 15 additional cities in
addition to the original 80. Also, the warm seasons are characterized differently in the two
studies. The U.S. 90 cities study defined summer as a three-month period of June through
August, whereas the 95 communities study defined the warm season as a 7-month period of
April through October. In addition, the results presented in the U.S. 90 cities study were from a
single-day lag model (1-day lag) while the estimates from the 95 communities study were
calculated using a constrained distributed 7-day lag model. Differences in seasonal O3 effects
observed in the two related studies might be attributable to some of these factors.
Many studies reported larger excess mortality risks in the warm (or summer) season than in
the cool (or winter) season (see Figure 7-21 in Section 7.4.5). These studies showed cool-season
risk estimates that were either smaller compared to warm-season estimates or slightly negative.
Of the studies that analyzed data by season, only one study in Pittsburgh, PA (Chock et al.,
2000) showed negative risk estimates for the summer. Ozone concentration data were not
provided for the Chock et al. (2000) study. The studies that observed larger, positive
associations between O3 and mortality in warm seasons are consistent with the expectation
that O3, if harmful, should have a stronger association with health outcomes for the summer
when exposure to O3 is higher. However, the negative O3-mortality associations seen for the
winter suggest that further examination of this issue is required. Specifically, if O3 levels in the
winter are shown to be negatively associated with factors (e.g., PM) that are positively
associated with mortality, then these potentially spurious negative O3-mortality associations can
be explained. Several examples of this phenomenon also exist in morbidity studies investigating
the effect of O3 on daily hospital admissions and emergency department visits (Anderson et al.,
1998; Burnett et al., 2001; Prescott et al., 1998; Thompson et al., 2001).
Unlike the time-series studies examining outcomes of mortality, hospital admissions, and
emergency department visits, most acute field studies did not perform year-round analyses.
The acute field studies that examined the relationship between O3 and lung function, respiratory
symptoms, and inflammation focused primarily on the O3 effect during the warm season
when O3 levels were expected to be high and subjects spent more time outdoors and were
physically active.
7-144
-------
There are seasonal (e.g., air conditioning use) or seasonally-modified (e.g., time spent
outdoors, air exchange rates) factors that affect the relationship between ambient concentrations
and personal exposures to O3, as discussed in Section 3.9. The influence of combinations of
these factors across seasons on air pollution health effects can become quite complex. For
example, longer time spent outdoors in the summer may increase personal exposure to O3 for
some segment of the population, but the increased use of air conditioning may reduce exposures
to ambient O3 for those who spend much of their time indoors. In the meta-analysis by Levy
et al. (2005), the combined risk estimate from the warm season was greater (3.38% [95% CI:
2.27, 4.42] per 40 ppb increase in 1-h max O3) compared to the estimate from all-year data
(1.64% [95% CI: 1.25, 2.03]). However, further analysis suggested that the O3-mortality risk
estimates were smaller in cities with high air conditioning prevalence. Seasonal factors such as
these that influence the relationship between ambient concentrations and personal exposures
make the interpretation of the concentration-response relationships obtained from analyses of
year-round data less straightforward.
In some cities, O3 is only monitored during the warm season. For example, 34% of the
communities in the U.S. 95 communities study only collected O3 data during the warm season
(Bell et al., 2004). The cities with larger populations and/or higher O3 concentrations generally
collected year-round data. There is some concern that differential data availability may also
contribute to the seasonal differences in O3 health effects observed.
The potential influence of season on O3 effect estimates was examined here using summary
density curves. The O3 effect observed in all-year data was compared to effects from warm-
season and cool-season only data (Figures 7-29 and 7-30). Summary probability density curves
were calculated to assess the effect estimates from the various studies (see Annex Section AX7.2
for further explanation of summary density curves). The summary density curves shown in
Figures 7-29 and 7-30 were smoothed by multiplying a constant to the standard error of each
effect estimate in the calculation of the individual distribution functions. Since the normal
distribution is unimodal, this constant will oversmooth when the density is multimodal.
In Figure 7-29, the summary density curves of O3-associated all-cause (nonaccidental) mortality
are presented (see Figure 7-21 in Section 7.4.5 for the effect estimates). The summary density
curves are calculated using results from 14 studies that reported at least two of the three
estimates. This figure indicates that 75% of the area under the density curve has a value greater
7-145
-------
=
"w
c
0)
Q
-5
% Change in Mortality
All year
Warm season
10
Cool season
15
% area under the density curve and >0 75% 78% 62%
Mean (SD) effect estimates 1.2% (2.8%) 1.3% (2.6%) 0.1% (3.1%)
Mode effect estimates 0.7% 0.9% 0.4%
Figure 7-29. Summary density curves of the percent change in all cause mortality for
all-year data and by season per standardized increment (see Section 7.1.3.2).
Effect estimates from 14 studies have been included in the summary density
curves (see Figure 7-21 in Section 7.4.5 for the effect estimates).
than zero for all-year data compared to 78% for warm-season data and 62% for cool-season data.
Therefore, both all-year and warm-season data generally indicate a positive O3 effect on
mortality. The mean effect estimates for all-year data and warm-season only data are 1.2%
(SD 2.8) and 1.3% (SD 2.6) excess risk in mortality per 40 ppb increase in 1-h max O3,
respectively. A slightly larger mode of effects is observed for warm-season data (0.9% excess
risk) compared to all-year data (0.7%). The cool-season-only data indicate that there is no
excess risk (mean 0.1% [SD 3.1]) associated with O3 concentrations.
Similar observations are made when examining the O3 effect on total respiratory hospital
admissions (Figure 7-30). Six studies provided season-specific estimates as well as all-year
results (see Figure 7-10 in Section 7.3.3 for the effect estimates). Once again, a large percent
of the area under the summary density curve is greater than zero when using all-year and
warm-season data, 92% and 84%, respectively, compared to cool-season data, 49%.
7-146
-------
0 15 30
% Change in Respiratory Hospital Admissions
45
All year
Warm season
Cool season
% area under the density curve and >0 92% 84% 49%
Mean (SD) effect estimates 6.5% (6.4%) 6.3% (9.1%) 0.8% (6.1%)
Mode effect estimates 1.8% 4.0% -1.3%
Figure 7-30. Summary density curves of the percent change in total respiratory hospital
admissions for all-year data and by season per standardized increment
(see Section 7.1.3.2). Effect estimates from six studies have been included
in the summary density curves (see Figure 7-10 in Section 7.3.3 for the
effect estimates).
The mean O3 effect estimates for warm-season-only data, 6.3% (SD 9.1) excess risk per 40 ppb
increase in 1-h max O3, and all-year analyses, 6.5% (SD 6.4) excess risk, are similar. A larger
mode of effects is observed for warm-season data (3.9% excess risk) compared to all-year data
(1.8% excess risk). A small O3 effect (0.8% [SD 6.1] excess risk) is observed when using cool-
season data only.
Integrating seasonal influences across the various health outcomes supports the view
that O3 effects are different in the cool and warm seasons, with greater effects observed during
the warm season. Warm-season data should be used to derive quantitative relationships for the
effect of O3 on health outcomes, because the attenuation of O3 exposure from ambient
concentrations is likely to be much less during this season. However, studying summer data
7-147
-------
only when all-year data are available weakens the power of the study, since less data are
analyzed. In addition, increased adverse health outcomes are observed in the winter, some of
which may be attributable to O3. The O3 effect in the wintertime may be masked by the effects
of PM due to the negative correlation between these two variables (see Section 7.6.4.2 for
further discussion). Therefore, analysis of all-year data may be improved by adjusting for PM
indices in addition to adequate adjustment of meteorological factors and temporal trends.
Seasonality influences the relationship between O3 and health outcomes, as it may serve as
an indicator for time-varying factors, such as temperature, copollutant concentrations,
infiltration, and human activity patterns. Given the potentially significant effect of season, O3
effect estimates computed for year-round data need to be interpreted with caution. Finding small
or no effects may simply reflect (a) the cancellation of positive associations in the summer and
negative associations in the winter or (b) the presence of confounding due to the strong seasonal
character of O3 concentrations.
7.6.4 Assessment of Confounding by Copollutants
Potential confounding by daily variations in copollutants is another analytical issue to be
considered. With respect to copollutants, daily variations in O3 tend to not correlate highly with
most other criteria pollutants (e.g., CO, NO2, SO2, PM10), but they may be more correlated with
secondary fine PM (e.g., PM2 5, sulfates) measured during the summer months. Assessing the
independent health effects of two pollutants that are correlated over time is not straightforward.
If high correlations between O3 and PM or other gaseous pollutants exist in a given area, then
disentangling their relative individual contributions to observed health effects associations
becomes very difficult. The changing relationship between O3 and other copollutants also is at
issue. Ito et al. (2005) described the relationship between O3 and PM by season in seven urban
U.S. cities: Cook County, IL; Detroit, MI; Houston, TX; Minneapolis-St. Paul, MN; New York
City; Philadelphia, PA; and St. Louis, MO (see Figure 7-28). With the exception of Houston,
TX, the O3-PM associations were positive in the summer and negative in the winter.
Relationships between O3 and copollutants and the potential confounding of the O3 effect by
copollutants are discussed in the following section.
7-148
-------
7.6.4.1 Relationship between Personal Exposure to Ozone and Copollutants
The correlation between ambient O3 concentrations and ambient levels of PM, NO2, SO2,
and CO measured at central monitoring sites has been reported in many studies; however, only a
limited number of studies have examined the association between personal O3 exposures and
personal exposures to other copollutants. The relationship between personal exposure to O3 and
personal exposure to PM2 5 was examined by Chang et al. (2000) in various microenvironments
in Baltimore, MD, using activity data from the National Human Activity Pattern Survey study
(Klepeis, 1999). Activities were scripted to simulate activities performed by older adults
(65+ years of age). Mean hourly personal O3 exposures were 15.0 ppb (SD 18.3) in the summer
and 3.6 ppb (SD 7.5) in the winter. Strong correlations were not observed between 1-h personal
O3 and PM2 5 concentrations in the various microenvironments, including the indoor residence
(n = 91), other indoor environments (n = 53), outdoor near the roadways (n = 21), outdoor away
from roads (n = 19), and in vehicles (n = 71). Spearman r ranged from -0.14 to 0.29 during the
summer and -0.28 to 0.05 during the winter (SEs not provided, all correlation coefficients noted
as statistically nonsignificant [p > 0.05]).
An issue of particular interest is the correlation between personal exposure to O3 and
personal exposure to the ambient component of PM2 5. In a study of susceptible populations
(older adults [n = 20], individuals with COPD [n = 15], and children [n = 21]) in Baltimore, MD,
a total of 800 person-days of exposure data were collected for the following pollutants: PM2 5,
PM10, O3, NO2, SO2, elemental carbon, organic carbon, and volatile organic compounds (Sarnat
et al., 2001). A subset of PM25 filters was analyzed for SO42" concentration to estimate the
personal exposure to PM25 of ambient origin. The mean ambient 24-h avg O3 concentration was
approximately 36 ppb. Sarnat et al. (2001) found that ambient 24-h avg O3 concentrations and
ambient 24-h avg PM2 5 levels, both measured at centrally located monitoring sites, were
positively associated in the summer (P = 0.84 [95% CI: 0.56, 1.12], Spearman r = 0.67
[p < 0.05]) and negatively associated in the winter (P = -0.67 [95% CI: -0.91, -0.43],
Spearman r = -0.72 [p < 0.05]). An association also was observed between ambient O3
concentrations and personal PM2 5 of ambient origin, with a mixed regression effect estimate of
P = 0.37 (95% CI: 0.25, 0.49) in the summer and B = -0.36 (95% CI: -0.41, -0.31) in the
winter. However, no significant relationship was found between 24-h avg personal O3 exposure
7-149
-------
and personal exposure to PM25 of ambient origin in either season (P = 0.22 [95% CI: -0.06,
0.50] in the summer and P = -0.18 [95% CI: -0.39, 0.03] in the winter).
In a related study conducted in Boston, MA, 20 healthy senior citizens and 23 school
children were monitored during six sessions for a total of 714 person-days in both the summer
and winter seasons (Sarnat et al., 2005). Personal SO42" exposures were used as indicators of
personal exposure to PM25 of ambient origin. Mean ambient 24-h avg O3 levels ranged from
22.7 to 31.6 ppb during the three summer sessions and 14.0 to 21.8 ppb during the three winter
sessions. Similar to the earlier study conducted in Baltimore, ambient 24-h avg O3 levels and
ambient 24-h avg PM25 levels were positively associated in the summer (P = 0.51 [95% CI:
0.34, 0.68]) but negatively associated in the winter (P = -0.53 [95% CI: -0.22, -0.85]).
In addition, an association was observed between ambient O3 concentrations and personal PM2 5
of ambient origin, but only in the summer (P = 0.24 [95% CI: 0.20, 0.28]). However, in contrast
to the Baltimore results, a significant association between 24-h avg personal O3 exposure and
personal exposure to PM25 of ambient origin was observed both in the summer (P = 0.35 [95%
CI: 0.22, 0.47]) and in the winter (P = 0.07 [95% CI: 0.01,0.13]). Neither of the Sarnat et al.
(2001, 2005) studies evaluated relationships between personal O3 exposures and personal
exposure to other gaseous pollutants.
Results from the two related Sarnat studies provide initial evidence that the relationship
between personal O3 exposure and personal exposure to PM2 5 of ambient origin may differ by
city/region and season. The authors noted that their results were for a small, nonrandom
selection of subjects living in the eastern United States, therefore caution should be exercised in
generalizing the results to other locations and cohorts. The results are likely to vary for locations
depending on indoor air exchange rates and amount of time spent outdoors.
7.6.4.2 Assessment of Confounding Using Multipollutant Regression Models
Multipollutant regression models are generally used to determine whether the pollutant-
specific effect is robust. However, because of the multicollinearity among O3 and pollutants and
the changing correlations by season, multipollutant models may not adequately adjust for
potential confounding, especially when using year-round data. These limitations need to be
considered when evaluating results from multipollutant models. Results from the U.S. 90 cities
study, which included 80 cities with O3 data, indicated that while the addition of PM10 in the
7-150
-------
model did not substantially change the O3-mortality risk estimates, slight declines in the O3
effect were observed, as shown in Figure 7-31 (Samet et al., 2000; reanalysis Dominici et al.,
2003). In the extended U.S. 95 communities study (Bell et al., 2004), the city-specific O3-
mortality effects were robust to adjustment for PM10, as indicated by the nearly 1:1 ratio between
estimates with and without PM10 adjustment shown in Figure 7-32. This finding suggested
that PM10 generally did not confound the association between O3 and mortality. Limited data
were available to examine the potential confounding effect of PM25 on the O3-mortality
relationship. A weighted second-stage linear regression indicated that there was no association
between long-term PM2 5 average and the community-specific O3-mortality effect estimate.
Several other mortality and morbidity studies have investigated confounding of O3 risk estimates
using multipollutant models with year-round data, and most have reported that O3 effects were
robust to adjustment for copollutants (see Figures 7-11 and 7-22 in Sections 7.3.3 and 7.4.6,
respectively).
0
O)
c
TO
O
1.0-
U.b-
n.
U .0 "
1.0-
t «
•
La
•
»
<
,
go
i
. <
^
i
i
> i
i
i
•
La
i I
! i <
,
i
91
i
4
i
;
1 V
;
•
r *
*
L<
1
> «
I
1
ig2
^ <
i
» <
i
i
ABCDEABCDEABCDE
Models
Figure 7-31. Posterior means and 95% probability intervals for the national average
estimate of O3 effects on total mortality from non-external causes per 10 ppb
increase in 24-h avg O3 at 0-, 1-, and 2-day lags within sets of 80 U.S. cities
with pollutant data available. Models A = O3 only; B = O3 + PM10; C = O3 +
PM10 + NO2; D = O3 + PM10 + SO2; E = O3 + PM10 + CO.
Source: Derived from Dominici et al. (2003).
7-151
-------
O)
re
re
Q
Q.
a
•o
15 -
10 -
5 -
0 -
-5 -
= National Average Effect
-5
10
15
Without PM10 Adjustment
(using only days with PM10 data)
Figure 7-32. Maximum likelihood estimates of O3-mortality for 95 U.S. communities,
determined using a constrained distributed lag model for lags 0 through
6 days. Same data set was used for O3 estimates with and without
adjustment for PM10.
Source: Derived from Bell et al. (2004).
The pollutant most correlated with O3 in the summer is sulfate (which is in the fine particle
size range), especially in the eastern United States. Therefore, the main potential confounders of
interest for O3 in the summer are PM2 5 and sulfate. Once again, the results from two-pollutant
regression models with O3 and sulfate (or PM2 5) should be interpreted with caution because both
of these pollutants are formed under the same atmospheric conditions and both are part of the
"summer haze" pollution mix. A simple two-pollutant regression model does not address their
possible synergistic effects, and the high correlation between the two pollutants may lead to
unstable and possibly misleading results. In any case, most studies that analyzed O3 with PM
indices did not have PM2 5 data and very few examined sulfate data. A mortality study by Lipfert
7-152
-------
et al. (2000a) found that all-year O3 risk estimates were not affected by the addition of sulfate
(3.2% versus 3.0% with sulfate per 45 ppb less background level of 1-h max O3). Among
studies with PM25 data, including Santa Clara County, CA (Fairley, 1999; reanalysis Fairley,
2003), Philadelphia, PA (Lipfert et al., 2000a), and Detroit, MI (Lippmann et al., 2000;
reanalysis Ito, 2003), most examined copollutant models for year-round data only, but
O3-mortality risk estimates were not substantially affected by the addition of PM2 5. An analysis
of Philadelphia and Detroit data by season suggested that O3-mortality risk estimates were not
sensitive to adjustment for PM2 5 in all-year or seasonal analyses (Ito et al., 2005). Including
both O3 and PM2 5 in the regression models for these cities tended to attenuate both pollutants'
coefficients, but not substantially. The models with both O3 and PM2 5 in these cities often
showed better fits (i.e., lower Akaike's Information Criteria) than those with either pollutant
alone. Ito et al. (2003) concluded that these results suggest that O3 and PM2 5 contribute
independently to mortality.
Other studies have estimated O3 health risks with copollutants in the model by season.
Respiratory hospitalization studies conducted during the warm season in Canada observed
consistent O3 risk estimates with the inclusion of PM25 in the model (Burnett et al., 1997b,
2001). In one of these studies (Burnett et al., 1997b), the effect of O3 also was adjusted for
sulfate. With the addition of sulfate in the model, the risk estimate for O3 on respiratory
hospitalizations remained relatively stable, from a 14.4% (95% CI: 8.7, 20.5) excess risk to a
11.7% (95% CI: 5.6, 18.0) excess risk per 25 ppb increase in 12-h avg O3 at a 1- to 3-day lag.
In contrast, the effects for sulfate were reduced in half after adjusting for O3. Amongst the
mortality studies (see Figure 7-23 in Section 7.4.6), adjusting for copollutants, in particular PM
indices, did not substantially change the warm-season O3-mortality effect estimates, with both
slight reductions and increases observed in the adjusted estimates. In the analysis using cool-
season-only data, the O3 effect estimates were generally negative, but none were statistically
significant. The O3 risk estimates all increased slightly with the adjustment of PM indices.
The inverse relationship between O3 and PM during the cool season most likely influenced
the O3-mortality effect estimates in the single-pollutant models.
In field studies, power to assess independent O3 effects may be limited by small sample
sizes and short follow-up times. Yet, the O3 effect also was robust to the addition of copollutants
in multipollutant models, with a few exceptions. For example, the effect of O3 on PEF was not
7-153
-------
robust to adjustments for PM25 and sulfate, in studies by Romieu et al. (1996) and Neas et al.
(1999). In general, however, O3 effects on respiratory symptoms (Romieu et al., 1996), lung
function parameters (Brauer et al., 1996; Gold et al., 1999), and asthma medication use (Gent
et al., 2003) were robust to inclusion of PM2 5. Further, the effects for O3 were observed to be
stronger than those for PM.
Multipollutant regression analyses indicated that O3 risk estimates, in general, were not
sensitive to the inclusion of copollutants, including PM2 5 and sulfate. These results suggest that
the effect of O3 on respiratory health outcomes appears to be robust and independent of the
effects of other copollutants.
7.6.5 Concentration-Response Function and Threshold
An important consideration in characterizing the public health impacts associated with O3
exposure is whether the concentration-response relationship is linear across the full
concentration range or instead shows nonlinearity. Of particular interest is the shape of the
concentration-response curve at and below the level of the current 8-h standard of 80 ppb. There
are limitations to identifying possible "thresholds" in air pollution epidemiologic studies,
including difficulties related to the low data density in the lower concentration range, possible
influence of measurement error, and individual differences in susceptibility to air pollution
health effects. Despite these difficulties, the slope of the O3 concentration-response relationship
has been explored in several studies.
To examine the shape of the concentration-response relationship between O3 and mortality,
Gryparis et al. (2004) used meta-smoothing to combine smooth curves across the 23 European
cities in a hierarchical model. The mean 1-h max O3 concentrations ranged from 44 ppb in
Tel Aviv to 117 ppb in Torino, Italy during the summer. For the summer period, while the
estimated concentration-response curve did not appear to deviate significantly from linearity,
there were indications of decreasing effects at lower exposures.
In the U.S. 95 communities study (Bell et al., 2004), effect estimates calculated using only
days with 24-h avg O3 levels less than 60 ppb were compared to those using all data. At a lag of
1 day, O3 was associated with an excess risk of 0.36% (95% PI: 0.12, 0.60) per 20 ppb increase
in 24-h avg O3 using data from all days and only a slightly smaller risk of 0.30% (95% PI: 0.08,
0.54) when data were limited to days less than 60 ppb. These results suggest that if there is a
7-154
-------
threshold, it must be notably lower than a 24-h avg O3 of 60 ppb. In a more recent study by Bell
et al. (2006; published online January 23, 2006), the shape of the concentration-response curve
for the O3-mortality relationship was evaluated in 98 U.S. urban communities for the period
1987 to 2000. The measure of exposure was the average of the same and previous days' ambient
24-h avg O3 concentration. Analytical methods included linear, subset, threshold, and spline
models. Results from all methods indicated that if a threshold did exist, it would have to be at
low concentrations, less than a 24-h avg O3 of 15 ppb.
Fairley (2003) reanalyzed the Santa Clara County mortality data using GAM with stringent
convergence criteria and examined a new exposure index for O3. He noted O3 concentrations
exceeding 60 ppb each hour and calculated a daily sum of these exceedances. Fairley's index
incorporates measures of concentration and exposure duration; thus, this index represents a
linear time-integrated concentration, also known as dosage. The O3 index with the 60 ppb
"threshold level" was found to be significantly associated with mortality in single-pollutant
models as well as in multipollutant models. Two other "threshold levels" were examined,
40 ppb and 80 ppb. Both produced statistically significant results in single-pollutant models.
These results suggest that the threshold for O3-mortality effects, if it exists, is likely less than
40 ppb. The implication for thresholds in terms of the three standard indices (i.e., 1-h max, 8-h
max, and 24-h avg) is unclear, but there may be an empirical relationship.
Vedal et al. (2003) observed that the annual mean 1-h daily max O3 concentration of
27.3 ppb (SD 10.2) in Vancouver, Canada, was lower than that in any of the 80 NMMAPS cities
(Samet et al., 2000); thus, a Vancouver study might provide a better focus on the shape of the O3
concentration-response curve at lower levels. An O3 effect was observed on total mortality at a
0-day lag during the summer. Ozone effects on respiratory mortality at a 2-day lag and
cardiovascular mortality at a 0-day lag also were observed in the summer in this study. The
effect of O3 on mortality was robust in two-pollutant models. Vedal et al. (2003) concluded
that O3 concentrations were associated with adverse effects on mortality even at low levels.
Although this study appears to support the argument that there are no threshold concentrations
below which adverse effects cannot be detected, the results must be interpreted with caution as
concerns remain. Vedal et al. (2003), for example, questioned if O3, other gaseous pollutants,
and PM may be acting as surrogate markers of pollutant mixes that contain more toxic
compounds, since the low measured concentrations were unlikely, in their opinion, to cause the
7-155
-------
observed effects. They further stated that measurement error and interference by meteorological
factors might have contributed to the inability to detect a threshold.
Kim et al. (2004) investigated the presence of a threshold in O3-mortality effects in Seoul,
Korea by analyzing data using a log linear GAM (linear model), a cubic natural spline model
(nonlinear model), and a B-mode splined model (threshold model). Models were stratified by
season and adjusted for PM10, long-term time trend, and meteorological variables. The mean 1-h
max O3 levels were 46.9 ppb (SD 22.5) during the summer and 21.3 ppb (SD 6.9) during the
winter. Threshold values were estimated as 1-h max O3 levels of 28 ppb when using all-year
data and 45 ppb for summer-only data. None of the other pollutants examined, including PM10,
SO2, NO2, and CO, had a nonlinear association with mortality. Using summer-only data, the
B-spline model resulted in an excess mortality risk of 7.1% (95% CI: 3.1, 11.2) per 40 ppb
increase in 1-h max O3, compared to an excess risk of 3.6% (95% CI, 0.5, 6.8) calculated using
the log linear model. If a threshold truly exists, results from the Kim et al. (2004) study suggest
that the use of log-linear models may underestimate the O3 effect on mortality at levels above the
threshold.
Other studies examining the effect of O3 on mortality also have found suggestive evidence
for a possible threshold level. In a London, England study (Anderson et al., 1996), an adjusted
O3-mortality bubble plot suggested that a threshold might exist around 50 ppb for 8-h avg O3.
A study by Simpson et al. (1997) in Brisbane, Australia observed a significant excess risk in
mortality only in the highest quintile of O3 exposure, which had a mean concentration of 42 ppb
for 1-h max O3.
Among several studies with morbidity outcomes, examination of the shape of the
concentration-response function indicated evidence of an effect threshold. In a study of all-age
respiratory hospital admissions in Toronto, Canada, effects of O3 appeared to become apparent
only above an approximate 30 ppb daily 1-h max O3 (Burnett et al., 1997b). Also, Ponce de
Leon et al. (1996) reported an indication of a threshold in the O3 effect on hospitalizations in
London, England at 40 to 50 ppb for 8-h max O3 and 50 to 60 ppb for 1-h max O3. In a study of
emergency department visits for asthma in St. John, Canada, effects observed in the over
15 years age group were apparent only when data above the 95th percentile (75 ppb daily 1-h
max O3) were included (Stieb et al., 1996). However, other morbidity studies observed a
monotonic increase in the concentration-response function, suggesting that there was no
7-156
-------
threshold in O3 effects on hospitalizations and emergency department visits (Burnett et al.,
1997a; Jaffe et al., 2003; Petroeschevsky et al., 2001; Tenias et al., 1998).
In a field study by Mortimer et al. (2002), the associations of ambient O3 levels with PEF
and asthma symptoms were investigated in eight urban cities in the United States. The mean 8-h
avg O3 was 48 ppb, with less than 5% of days exceeding 80 ppb. Analysis performed using all
data indicated that a 15 ppb change in 8-h avg O3 was associated with decrements in PEF
(-0.59% [95% CI: -1.05, -0.13]) and increased incidence of respiratory symptoms (odds ratio
of 1.16 [95% CI: 1.02, 1.30]) over multiday lag periods. When data were restricted to days
when ambient O3 concentrations were less than 80 ppb, the O3 effects persisted, with a
significant PEF decline (-0.70% [95% CI: -1.29, -0.12]) and incidence of morning symptoms
(odds ratio of 1.17 [95% CI: 1.01, 1.35]). A study by Chen et al. (1999) also found that there
was no clear threshold in the O3 effect on FEVj and FVC in Taiwanese school children. The 1-h
max O3 concentration ranged from 19.7 to 110.3 ppb.
The studies of both Brauer et al. (1996) and Korrick et al. (1998) demonstrate that
exposure duration and exercise level, in addition to O3 concentration, must be considered when
evaluating thresholds. In the study by Brauer et al., the mean O3 concentration during the
11-hour work shift was 26.0 ppb (SD 11.8). Workers experienced a change of -180.0 mL (95%
CI: -227.0, -133.0) in FEVj levels the next morning per 40 ppb increase in 1-h max O3. The
hikers in the study by Korrick et al. (1998) were exposed to mean O3 levels of 40 ppb (SD 12)
over the duration of their hike (mean 8 hours). Korrick et al. observed a mean change of
-62.5 mL (95% CI: -115.3, -9.7) in pre-hike to post-hike FEVj per 30 ppb increase in 8-h
avg O3 when all hikers were included in the analysis; however, when analysis was restricted to
hikers with wheeze or asthma, a larger change of-182.5 mL (95% CI: -312.2, -52.9) was
observed. In both studies, large reductions in lung function were observed in subjects exposed to
relatively low levels of O3 over multiple hours while active outdoors.
Other studies that may provide information concerning the concentration-response
relationship are those that reported larger excess health risks in the warm (or summer) season
than in the cold (or winter) season (as discussed in Section 7.6.3.2). During the cold season,
when O3 concentrations are generally low, no consistent O3 effects were observed across the
various health outcomes. These results appear to provide epidemiologic evidence of the
presence of a threshold in O3 health effects. However, other factors also may contribute to the
7-157
-------
observed lack of an association between O3 and health outcomes in the cold season. First,
potential confounding by copollutants may mask the effect of O3 during the cool season. For
example, O3 levels have been found to be inversely related to PM concentrations during the
winter in some cities, e.g., New York City and Philadelphia (Ito et al., 2005). If PM is also
associated with the health outcome of interest, inadequate control of confounding of the PM
effect would result in negative or null associations for O3 during the winter. In the few mortality
studies that adjusted for PM in season-stratified analyses (see Figure 7-23), the cool-season O3
risk estimates all increased slightly with the adjustment of PM indices; however, it should be
noted that none achieved statistical significance. A second possible reason for the difference
in O3 effects observed by season may be the changing relationship between O3 concentrations
and personal exposure across seasons. The ambient O3 levels are lower in the cold season, but
people are likely to be exposed to even lower levels of O3 during this season because of the
shorter time spent outdoors and the longer time spent indoors with closed windows. Sarnat et al.
(2005) observed that a 1 ppb increase in ambient O3 concentration was associated with a
0.27 ppb (95% CI: 0.18, 0.37) increase in person O3 exposure during the summer and 0.04 ppb
(95% CI: 0.00, 0.07) increase during the winter.
A more "representative" concentration-response relationship may need to be examined in a
summer-only data set as personal exposures tend to be better represented by ambient O3
concentrations during that season. Even for summer data, however, an interpretation of the
relationship is not straightforward because of the possible influence of the use of air conditioning
(an effective remover of O3). Greater use of air conditioning is expected on hot days when
the O3 level is higher, but the use of air conditioning may also vary from city to city and across
social class within a city. Brauer et al. (2002) observed that surrogate measures of exposure (i.e.,
those from centrally-located ambient monitors) that were not highly correlated with personal
exposures obscured the presence of thresholds in epidemiologic studies at the population level,
even if a common threshold exists for individuals within the population.
Obscuring of thresholds would be even greater if thresholds varied across individuals
(Brauer et al., 2002). Effects occur at the molecular level and are eventually manifested at the
clinical level. One possible molecular mechanism for adverse physiologic effects from O3 is its
reaction with the carbon double bonds of lipids in the lung lining fluid (Levy et al., 2001). The
molecular effects of O3 are partially mitigated by the antioxidant defense system, which varies in
7-158
-------
individuals with age, diet, genetic factors, and disease status. This raises the possibility that
there may be a level below which O3 would have few or no adverse effects as well as
contributing to individual variability in the threshold level. The antioxidant defense system is
one of many factors that contribute to the large intersubject variability in lung function responses
to O3 reported in human clinical studies (discussed in Sections 6.4 and 6.5). As individual
sensitivities to O3 health effect vary, a common threshold may not be observable at the
population level in epidemiologic studies.
As noted earlier, demonstrating a clear threshold at a population level is difficult due to
low data density in the lower O3 concentration range, measurement error resulting from person-
to-person differences in the relationship between personal exposures and monitored ambient
concentrations, and individual differences in susceptibility to air pollution health effects. From
1990 to 2004, the 10th percentile values (which represent the lower concentration range) of the
nationwide mean daily 8-h max O3 concentrations were approximately 40 ppb during the warm
season (May to September) (see Figure 3-17 in Section 3.2). While no fully confident
conclusion can be made regarding the threshold issue from epidemiologic studies alone, the
limited currently available evidence suggests that if a population threshold level exists in O3
health effects, it is likely near the lower limit of ambient O3 concentrations in the United States.
7.6.6 Heterogeneity of Ozone Health Effects
As described in Chapter 3 of this AQCD, O3 concentrations tend to be more spatially
variable than PM2 5 concentrations in urban areas. In addition, relative personal exposures to O3
likely vary by region. The geographic variability in O3 concentrations and personal exposures
may contribute to the heterogeneity in observed O3 health effects. The degree of influence of the
geographic variability on heterogeneity in effects will vary by study, as study design affects
different aspects of exposure (e.g., time period and duration of exposure).
More than 80% of the O3-mortality estimates from the various studies conducted in North
America, South America, Europe, and Australia were between 0.5 and 5% excess risk per 40 ppb
increase in 1-h max O3 using year-round data. In general, the O3-mortality estimates were
greater when using summer only data compared to year-round data. Though not all statistically
significant, most of the O3-mortality estimates were greater than zero, indicating a positive
relationship between O3 exposure and mortality. The O3 risk estimates from the numerous
7-159
-------
hospitalization and emergency department visit studies were generally larger in magnitude and
more variable from study to study compared to the mortality studies. These differences in
the O3 effect estimates may be attributable to the greater variability in the outcome measure in
hospitalization studies compared to mortality studies, as a result of the use of more subcategories
of outcome and varying degrees of severity.
Three recent meta-analyses that included both U.S. and non-U.S. studies found consistent
all-year combined point estimates: 1.75% (95% PI: 1.10, 2.40), 1.6% (95% CI: 1.1,2.0), and
1.64% (95% CI: 1.25, 2.03) per 20 ppb increase in 24-h average O3, for Bell et al. (2005),
Ito et al. (2005), and Levy et al. (2005), respectively. Bell et al. further observed that the pooled
estimate for U.S. studies (11 estimates), 1.69% (95% PI: 0.94, 2.78), was similar to the pooled
estimate for the non-U.S. studies (30 estimates), 1.85% (95% PI: 0.94, 2.78). Levy et al.
compared North American studies to European studies and also found nearly identical effect
estimates.
As differences in study design, population, and data analysis may affect risk estimates,
studies that were conducted in multiple cities using standardized methods were further examined
to investigate the geographic heterogeneity of O3 effects. Bell et al. (2004) conducted a time-
series analysis of O3 and mortality in 95 U.S. communities from 1987 to 2000. A 20-ppb
increase in 24-h avg O3 levels in the previous week was associated with an increase of 1.04%
(95% PI: 0.54, 1.55) excess risk of mortality in the pooled analysis of 95 communities using all
available data. The median 24-h avg O3 concentrations varied from 14.4 ppb in Newark, NJ to
37.3 ppb in Bakersfield, CA. Intercommunity heterogeneity in O3 effects was observed among
the 95 communities (see Figure 7-17 of Section 7.4.3), which the authors noted as plausible
given the community-specific differences in pollution characteristics, the use of air conditioning,
time-activity patterns, and socioeconomic factors. One factor that appears to explain some of the
intercity heterogeneity in the 95 U.S. communities study is long-term O3 concentration.
A weighted regression of the community-specific mortality risk from acute O3 exposure and
long-term average 24-h avg O3 was significant (p = 0.01). These results indicated a higher
excess risk of mortality per incremental change in 24-h avg O3 in the previous week was
observed in communities with lower long-term average O3. Note that this analysis only includes
the 40 communities with warm-season data. This relationship can be explained in a three
parameter Gompertz concentration-response model of the cumulative percent excess risk in
7-160
-------
mortality by long-term average 24-h avg O3 levels (Figure 7-33). The method of estimating this
curve is discussed in Annex Section AX7.3 along with limitations to inference concerning a
threshold. The curve has shallower slopes above the median effective concentration (inflection
point) than below that concentration point. The horizontal line is the 95% confidence interval on
the long-term average 24-h avg O3 associated with a 0.05% excess risk in mortality (EC005o/0).
The EC0 05o/0 is a risk that is believed to be low enough to be in the noise band. The Gompertz
model presents a threshold of effects at low concentrations and a flat curve at higher
concentrations. Thus, to the extent that the estimated concentration associated with a 0.05%
excess risk is realistic, the fit of the model supports the observed relationship between greater
mortality risk per incremental change in acute O3 exposure at lower long-term O3 levels.
1.0
go.8
(0
o>
O r, ~
X, 0.6
HI
> 0.4
3
| 0.2
O
10 20 30
Long-Term Average 24-h avg O3 (ppb)
40
Figure 7-33. The fitted Gompertz model of the cumulative percent excess risk in mortality
by long-term average 24-h avg O3 concentrations using data from the 95 U.S.
communities study (Bell et al., 2004). Only the 40 communities with warm-
season data are included. Highlighted is the actual data range of the long-
term average 24-h avg O3 concentrations. The horizontal line is the 95% CI
on the long-term average 24-h avg O3 associated with a 0.05% excess risk in
mortality (EC0.OS%).
7-161
-------
Greater heterogeneity was observed in the European study of 23 cities in 14 countries
(Gryparis et al., 2004). In the year-round analyses, only 8 of the 23 cities had positive
O3-mortality effect estimates. However, in the analyses using summer-only data, the risk
estimates were positive in 19 of the 23 cities, with a range of 0.8 to 8% excess risk per 40 ppb
increase in 1-h max O3. Median 1-h max O3 concentrations ranged from 44 ppb in Tel Aviv,
Israel to 117 ppb in Torino, Italy during the summer. The heterogeneity may be attributable to
the considerable variability among countries in factors that may influence the relationship
between ambient O3 concentrations and personal exposure to O3, such as climate, use of air
conditioning, personal activity patterns, and socioeconomic factors. In addition, the variability
in the concentration and composition of copollutants by cities or countries may contribute to the
heterogeneity in the O3-mortality effects. For example, concentrations of NO2 may vary widely
by region, depending on the differences in traffic density.
Among the hospitalization studies, Burnett et al. (1997a) conducted the largest study of
16 Canadian cities. The mean daily 1-h max O3 was 31 ppb (range 26-38) in the 16 cities. The
pooled O3 estimate was 5.6% (95% CI: 3.4, 7.9) excess risk in respiratory hospitalization per
40 ppb increase in 1-h max O3 using warm-season data (April to December). The risk estimates
were fairly homogenous across the 16 Canadian cities, ranging from 3.1% for Vancouver to
7.7% for Quebec City.
Anderson et al. (1997) investigated the association between O3 and hospital admissions for
COPD in five European cities (London, Paris, Amsterdam, Rotterdam, and Barcelona). The
pooled effect estimate was 5.0% (95% CI: 2.6, 7.6) excess risk per 30 ppb increase in 8-h
max O3 for year-round data. Results from the APHEA study showed similar variability to that
from the Burnett et al. (1997a) study. The year-round effects estimates were lower in the two
Dutch cities (2.5% excess risk) compared to that in Paris (7.7% excess risk); however, analyses
indicated that there was no significant heterogeneity in effects by city. The authors further noted
that among the pollutants examined (O3, BS, TSP, SO2, and NO2), O3 had the most consistent
and significant findings.
Among the field studies, various respiratory health outcomes were examined, including
PEF, spirometric parameters, respiratory symptoms, and medication use. Only one field study
investigated the O3 effect in several locations (Mortimer et al., 2002). Mortimer et al. (2002)
investigated the association of ambient O3 concentrations with PEF and asthma symptoms in
7-162
-------
asthmatic children living in eight urban cities in the United States: St. Louis, MO; Chicago, IL;
Detroit, MI; Cleveland, OH; Washington, DC; Baltimore, MD; East Harlem, NY; and Bronx,
NY. In the analysis pooling data from all eight cities, a 30 ppb increase in 8-h avg O3 was
associated with a decrement of-1.18% (95%CI: -2.10, -0.26) in morning PEF for a 5-day
cumulative lag period. The percent changes in PEF were negative in all cities except for
Baltimore, 0.49%. Among the other seven cities, the percent changes in PEF were quite
homogenous, with values ranging from -1.08% for Washington, DC to -1.71% for St. Louis.
A 30 ppb increase in 8-h avg O3 also was associated with an increased incidence of morning
symptoms in the pooled analysis (odds ratio of 1.35 [95% CI: 1.04, 1.69] for a 4-day cumulative
lag period). In all cities except for St. Louis, there was an increase in the incidence of morning
symptoms. In these cities, the odds ratios for incidence of morning symptoms were somewhat
more varied than for the PEF measurements, ranging from 1.19 for Chicago to 2.96 for Detroit.
Most of the multicity and meta-analyses studies consistently found positive associations
between O3 and mortality. Consistent O3 effects on hospitalizations and various respiratory
health outcomes also were found. The observed heterogeneity of O3 effects may be partially
attributable to the use of centrally-located ambient monitors to assess exposure. There may be
differences in relative personal exposures to O3 due to varying factors, such as use of air
conditioning and activity patterns, that affect the relationship between personal exposure and
ambient concentrations. For example, Levy et al. (2005) found suggestive evidence that air
conditioning prevalence was a predictor of heterogeneity in O3 risk estimates in their meta-
analysis. The variability in the concentration and composition of other pollutants present also
may contribute to the heterogeneity of the effect of O3 on health outcomes as confounding by
copollutants may vary by region.
7.6.7 Health Effects of Ozone in Susceptible and Vulnerable Populations
In this section, the effects of O3 on morbidity and mortality in potentially susceptible and
vulnerable populations will be examined. In epidemiologic studies of O3 health effects, the most
widely studied subpopulation was asthmatics. Also of interest were the observed health effects
of O3 on different age groups, particularly children and the elderly. Other groups that are
vulnerable to O3 health effects are those that spend a lot of time outdoors at higher exertion
7-163
-------
levels, such as outdoor workers. This section begins with a discussion of the O3-related health
effects in asthmatics.
7.6.7.1 Health Effects Associated with Ambient Ozone Exposure in Asthmatics
Epidemiologic studies of health effects from acute O3 exposure in asthmatics have
examined a range of outcomes: pulmonary function, respiratory symptoms, inflammation,
emergency room visits, hospital admissions, and mortality. Chronic O3 exposure studies have
investigated similar outcomes, with the exception of emergency room visits and hospitalizations.
Both are discussed in the earlier text. This subsection draws together this information to
examine whether the evidence indicates that O3 exposure impacts asthmatics.
In Germany and Mexico City, O3 exposure was associated with a decline in FEVj in
asthmatic adults and children (Hoppe et al., 1995a, 2003; Romieu et al., 2002). Change in FEVj
also was examined in a group of asthmatic hikers in Mount Washington, NH (Korrick et al.,
1998). The mean hourly O3 concentration during each hike was 40 ppb (range 21-74).
Compared to the healthy subjects, the asthmatic subjects experienced a 4-fold greater decline in
FEVj with the same exposure to O3 (mean change of -1.08% [95% CI: -2.49, 0.33] versus
-4.47% [95% CI: -7.65, -1.29] per 30 ppb increase in 8-h avg O3). The results from the hiker
study are consistent with those observed in a controlled human exposure study, which observed
an approximately 2-fold greater decrement in FEVj among mild-to-moderate asthmatics versus
nonasthmatic subjects performing light exercise during a 7.6-h exposure period at 0.16 ppm O3
(Horstman et al., 1995).
PEF was examined in panels of asthmatic children in several field studies (see Figures 7-1
and 7-2). Collectively, most of the studies indicated decrements of morning PEF, though only a
few estimates were statistically significant. One multicity study of eight urban areas in the
United States observed O3-related reductions in morning PEF that were not significant in each
individual city (Mortimer et al., 2002); however, the analysis combining data from all eight cities
indicated a significant decline in PEF with a cumulative lag of 1 to 5 days of O3 exposure. The
median 8-h avg O3 levels ranged from 34 to 58 ppb across the eight cities. The odds ratio for the
incidence of > 10% decline in morning PEF was greater than one, which was discussed by the
author as an indication that O3 exposure might be associated with clinically important changes in
7-164
-------
PEF in asthmatic children. The study examined 846 asthmatic children, the largest asthma panel
study reported.
Mortimer et al. (2000) observed that the subpopulation of asthmatic children with a history
of low birth weight or premature birth had greater O3-associated declines in PEF (mean change
of -3.66% [95% CI: -5.30, -2.02] per 30 ppb increase in 8-h avg O3) than normal birth weight
children (-0.60% [95% CI: -1.58, 0.38]). Low birth weight and prematurity are associated with
reduced lung function, higher levels of airway reactivity and increased susceptibility to lung
damage (Barker et al., 1993; Rona et al., 1993), which may explain why these factors are found
to increase susceptibility to respiratory insults of air pollution in children.
Lung function parameters have been evaluated for clinical significance. A reversible 5 to
15% decline in FEVj in an individual may have clinical importance to asthma morbidity
(American Thoracic Society, 1991; Lebowitz et al., 1987; Lippmann, 1988). The National
Institutes of Health (1997) has stated that PEF below 80% of the personal best indicates a need
for additional medication use in asthmatics. At a population level the mean changes in lung
function attributable to O3 exposure do not generally exceed 10% changes in FEVj or PEF per
standardized increment of O3. At an individual level, a subpopulation of susceptible asthmatics
are likely experiencing clinically significant declines in lung function. Hoppe et al. (2003)
investigated the effects of O3 on the lung function of potential risk groups, including asthmatics,
children, athletes, and the elderly. The mean afternoon (1 p.m.-4 p.m.) lA>-h max O3 levels on
high O3 days were 66.9 ppb (range 51-91) for the asthmatics and 65.2 ppb (range 43-88) for the
children. The afternoon Va-h max O3 levels in these two groups were similar to those
experienced by the elderly and athletes. Consistent associations between O3 and group means of
lung function endpoints were not observed for the various risk groups. However, a potential
pattern of O3 sensitivity was observed when individual data was examined. About 20% of the
asthmatics and children were regarded as O3 responders (i.e., individuals with >10% change
in FEVj) compared to only 5% of the elderly and athletes. In these responders, a significant O3
concentration-response relationship was observed in the regressions using repeated
measurements from individuals. These results indicated that while the population as a whole
was not reacting to O3, susceptible individuals were experiencing clinically significant declines
in lung function in response to O3 exposure.
7-165
-------
Respiratory symptom increases in asthma panels were examined in several field studies,
some of which also examined PEF as discussed above. The outcome definition of symptoms
varied among these studies. Collectively, the results are suggestive of a potential O3 effect on
respiratory symptoms, but the evidence is not strong in the available studies. Two U.S. studies
that examined larger panels might be better to draw inferences from as the large sample size
provided greater power to examine the effect of O3 on respiratory symptoms. The eight U.S.
urban cities study reported that morning symptoms in the 846 asthmatic children were most
strongly associated with a 4-day cumulative lag of O3 concentrations (Mortimer et al., 2002).
A New England study examined 271 asthmatic children and observed an O3 effect on a variety
of respiratory symptoms at a lag of 1 day among the 130 subjects who used maintenance asthma
medications (Gent et al., 2003). The mean 1-h max O3 was 58.6 ppb (SD 19.0).
Only a few epidemiologic studies have examined airway inflammation in asthmatics.
A Mexico City study indicated that supplementation with antioxidants may modulate the impact
of O3 exposure on the small airways of children with moderate-to-severe asthma (Romieu et al.,
2002). The mean 1-h max O3 was 102 ppb (SD 47). A related study indicated that asthmatic
children with GSTM1 null genotype were found to be more susceptible to the impact of O3
exposure on small airways (Romieu et al., 2004).
Emergency department visits for asthmatics have been examined in several studies and
range from negative to positive results (see Figure 7-8 in Section 7.3.2). Studies of mostly
year-long data tended to produce inconsistent results, with some finding negative estimates
(Atkinson et al., 1999a; Castellsague et al., 1995; Thompson et al., 2001; Tobias et al., 1999).
Warm-season studies tended to yield positive outcomes, as expected based on earlier
discussions. Two studies in Atlanta, GA (Tolbert et al., 2000) and Valencia, Spain (Tenias et al.,
1998) indicated positive effects in warm-season analyses. Further, a Canadian study, one of the
larger studies conducted in the summertime, reported a large increase in asthma emergency
department visits when the daily 1-h max O3 concentration exceeded 75 ppb (Stieb et al., 1996).
A three-city study in Ohio also indicated an increased risk of asthma visits during the summer
(Jaffe et al., 2003). The mean 8-h max O3 levels ranged from 50 to 60 ppb in the three cities.
Hospital admission studies that specifically examined asthmatics were fewer in number
than those that examined total respiratory diseases. Associations were noted in all age groups in
studies conducted in Seattle, WA (Sheppard, 2003), New Jersey (Weisel et al., 2002), Toronto,
7-166
-------
Canada (Burnett et al., 1999), London, England (Anderson et al., 1998), Brisbane, Australia
(Petroeschevsky et al., 2001), and Hong Kong (Wong et al., 1999a). However, several other
studies, mostly examining the effect of O3 on asthmatic children, did not observe a significant
relationship (Gouveia and Fletcher, 2000a; Lin et al., 2003; Morgan et al., 1998; Nauenberg and
Basu, 1999; Schouten et al., 1996).
Acute mortality related to asthma was examined in Barcelona, Spain (Saez et al., 1999;
Sunyer et al., 2002). In the study by Sunyer et al. (2002), severe asthmatics with more than one
asthma emergency visit were found to have the strongest mortality associations with O3. The
median 1-h max O3 level was 35.8 ppb (range 3.4-146.1).
Recent reports from longitudinal cohort studies in California have reported associations
between the onset of asthma and long-term O3 exposures (Greer et al., 1993; McConnell et al.,
2002; McDonnell et al., 1999). In adult studies, associations were seen in males but not females
(Greer et al., 1993; McDonnell et al., 1999). Among children residing in high O3 communities
(mean 8-h avg O3 of 59.6 ppb [range 55.8-69.0]), McConnell et al. (2002) observed that asthma
risk was elevated for those who played three or more sports as compared with those who did not
play sports. Playing sports may indicate outdoor activity and an increased ventilation rate which
leads to increased dose of O3. These outcomes would benefit from replication in other cohorts in
regards to indicating weight of a causal interpretation.
A few studies provide limited discussion of concentration-response functions and
thresholds. In the eight U.S. urban cities study of asthmatic children, the odds ratios for
incidence of >10% decline in morning PEF and incidence of morning symptoms when excluding
days with 8-h avg O3 greater than 80 ppb were nearly identical to those including data from all
days (Mortimer et al., 2002). In the New England asthma panel study (Gent et al., 2003), some
of the associations for symptoms occurred at 1-h max O3 levels below 60 ppb. In the St. John,
Canada study (Stieb et al., 2003), an effect of O3 on emergency department visits was reported
with evidence of a threshold somewhere in the range below a 1-h max O3 of 75 ppb in the
15 years and over age group.
Overall, subjects with asthma have been examined across most health endpoints of interest.
The results reported in these studies vary with some indicating a positive excess risk associated
with O3. While no endpoint in itself seems to indicate an unquestionable demonstration of an
association, studies with adequate sample size and power consistently provide positive results,
7-167
-------
especially during the summer months when higher O3 levels occur. This view is strengthened as
positive results are obtained cohesively across the varied outcomes. Therefore, based on the
evidence, it seems prudent to consider asthmatics as a susceptible group that requires more
protection from O3 exposures than the general public.
7.6.7.2 Age-Related Differences in Ozone Effects
The American Academy of Pediatrics (2004) notes that children and infants are among the
most susceptible to many air pollutants, including O3. Eighty percent of alveoli are formed
postnatally and changes in the lung continue through adolescence; the developing lung is highly
susceptible to damage from exposure to environmental toxicants (Dietert et al., 2000). Children
also have increased vulnerability as they spend more time outdoors, are highly active, and have a
high minute ventilation, which collectively increase their dose (Plunkett et al., 1992; Wiley et al.,
1991a,b). In addition to children, the elderly are frequently classified as being particularly
susceptible to air pollution. The basis of the increased sensitivity in the elderly is not known but
one hypothesis is that it may be related to changes in the respiratory tract lining fluid antioxidant
defense network (Kelly et al., 2003).
Several mortality studies have investigated age-related differences in O3 effects. Among
the studies that observed positive associations between O3 and mortality, a comparison of all age
or younger age (<65 years of age) O3-mortality risk estimates to that of the elderly population
(>65 years) indicates that, in general, the elderly population is more susceptible to O3 effects
(Borja-Aburto et al. 1997; Bremner et al., 1999; Gouveia and Fletcher 2000b; O'Neill et al.,
2004; Simpson et al., 1997; Sartor et al., 1995; Sunyer et al., 2002). For example, a study by
Gouveia and Fletcher (2000b) examined the O3-mortality effect by age in Sao Paulo, Brazil. The
mean 1-h max O3 level was 35.1 ppb (SD 21.7). There were 151,756 deaths for all non-violent
causes over the period of 1991 to 1993, of which 49% occurred in the elderly. Among all
ages, O3 was associated with a 0.6% (95% CI: -0.8, 2.0) excess risk in all cause mortality per
40 ppb increase in 1-h max O3. In comparison, in the elderly population, the O3-mortality risk
estimate was nearly 3-fold greater, 1.7% (95% CI: 0.0, 3.3). Similarly, a Mexico City study
found that O3-mortality risk estimates were 1.3% (95% CI: 0.04, 2.6) and 2.8% (95% CI: 1.0,
4.6) per 20 ppb increase in 24-h avg O3 concentration in all ages and the elderly, respectively
(O'Neill et al., 2004). The mean 24-h avg O3 level was 35.3 ppb (SD 11.0).
7-168
-------
The meta-analysis by Bell et al. (2005) found a larger effect estimate for the elderly (2.92%
[95% PI: 1.34, 4.51] per 20 ppb increase in 24-h avg O3) than for all ages (1.75% [95% PI:
1.10, 2.37]). In the large U.S. 95 communities study (Bell et al., 2004), effect estimates were
slightly higher for those aged 65 to 74 years, 1.40% (95% PI: 0.56, 2.25) excess risk per 20 ppb
increase in 24-h avg O3, compared to individuals less than 65 years and 75 years or greater,
1.00% (95% PI: 0.20, 1.85) and 1.04% (95% PI: 0.36, 1.75), respectively, using a constrained
distributed 7-day lag model. Bell et al. (2004) notes that despite somewhat similar effect
estimates, the absolute effect of O3 is substantially greater in the elderly population due to the
higher underlying mortality rates, which leads to a larger number of extra deaths for the elderly
compared to the general population.
A few mortality studies examined another potentially susceptible age group, young
children under the age of 5 years. The results were mixed, with one Mexico City study showing
a lower risk of O3-related all cause mortality in young children compared to all ages and the
elderly (Borja-Aburto et al., 1997) and one Sao Paulo, Brazil study showing a greater risk in
respiratory mortality in young children compared to the elderly (Gouveia and Fletcher, 2000b).
Another study in Mexico City by Loomis et al. (1999) observed a positive but nonsignificant
association between ambient O3 concentrations and infant mortality (used Poisson GAM with
default convergence criteria). Only a limited number of studies have focused on air pollution
effects on mortality in children using a time-series approach. This is probably because the
numbers involved are usually not adequate for such an analysis. Approximately 10% of
mortality occurs in young children. Gouveia and Fletcher (2000b) noted that in the case of
Sao Paulo, the mean number of daily deaths for respiratory causes in children was 2.2, which
was unlikely to provide statistical power to detect the effects of air pollution even if they existed.
In addition, there are other competing causes of mortality in young children, especially those in
developing countries, which are together more important than air pollution.
With respect to age-specificity of associations between O3 and acute respiratory
hospitalizations or emergency department visits, no clear pattern emerges from recent studies.
Associations have been reported for all ages (Anderson et al., 1997; Burnett et al., 1995, 1997b,
1999; Weisel et al., 2002), adults or elderly (Burnett et al., 1997a; Delfino et al., 1997b, 1998b;
Moolgavkar et al., 1997; Schwartz et al., 1996; Yang et al., 2003), and children (Burnett et al.,
2001; Gouveia and Fletcher, 2000a; Lin et al., 1999; Ponka and Virtanen, 1996; Tolbert et al.,
7-169
-------
2000; Yang et al., 2003). Interestingly, studies that have examined effects in multiple age strata
often have seen effects only in non-pediatric strata (Delfino et al., 1997b, 1998b; Stieb et al.,
1996; Jones et al., 1995). Several studies that focused on children did not report significant O3
effects, though in some cases these studies are limited by small size, inadequate control of
seasonal patterns, or very low O3 levels (Lierl and Hornung, 2003; Lin et al., 2003; Thompson
et al., 2001). If O3 is causally related to exacerbations of respiratory diseases leading to hospital
usage, one would expect to see effects most prominently among children, for whom asthma is
more prevalent and O3 exposures may be greater. However, once again, other competing causes
and the small numbers of hospitalizations in children likely limit the ability to examine O3-
related health effects.
A few field studies compared the effect of O3 in different age groups. Korrick et al. (1998)
examined changes in FEVj and FVC related to O3 exposure in a group of hikers ranging in age
from 18 to 64 years, and found that there was no association between O3 responsiveness and age.
Brauer et al. (1996), in a study of berry pickers aged 10 to 69 years, also observed that subject
age was not significantly associated with O3-related changes in lung function. However, a study
by Hoppe et al. (2003) observed that children, but not seniors (69 to 95 years of age),
experienced a decline in lung function associated with O3 exposure. Approximately 20% of the
children and juvenile asthmatics experienced a greater than 10% change in FEVl3 compared to
only 5% of the elderly population. The results by Hoppe et al. are consistent with the
diminishing responses to O3 exposure with increasing age observed in clinical studies. The
clinical studies by Drechsler-Parks (1995) and Bedi et al. (1989) found that subjects aged 56 to
89 years had markedly reduced responses to O3 exposure compared to young adults.
Many field studies focused on the effect of O3 on the respiratory health of school children.
In general, children experienced decrements in pulmonary function parameters, including
PEF, FEVb and FVC (Castillejos et al., 1995; Chen et al., 1999; Gielen et al., 1997; Gold et al.,
1999; Jalaludin et al., 2000; Mortimer et al., 2002; Romieu et al., 1996; Thurston et al., 1997).
Increases in respiratory symptoms (Delfino et al., 2003; Gold et al., 1999; Neas et al., 1995;
Romieu et al., 1996, 1997; Thurston et al., 1997) and asthma medication use (Delfino et al.,
1996; Just et al., 2002; Ostro et al., 2001; Thuston et al., 1997) also were observed in children.
These respiratory health effects were largely observed in asthmatic children. Ozone-associated
lung function declines were found in healthy children as well.
7-170
-------
Collectively, there is supporting evidence of age-related differences in susceptibility
to O3 health effects. The elderly population (>65 years of age) appear to be at increased risk
of O3-related mortality and hospitalizations, and children (<18 years of age) experience other
potentially adverse respiratory health outcomes with increased O3 exposure. One epidemiologic
study also found that the lung function response to O3 exposure may be diminished in elderly
populations; this finding is further supported by evidence from clinical studies.
7.6.7.3 Vulnerability of Outdoor Workers and Others Who Participate in Outdoor
Activities to Ozone Health Effects
The health effects of O3 on outdoor workers and others who participate in outdoor
activities have been investigated in various field studies. The most common endpoint examined
in these studies is lung function. These individuals are typically exposed to high doses of O3 as
they spend long hours outdoors often at elevated exertion levels. Their increased vulnerability
to O3 health effects has been noted in epidemiologic studies.
Brauer et al. (1996) repeatedly measured spirometric lung function before and after
outdoor summer work shifts over 59 days on a group of 58 berry pickers (mean age 44 years
[range 10-69]) in Fraser Valley, British Columbia, Canada. Outdoor work shifts averaged
11 hours in duration. The mean ambient 1-h max O3 was 40.3 ppb (SD 15.2) over the study
period. Heart rates during the work shift averaged 36% higher than resting levels, indicating
elevated exertion levels while working outdoors. The lung function changes experienced by the
workers in this study are large compared to those from other field studies (see Table 7-lb). The
next morning FEVj declined by 6.36% (95% CI: 4.70, 8.02) per 40 ppb increase in 1-h max O3.
These results indicate that extended exposures to O3 at elevated exertion levels produce more
marked effects on lung function.
Mexico City outdoor street workers (n = 47) were repeatedly monitored for lung function
changes at the end of the work shift over a two-month period (Romieu et al., 1998). Workers
were exposed to outdoor ambient O3 levels for a mean of 7.4 hours during the day. The mean
1-h max O3 was 123 ppb (SD 40). Among those who had never taken an antioxidant
supplement, same day O3 concentrations were associated with decreases in afternoon FEVj.
A mean change of-3.55% (95% CI: -6.28, -0.82) was observed per 40 ppb increase in
1-h max O3.
7-171
-------
Hoppe et al. (1995a) examined forestry workers (n = 41) for O3-related changes in
pulmonary function in Munich, Germany. The average time they spent outdoors was not
presented. Ventilation rates were estimated from the average activity levels. Their ventilation
rates were elevated at 40 L/min (compared to 15 L/min in the control group). When
comparisons were made between high O3 days (mean !/2-h max O3 of 64 ppb) and low O3 days
(mean !/2-h max O3 of 32 ppb), 59% of the forestry workers experienced a remarkable decrement
in lung function (i.e., at least a 20% increase in specific airway resistance or at least a 10%
decrease in FEVl3 FVC, or PEF) on high O3 days. None experienced an improvement in lung
function. A change of -56.0 mL (95% CI: -118.4, 6.4) in FEVj was observed per 40 ppb
increase in lA>-h max O3.
In addition to forestry workers, Hoppe et al. (1995a) also monitored athletes (n = 43) in the
afternoon following a two-hour outdoor training period. Athletes had a fairly high ventilation
rate of 80 L/min. Compared to the forestry workers a smaller percentage of athletes experienced
a remarkable decrement in lung function, 14%, on high O3 days; 19% of the athletes actually
showed an improvement. Overall, a significant change in FEVj was observed, -60.8 mL (95%
CI: -115.2, -6.4) per 40 ppb increase in !/2-h max, in the athletes. In a subsequent study, Hoppe
et al. (2003) reanalyzed the results of the athletes after stratifying the spirometric data by time of
day (morning versus afternoon) and at different lag periods (lags of 0 to 2 days). The reanalysis
indicated that O3-related decrements were observed only with the afternoon FEVj at a 0-day lag,
-1.26% (95% CI: -2.63, 0.10) change in FEVj per 40 ppb increase in !/2-h max O3.
A study by Korrick et al. (1998) also examined the effects of multihour O3 exposures on
adults exercising outdoors. A total of 530 hikers (mean age 35 years [range 18-64]) of Mount
Washington, NH performed spirometry before and after hiking for a mean of 8 hours (range
2-12). The mean of the hourly O3 concentrations during the hike was 40 ppb (range 21-74).
After the hike, all subjects combined experienced a small mean decline of 1.53% (95% CI: 0.24,
2.82) in FEVj per 30 ppb increase in the mean of the hourly O3 concentrations during the hike.
Ozone-related changes in lung function parameters were estimated after stratifying by hiking
duration. Subjects who hiked 8 to 12 hours (n = 265) experienced a 2.07% (95% CI: 0.36, 3.78)
decline in FEVj per 30 ppb increase in the mean of the hourly O3 concentrations; those who
hiked 2 to 8 hours (n = 265) experienced a smaller decline of 0.99% (95% CI: -0.72, 2.70).
7-172
-------
Each hour hiked, which may reflect dose, was associated with a decline of 0.3% (p = 0.05)
in FEVl3 after adjusting for O3.
The O3 effect attributable to exercise in children (n = 40) was investigated in a Mexico City
study (Castillejos et al., 1995). These children were chronically exposed to moderate to high
levels of O3. The mean ambient 1-h max O3 level during the study period was 112.3 ppb, with
a maximum value of 365 ppb. Spirometry was performed by the children before and after a
one-hour intermittent exercise session outdoors. Children were repeatedly tested up to 8 times.
A 0.48% (95% CI: 0.24, 0.72) decline in FEVj was experienced by the children after exercising
outdoors. However, stratified analyses indicated that significant changes were observed only
with higher quintiles of O3 exposure. At the highest quintile of exposure (183-365 ppb),
a -2.85% (95% CI: -1.30, - 4.40) change in FEVj was observed postexposure. Therefore,
children exercising outdoors when ambient O3 levels were high experienced declines in
pulmonary function despite the repeated daily exposure to moderate and high levels of O3 in
Mexico City.
In the southern California Children's Health Study, a total of 3,535 initially nonasthmatic
children (ages 9 to 16 years at enrollment) were followed for up to 5 years to identify new-onset
asthma cases (McConnell et al., 2002). Communities were stratified by pollution levels, with six
high-O3 communities (mean 1-h max O3 of 75.4 ppb [SD 6.8] over four years) and six low-O3
communities (mean 50.1 ppb [SD 11.0]). Asthma risk was not found to be higher for residents
of the six high-O3 communities versus residents of the six low-O3 communities. However,
within the high-O3 communities, asthma risk was 3.3 (95% CI: 1.9, 5.8) times greater for
children who played three or more sports as compared with children who played no sports. This
association was absent in the low-O3 communities (relative risk of 0.8 [95% CI: 0.4, 1.6]).
Thus, similar to the results observed in the Mexico City study by Castillejos et al. (1995), greater
effects were seen for susceptible individuals who exercised more outdoors at high O3 levels.
The studies discussed above indicate that prolonged exposure periods, combined with
elevated levels of exertion or exercise, may magnify the effect of O3 on lung function. Results
from these studies are consistent with the earlier summer camp studies (Avol et al., 1990;
Higgins et al., 1990; Raizenne et al., 1987, 1989; Spektor et al., 1988a, 1991) which also
indicated large O3-related changes in lung function parameters in children who spent long hours
outdoors. The large observed effects further suggest that other components in the ambient air
7-173
-------
pollution mixture may potentiate the effects of O3. In conclusion, outdoor workers who spend,
on average, 8 to 10 hours outdoors daily during the time of day when high peak O3
concentrations are expected appear to be particularly vulnerable to O3 health effects and may
need protection from O3 exposures.
7.6.8 Summary of Key Findings and Conclusions Derived from Ozone
Epidemiologic Studies
In the previous 1996 O3 AQCD, there was considerable evidence of O3-related respiratory
health effects from individual-level camp and exercise studies, as well as some consistent
evidence from time-series studies of emergency room visits and hospitalizations. Since the 1996
document, more field studies have been conducted, with some emphasis on additional outcome
markers such as respiratory symptoms and asthma medication use. Another significant addition
to the current O3 AQCD is the substantial number of short-term O3 mortality studies. The recent
publication of an analysis examining the relationship between O3 and mortality in 95 U.S.
communities (Bell et al., 2004) and three meta-analysis on O3-mortality associations (Bell et al.,
2005; Ito et al., 2005; Levy et al., 2005) also contribute significantly to the evidence base.
Considering the wide variability in possible study designs and statistical model specification
choices, the reported O3 risk estimates for the various health outcomes are in reasonably good
agreement. In the case of O3-mortality time-series studies, combinations of choices in model
specifications (e.g., the number of weather terms and degrees of freedom for smoothing of
mortality-temporal trends) alone may explain the extent of the difference in O3 risk estimates
across studies. As use of time-series studies to investigate air pollution effects has become more
common, there has been a great effort to evaluate the issues surrounding these studies.
The epidemiologic studies discussed in this chapter provide important information on the
associations between health effects and exposures of human populations to ambient O3.
A variety of oxidants in both the gaseous and paniculate phases have not been examined in
the available epidemiologic literature. The associations observed between ambient O3
concentrations and health outcomes may represent O3 effects, per se, or O3 may be serving as a
surrogate measure for the overall ambient photochemical oxidant mix.
In this section, conclusions regarding O3 health effects from the epidemiologic evidence
and the issues that may affect the interpretation of the effect estimates are briefly summarized.
7-174
-------
A more integrative synthesis of all relevant information will be presented in Chapter 8 of
this AQCD.
(1) Field/panel studies of acute O? effects. Results from recent field/panel studies
continue to confirm that short-term O3 exposure is associated with acute decrements
in lung function and increased respiratory symptoms, particularly in children and
asthmatics. There is also suggestive evidence that O3 is related to increased asthma
medication use. Strong O3 effects on lung function also were observed in outdoor
workers. Taken together with the evidence from controlled human exposure studies,
O3 is likely causally related to the various respiratory health outcomes. The current
evidence is much more limited, though suggestive, of a potential effect of O3 on heart
rate variability, ventricular arrhythmias, and the incidence of myocardial infarctions.
(2) Acute O? effects on emergency department visits and hospitalizations. Large multicity
studies, as well as many studies from individual cities have reported an effect of O3 on
respiratory hospital admissions. Studies using year-round data noted some
inconsistencies; however, studies with data restricted to the summer or warm season,
in general, indicated positive and robust associations between ambient O3
concentrations and respiratory hospital admissions. Effects of O3 on asthma
emergency department visits also were observed during the warm season.
(3) Acute O3 effects on mortality. The majority of the studies suggest an elevated risk of
all-cause mortality associated with acute exposure to O3, especially in the summer or
warm season when O3 levels are typically high. Slightly greater O3 effects were
observed for cardiovascular mortality. Results from recent large U.S. multicity time-
series studies provide the strongest evidence to-date for O3 effects on acute mortality.
Recent meta-analyses also indicate positive risk estimates that are unlikely to be
confounded by PM; however, future work is needed to better understand the influence
of model specifications on the risk coefficient.
(4) Chronic O3 exposure effects on morbidity and mortality. Fewer studies have
investigated the effect of chronic O3 exposure on morbidity and mortality. The
strongest evidence is for seasonal effects of extended O3 exposures on lung function in
children, i.e., reduced lung function growth being associated with higher ambient O3
levels. Longer-term studies investigating the association of chronic O3 exposure on
yearly lung function, asthma incidence, and respiratory symptoms are inconclusive.
Chronic O3-mortality studies observed inconsistent results across exposure periods and
cause-specific mortality outcomes.
(5) Exposure assessment. Exposure misclassification may result from the use of
stationary ambient monitors to determine exposure in population studies. Although
central ambient monitors do not explain the variance of individual personal exposures,
7-175
-------
significant correlations are found between aggregate personal O3 measurements and O3
concentrations from ambient monitors. A simulation study indicated that the use of
ambient monitor data will tend to underestimate the O3 effect. A better understanding
of the factors that affect the relationship between ambient concentrations and personal
exposures will improve interpretation of the O3 effect estimates.
(6) Ozone exposure indices. The three most commonly used daily O3 exposure indices,
1-h max O3, 8-max O3, and 24-h avg O3, were found to be highly correlated in studies
conducted in various regions. In addition, the effect estimates and significance of
associations across all health outcomes were comparable when using the standardized
distributional increment of 40 ppb, 30 ppb, and 20 ppb for mean 1-h max O3, mean 8-h
max O3, and mean 24-h avg O3, respectively.
(7) Lag structures for O? exposure and effect. The lag time between O3 exposure and
effect may differ depending on various factors such as the specific health outcome of
interest, the mechanism of effect, and preexisting health conditions. The majority of
the studies found an immediate O3 effect, with the strongest associations observed
between health outcomes and O3 exposure on the same day and/or previous day. Some
studies found large cumulative effects of O3 over longer lag periods, indicating that
multiday lags also may be relevant for some health outcomes, including mortality and
asthma symptoms in children.
(8) Sensitivity to model specifications for temporal trends. Ozone effect estimates
that were reported in studies whose main focus was PM often were calculated using
the same model specifications as PM to adjust for temporal trends. While the
sensitivity of the O3 risk estimates to alternative model specifications has not been
throughly investigated, limited evidence indicates that O3 effects may be robust to
various model specifications for temporal trend adjustment.
(9) Sensitivity to model specifications for meteorological effects. Ozone risk estimates
were generally more sensitive to alternative weather models than to varying degrees of
freedom for temporal trend adjustment. In studies in which alternative weather models
(e.g., quintile indicator model and four-smoother model) were considered, up to a
factor of two difference in the O3 risk estimates was observed. Further research is
needed to reduce the uncertainties related to confounding by weather influences.
(10) Influence of seasonal factors. An evaluation of the confounding effects of
meteorologic factors and copollutants on O3 risk estimates is complicated by their
changing relationships with O3 across seasons. In addition, seasonal or seasonally-
modified factors (e.g., air conditioning use, time spent outdoors) complicate
interpretation of all-year effect estimates as they affect the relationship between
7-176
-------
ambient concentrations and personal exposures. Given the potentially significant
influence of season, season-specific analyses are more informative in assessing O3
health risks.
(11) Confounding by copollutants. Multipollutant regression models often are used to
adjust for confounding by copollutants. Although there is some concern regarding the
use of multipollutant models given the varying concurvity across pollutants, results
generally suggest that the inclusion of copollutants into the models do not substantially
affect O3 risk estimates. These findings indicate that effects of O3 on various health
outcomes are robust and independent of the effects of other copollutants.
(12) Concentration-response function. In the limited mortality and morbidity studies that
have specifically examined the O3 concentration-response relationship, the evidence is
inconclusive regarding the presence of an effect threshold. Factors such as exposure
measurement error may reduce the ability to detect a threshold in population studies.
The limited evidence suggests that if a population threshold exists in O3 health effects,
it is likely near the lower limit of ambient O3 concentrations in the United States.
(13) Heterogeneity of O3 health effects. Consistent O3 effect estimates generally were
observed for mortality, hospitalizations, and other respiratory health outcomes in
multicity studies. Some of the observed geographic heterogeneity in effects may be
attributable to the differences in relative personal exposure to O3, which is affected by
factors such as air conditioning prevalence and activity patterns, as well as varying
concentrations and compositions of copollutants present by region.
(14) Ozone health effects in asthmatics. The effects of O3 on asthmatics have been
examined widely in both time-series studies and panel studies. Associations of
O3 with various respiratory health outcomes, including lung function declines,
increased respiratory symptoms, and emergency department visits, were observed.
These findings, along with the pathophysiologic understanding of asthma as a chronic
inflammatory disease, indicate that asthmatics are likely a susceptible population that
requires protection from O3 exposures.
(15) Age-related differences in (X health effects. Supporting evidence exists for
heterogeneity in the effects of O3 by age. The elderly population (>65 years of age)
appears to be at greater risk of O3-related mortality and hospitalizations compared to
all age populations. In addition, potentially adverse respiratory health outcomes are
associated with O3 exposure in children (<18 years of age). Lung function responses
to O3 exposure are diminished in the elderly population, as observed in an
epidemiologic study and numerous human clinical studies.
7-177
-------
REFERENCES
Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. I; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir. Crit.
Care Med. 159: 373-382.
American Academy of Pediatrics, Committee on Environmental Health. (2004) Ambient air pollution: health
hazards to children. Pediatrics 114: 1699-1707.
American Thoracic Society. (1991) Lung function testing: selection of reference values and interpretative strategies
Am. Rev. Respir. Dis. 144: 1202-1218.
Anderson, H. R.; Ponce de Leon, A.; Bland, J. M; Bower, J. S.; Strachan, D. P. (1996) Air pollution and daily
mortality in London: 1987-92. Br. Med. J. 312: 665-669.
Anderson, H. R.; Spix, C.; Medina, S.; Schouten, J. P.; Castellsague, J.; Rossi, G.; Zmirou, D.; Touloumi, G.;
Wojtyniak, B.; Ponka, A.; Bacharova, L.; Schwartz, J.; Katsouyanni, K. (1997) Air pollution and daily
admissions for chronic obstructive pulmonary disease in 6 European cities: results from the APHEA project.
Eur. Respir. J. 10: 1064-1071.
Anderson, H. R.; Ponce de Leon, A.; Bland, J. M.; Bower, J. S.; Emberlin, J.; Strachen, D. P. (1998) Air pollution,
pollens, and daily admissions for asthma in London 1987-92. Thorax 53: 842-848.
Anderson, H. R.; Bremner, S. A.; Atkinson, R. W.; Harrison, R. M.; Walters, S. (2001) Paniculate matter and daily
mortality and hospital admissions in the west midlands conurbation of the United Kingdom: associations with
fine and coarse particles, black smoke and sulphate. Occup. Environ. Med. 58: 504-510.
Atkinson, R. W.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Bremner, S. A.; Ponce de Leon, A. (1999a)
Short-term associations between outdoor air pollution and visits to accident and emergency departments in
London for respiratory complaints. Eur. Respir. J. 13: 257-265.
Atkinson, R. W.; Bremner, S. A.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Ponce de Leon, A. (1999b)
Short-term associations between emergency hospital admissions for respiratory and cardiovascular disease
and outdoor air pollution in London. Arch. Environ. Health 54: 398-411.
Atkinson, R. W.; Anderson, H. R.; Sunyer, J.; Ayres, J.; Baccini, M.; Vonk, J. M.; Boumghar, A.; Forastiere, F.;
Forsberg, B.; Touloumi, G.; Schwartz, J.; Katsouyanni, K. (2001) Acute effects of paniculate air pollution on
respiratory admissions: results from APHEA 2 project. Am. J. Respir. Crit. Care Med. 164: 1860-1866.
Avol, E. L.; Trim, S. C.; Little, D. E.; Spier, C. E.; Smith, M. N.; Peng, R.-C.; Linn, W. S.; Hackney, J. D.; Gross,
K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone exposure and lung function in children
attending a southern California summer camp. Presented at: 83rd annual meeting and exhibition of the Air &
Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management
Association; paper no. 90-150.3.
Avol, E. L.; Navidi, W. C.; Rappaport, E. B.; Peters, J. M. (1998) Acute effects of ambient ozone on asthmatic,
wheezy, and healthy children. Cambridge, MA: Health Effects Institute; research report no. 82.
Avol, E. L.; Gauderman, W. J.; Tan, S. M.; London, S. J.; Peters, J. M. (2001) Respiratory effects of relocating to
areas of differing air pollution levels. Am. J. Respir. Crit. Care Med. 164: 2067-2072.
Ballester, F.; Tenias, J. M.; Perez-Hoyos, S. (2001) Air pollution and emergency hospital admissions for
cardiovascular diseases in Valencia, Spain. J. Epidemiol. Community Health 55: 57-65.
Barker, D. J. P.; Gluckman, P. D.; Godfrey, K. M.; Harding, J. E.; Owens, J. A.; Robinson, J. S. (1993) Fetal
nutrition and cardiovascular disease in adult life. Lancet 341: 938-941.
Bates, D. V. (2005) Ambient ozone and mortality. Epidemiology 16: 427-429.
Bates, D. V.; Sizto, R. (1983) Relationship between air pollutant levels and hospital admissions in Southern Ontario.
Can. J. Public Health 74: 117-122.
Bates, D. V.; Sizto, R. (1987) Air pollution and hospital admissions in southern Ontario: the acid summer haze
effect. Environ. Res. 43: 317-331.
Bates, D. V.; Sizto, R. (1989) The Ontario Air Pollution study: identification of the causative agent. Environ. Health
Perspect. 79: 69-72.
Bates, D. V.; Baker-Anderson, M.; Sizto, R. (1990) Asthma attack periodicity: a study of hospital emergency visits
in Vancouver. Environ. Res. 51: 51-70.
Bedi, J. F.; Horvath, S. M.; Drechsler-Parks, D. M. (1989) Adaptation by older individuals repeatedly exposed to
0.45 parts per million ozone for two hours. JAPCA 39: 194-199.
Beeson, W. L.; Abbey, D. E.; Knutsen, S. F. (1998) Long-term concentrations of ambient air pollutants and incident
lung cancer in California adults: results from the AHSMOG study. Environ. Health Perspect. 106: 813-823.
7-178
-------
Bell, M. L.; McDermott, A.; Zeger, S. L.; Samet, J. M.; Dominici, F. (2004) Ozone and short-term mortality in
95 US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292: 2372-2378.
Bell, M. L. (2006) Community-specific maximum likelihood estimates of O3-related excess risk in mortality for the
NMMAPS U.S. 95 communities study [personal communication with attachments to Jee Young Kim].
New Haven, CT: Yale University School of Forestry and Environmental Studies; January 6.
Bell, M. L.; Dominici, F.; Samet, J. M. (2005) A meta-analysis of time-series studies of ozone and mortality with
comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
Bell, M. L.; Peng, R. D.; Dominici, F. (2006) The exposure-response curve for ozone and risk of mortality and the
adequacy of current ozone regulations. Environ. Health Perspect: in press, doi:10.1289/ehp.8816. Available:
http://dx.doi.org/ [23 January, 2006].
Berhane, K.; Thomas, D. C. (2002) A two-stage model for multiple time series data of counts. Biostatistics 3: 21-32.
Bobak, M. (2000) Outdoor air pollution, low birth weight, and prematurity. Environ. Health Perspect. 108: 173-176.
Bobak, M.; Leon, D. A. (1992) Air pollution and infant mortality in the Czech Republic, 1986-1988. Lancet
(8826): 1010-1014.
Bobak, M.; Leon, D. A. (1999) Pregnancy outcomes and outdoor air pollution: an ecological study in districts of the
Czech Republic 1986-8. Occup. Environ. Med. 56: 539-543.
Borja-Aburto, V. H.; Loomis, D. P.; Bangdiwala, S. I.; Shy, C. M.; Rascon-Pacheco, R. A. (1997) Ozone, suspended
particulates, and daily mortality in Mexico City. Am. J. Epidemiol. 145: 258-268.
Borja-Aburto, V. H.; Castillejos, M.; Gold, D. R.; Bierzwinski, S.; Loomis, D. (1998) Mortality and ambient fine
particles in southwest Mexico City, 1993-1995. Environ. Health Perspect. 106: 849-855.
Borsboom, G. J. J. M.; Van Pelt, W.; Van Houwelingen, H. C.; Van Vianen, B. G.; Schouten, J. P.; Quanjer, P. H.
(1999) Diurnal variation in lung function in subgroups from two Dutch populations. Consequences for
longitudinal analysis. Am. J. Respir. Crit. Care Med. 159: 1163-1171.
Bourcier, T.; Viboud, C.; Cohen, J.-C.; Thomas, F.; Bury, T.; Cadiot, L.; Mestre, O.; Flahault, A.; Borderie, V;
Laroche, L. (2003) Effects of air pollution and climatic conditions on the frequency of ophthalmological
emergency examinations. Br. J. Ophthalmol. 87: 809-811.
Braga, A. L. F.; Conceicao, G. M. S.; Pereira, L. A. A.; Kishi, H. S.; Pereira, J. C. R.; Andrade, M. F.; Goncalves,
F. L. T.; Saldiva, P. H. N.; Latorre, M. R. D. O. (1999) Air pollution and pediatric respiratory hospital
admissions in Sao Paulo, Brazil. J. Environ. Med. 1: 95-102.
Brauer, M.; Brook, J. R. (1997) Ozone personal exposures and health effects for selected groups residing in the
Fraser Valley. In: Steyn, D. G.; Bottenheim, J. W., eds. The Lower Fraser Valley Oxidants/Pacific '93 Field
Study. Atmos. Environ. 31: 2113-2121.
Brauer, M.; Blair, J.; Vedal, S. (1996) Effect of ambient ozone exposure on lung function in farm workers. Am.
J. Respir. Crit. Care Med. 154: 981-987.
Brauer, M.; Brumm, J.; Vedal, S.; Petkau, A. J. (2002) Exposure misclassification and threshold concentrations in
time series analyses of air pollution health effects. Risk Anal. 22: 1183-1193.
Bremner, S. A.; Anderson, H. R.; Atkinson, R. W.; McMichael, A. J.; Strachan, D. P.; Bland, J. M.; Bower, J. S.
(1999) Short term associations between outdoor air pollution and mortality in London 1992-4. Occup.
Environ. Med. 56: 237-244.
Brook, R. D.; Franklin, B.; Cascio, W.; Hong, Y.; Howard, G.; Lipsett, M.; Luepker, R.; Mittleman, M.; Samet, J.;
Smith, S. C., Jr.; Tager, I. (2004) Air pollution and cardiovascular disease. A statement for healthcare
professionals from the Expert Panel on Population and Prevention Science of the American Heart
Association. Circulation 109: 2655-2671.
Buchdahl, R.; Parker, A.; Stebbings, T.; Babiker, A. (1996) Association between air pollution and acute childhood
wheezy episodes: prospective observational study. Br. Med. J. 312: 661-664.
Buchdahl, R.; Willems, C. D.; Vander, M.; Babiker, A. (2000) Associations between ambient ozone, hydrocarbons,
and childhood wheezy episodes: a prospective observational study in south east London. Occup. Environ.
Med. 57: 86-93.
Burchfiel, C. M.; Marcus, E. B.; Curb, J. D.; Maclean, C. J.; Vollmer, W. M.; Johnson, L. R.; Fong, K.; Rodriguez,
B. L.; Masaki, K. H.; Buist, A. S. (1995) Effects of smoking and smoking cessation on longitudinal decline in
pulmonary function. Am. J. Respir. Crit. Care Med. 151: 1778-1785.
Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M.;
Dann, T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of respiratory
admissions to Ontario hospitals. Environ. Res. 65: 172-194.
7-179
-------
Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
paniculate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am. J. Epidemiol.
142: 15-22.
Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997a) Association between ozone and
hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72: 24-31.
Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997b) The role of paniculate size and chemistry in the
association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
Environ. Health Perspect. 105: 614-620.
Burnett, R. T.; Smith-Doiron, M; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of paniculate and gaseous air
pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54: 130-139.
Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.; Leech, J. A.; Cakmak, S.;
Krewski, D. (2001) Association between ozone and hospitalization for acute respiratory diseases in children
less than 2 years of age. Am. J. Epidemiol. 153: 444-452.
Burr, D.; Doss, H. (2005) A Bayesian semiparametric model for random-effects meta-analysis. J. Am. Stat. Assoc.
100:242-251.
Calderon-Garciduenas, L.; Rodriguez-Alcaraz, A.; Garcia, R.; Ramirez, L.; Barragan, G. (1995) Nasal inflammatory
responses in children exposed to a polluted urban atmosphere. J. Toxicol. Environ. Health 45: 427-437.
Cassino, C.; Ito, K.; Bader, L; Ciotoli, C.; Thurston, G.; Reibman, J. (1999) Cigarette smoking and ozone-associated
emergency department use for asthma by adults in New York City. Am. J. Respir. Crit. Care Med. 159:
1773-1779.
Castellsague, J.; Sunyer, J.; Saez, M.; Anto, J. M. (1995) Short-term association between air pollution and
emergency room visits for asthma in Barcelona. Thorax 50: 1051-1056.
Castillejos, M.; Gold, D. R.; Damokosh, A. L; Serrano, P.; Allen, G.; McDonnell, W. F.; Dockery, D.; Velasco,
S. R.; Hernandez, M.; Hayes, C. (1995) Acute effects of ozone on the pulmonary function of exercising
schoolchildren from Mexico City. Am. J. Respir. Crit. Care Med. 152: 1501-1507.
Chan, C.-C.; Wu, T.-H. (2005) Effects of ambient ozone exposure on mail carriers' peak expiratory flow rates.
Environ. Health Perspect. 113: 735-738.
Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H. H. (2000) Hourly personal exposures to fine particles and
gaseous pollutants-results from Baltimore, Maryland. J. Air Waste Manage. Assoc. 50: 1223-1235.
Chang, C.-C.; Tsai, S.-S.; Ho, S.-C.; Yang, C.-Y. (2005) Air pollution and hospital admissions for cardiovascular
disease in Taipei, Taiwan. Environ. Res. 98: 114-119.
Charpin, D.; Pascal, L.; Birnbaum, J.; Armengaud, A.; Sambuc, R.; Lanteaume, A.; Vervloet, D. (1999) Gaseous air
pollution and atopy. Clin. Exp. Allergy 29: 1474-1480.
Chen, P.-C; Lai, Y.-M.; Wang, J.-D.; Yang, C.-Y.; Hwang, J.-S.; Kuo, H.-W.; Huang, S.-L.; Chan, C.-C. (1998)
Adverse effect of air pollution on respiratory health of primary school children in Taiwan. Environ. Health
Perspect. 106:331-335.
Chen, P.-C.; Lai, Y.-M.; Chan, C.-C.; Hwang, J.-S.; Yang, C.-Y.; Wang, J.-D. (1999) Short-term effect of ozone on
the pulmonary function of children in primary school. Environ. Health Perspect. 107: 921-925.
Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. (2000) Elementary school absenteeism and air pollution.
Inhalation Toxicol. 12: 997-1016.
Chen, L.; Yang, W.; Jennison, B. L.; Goodrich, A.; Omaye, S. T. (2002) Air pollution and birth weight in northern
Nevada, 1991-1999. Inhalation Toxicol. 14: 141-157.
Chew, F. T.; Goh, D. Y. T.; Ooi, B. C.; Saharom, R.; Hui, J. K. S.; Lee, B. W. (1999) Association of ambient
air-pollution levels with acute asthma exacerbation among children in Singapore. Allergy (Copenhagen)
54: 320-329.
Chock, D. P.; Winkler, S. L.; Chen, C. (2000) A study of the association between daily mortality and ambient air
pollutant concentrations in Pittsburgh, Pennsylvania. J. Air Waste Manage. Assoc. 50: 1481-1500.
Cifuentes, L. A.; Vega, J.; Kopfer, K.; Lave, L. B. (2000) Effect of the fine fraction of paniculate matter versus the
coarse mass and other pollutants on daily mortality in Santiago, Chile. J. Air Waste Manage. Assoc. 50:
1287-1298.
Clyde, M. A. (1999) Bayesian model averaging and model search strategies. In: Bernardo, J. M.; Berger, J. O.;
Dawid, A. P.; Smith, A. F. M., eds. Bayesian Statistics 6: proceedings of the Sixth Valencia International
Meeting, June; pp. 157-185. Oxford, UK. Oxford, UK: Clarendon Press.
Clyde, M. (2000) Model uncertainty and health effect studies for paniculate matter. Environmetncs 11: 745-763.
7-180
-------
Clyde, M. A.; Guttorp, P.; Sullivan, E. (2000) Effects of ambient fine and coarse particles on mortality in Phoenix,
Arizona. Seattle, WA: University of Washington, National Research Center for Statistics and the
Environment; NRCSE technical report series, NRCSE-TRS no. 040. Available:
http://www.nrcse.washington.edu/pdf/trs40_pm.pdf [18 October, 2004].
Cody, R. P.; Weisel, C. P.; Birnbaum, G.; Lioy, P. J. (1992) The effect of ozone associated with summertime
photochemical smog on the frequency of asthma visits to hospital emergency departments. Environ. Res.
58: 184-194.
Cross, D.; Nelson, H. S. (1991) The role of the peak flow meter in the diagnosis and management of asthma.
J. Allergy Clin. Immunol. 87: 120-128.
Cuijpers, C. E. J.; Swaen, G. M. H.; Wesseling, G.; Wouters, E. F. M. (1994) Acute respiratory effects of summer
smog in primary school children. Toxicol. Lett. 72: 227-235.
Dab, W.; Medina, S.; Quenel, P.; Le Moullec, Y.; Le Tertre, A.; Thelot, B.; Monteil, C.; Lameloise, P.; Pirard, P.;
Momas, I.; Ferry, R.; Festy, B. (1996) Short term respiratory health effects of ambient air pollution: results of
the APHEA project in Paris. In: St Leger, S., ed. The APHEA project. Short term effects of air pollution on
health: a European approach using epidemiological time series data. J. Epidemiol. Commun. Health
50(suppl. 1): S42-S46.
Dejmek, J.; Selevan, S. G.; Benes, I.; Solansky, I.; Sra'm, R. J. (1999) Fetal growth and maternal exposure to
paniculate matter during pregnancy. Environ. Health Perspect. 107: 475-480.
Dejmek, J.; Solansky, I.; Benes, I.; Lenicek, J.; Sra'm, R. J. (2000) The impact of polycyclic aromatic hydrocarbons
and fine particles on pregnancy outcome. Environ. Health Perspect. 108: 1159-1164.
De Leon, S. F.; Thurston, G. D.; Ito, K. (2003) Contribution of respiratory disease to nonrespiratory mortality
associations with air pollution. Am. J. Respir. Crit. CareMed. 167: 1117-1123.
Delfino, R. J.; Coate, B. D.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Koutrakis, P. (1996) Daily asthma severity in
relation to personal ozone exposure and outdoor fungal spores. Am. J. Respir. Crit. Care Med. 154: 633-641.
Delfino, R. J.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Matteucci, R. M.; Anderson, P. R.; Koutrakis, P. (1997a)
The effect of outdoor fungal spore concentrations on daily asthma severity. Environ. Health Perspect.
105: 622-635.
Delfino, R. J.; Murphy-Moulton, A. M.; Burnett, R. T.; Brook, J. R.; Becklake, M. R. (1997b) Effects of air
pollution on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J. Respir. Crit. Care
Med. 155: 568-576.
Delfino, R. J.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H. (1998a) Symptoms inpediatric asthmatics and air pollution:
differences in effects by symptom severity, anti-inflammatory medication use and paniculate averaging time.
Environ. Health Perspect. 106: 751-761.
Delfino, R. J.; Murphy-Moulton, A. M.; Becklake, M. R. (1998b) Emergency room visits for respiratory illnesses
among the elderly in Montreal: association with low level ozone exposure. Environ. Res. 76: 61-11.
Delfino, R. J.; Gone, H.; Linn, W. S.; Pellizzari, E. D.; Hu, Y. (2003) Asthma symptoms in Hispanic children and
daily ambient exposures to toxic and criteria air pollutants. Environ. Health Perspect. Ill: 647-656.
Delfino, R. J., Quintana, P. J. E.; Floro, J.; Gastanaga, V. M.; Samimi, B. S.; Kleinman, M. T.; Liu, L.-J. S.;
Bufalino, C.; Wu, C.-F.; McLaren, C. E. (2004) Association of FEVl in asthmatic children with personal and
microenvironmental exposure to airborne paniculate matter. Environ. Health Perspect. 112: 932-941.
Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Squinazi, F.; Momas, I. (2002a) Short-term effects of low-level air
pollution on respiratory health of adults suffering from moderate to severe asthma. Environ. Res. A 89: 29-37.
Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Le Moullec, Y.; Momas, I. (2002b) Effects of air pollution on adults
with chronic obstructive pulmonary disease. Arch. Environ. Health 57: 554-560.
Detels, R.; Tashkin, D. P.; Sayre, J. W.; Rokaw, S. N.; Coulson, A. H.; Massey, F. J., Jr.; Wegman, D. H. (1987)
The UCLA population studies of chronic obstructive respiratory disease: 9. lung function changes associated
with chronic exposure to photochemical oxidants; a cohort study among never-smokers. Chest 92: 594-603.
Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the
lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
Diaz, J.; Garcia, R.; Ribera, P.; Alberdi, J. C.; Hernandez, E.; Pajares, M. S.; Otero, A. (1999) Modeling of air
pollution and its relationship with mortality and morbidity in Madrid, Spain. Int. Arch. Occup. Environ.
Health 72: 366-376.
7-181
-------
Dietert, R. R.; Etzel, R. A.; Chen, D.; Halonen, M.; Holladay, S. D.; Jarabek, A. M; Landreth, K.; Peden, D. B.;
Pinkerton, K.; Smialowicz, R. I; Zoetis, T. (2000) Workshop to identify critical window of exposure for
children's health: immune and respiratory systems work group summary. Environ. Health Perspect. Suppl.
108(3): 483-490.
Dockery, D. W.; Schwartz, I; Spengler, J. D. (1992) Air pollution and daily mortality: associations with particulates
and acid aerosols. Environ. Res. 59: 362-373.
Dockery, D. W.; Luttmann-Gibson, H.; Rich, D. Q.; Link, M. S.; Mittleman, M. A.; Gold, D. R.; Koutrakis, P.;
Schwartz, J. D.; Verrier, R. L. (2005) Association of air pollution with increased incidence of ventricular
tachyarrhythmias recorded by implanted cardioverter defibrillators. Environ. Health Perspect. 113: 670-674.
Dockins, C.; Maguire, K.; Simon, N.; Sullivan, M. (2004) Value of statistical life analysis and environmental policy:
a white paper. Washington, DC: U.S. Environmental Protection Agency, National Center for Environmental
Economics; report no. EE-0483. Available:
http://yosemite.epa.gov/ee/epa/eermfile.nsf/vwAN/EE-0483-01.pdf/$File/EE-0483-01.pdf [12 January, 2006].
Dominici, F.; McDermott, A.; Zeger, S. L.; Samet, J. M. (2002) On the use of generalized additive models in
time-series studies of air pollution and health. Am. J. Epidemiol. 156: 193-203.
Dominici, F.; McDermott, A.; Daniels, M.; Zeger, S. L.; Samet, J. M. (2003) Mortality among residents of 90 cities.
In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA: Health
Effects Institute; pp. 9-24. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
Drechsler-Parks, D. M. (1995) The dose-response relationship in older men exposed to ozone. Exp. Gerontol.
30: 65-75.
Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
Perspect. 107: 637-641.
Fairley, D. (2003) Mortality and air pollution for Santa Clara County, California, 1989-1996. In: Revised analyses of
time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute; pp.
97-106. Available: http://www.healtheffects.org/news.htm [16 May, 2003].
Flachaire, E.; Nunez, O. (2002) Estimation of income distribution and detection of subpopulations: an explanatory
model. Paris, France: Document de travail de la MSB, EUREQua 2002.86, Universite Parisl
Pantheon-Sorbonne. Available:
http://eurequa.univ-parisl.fr/membres/flachaire/research/Flachaire_Nunez_02.pdf [14 June, 2005].
Friedman, M. S.; Powell, K. E.; Hutwagner, L.; Graham, L. M.; Teague, W. G. (2001) Impact of changes in
transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality and
childhood asthma. JAMA J. Am. Med. Assoc. 285: 897-905.
Frischer, T. M.; Ktihr, J.; Pullwitt, A.; Meinert, R.; Forster, J.; Studnicka, M.; Koren, H. (1993) Ambient ozone
causes upper airways inflammation in children. Am. Rev. Respir. Dis. 148: 961-964.
Frischer, T.; Pullwitt, A.; Ktihr, K.; Meinert, R.; Haschke, N.; Studnicka, M.; Lubec, G. (1997) Aromatic
hydroxylation in nasal lavage fluid following ambient ozone exposure. Free Radical Biol. Med. 22: 201-207.
Frischer, T.; Studnicka, M.; Gartner, C.; Tauber, E.; Horak, F.; Veiter, A.; Spengler, J.; Ktihr, J.; Urbanek, R. (1999)
Lung function growth and ambient ozone: a three-year population study in school children. Am. J. Respir.
Crit. Care Med. 160: 390-396.
Frischer, T.; Studnicka, M.; Halmerbauer, G.; Horak, F.; Gartner, C.; Tauber, E.; Koller, D. Y. (2001) Ambient
ozone exposure is associated with eosinophil activation in healthy children. Clin. Exp. Allergy 31: 1213-1219.
Fuhlbrigge, A.; Kitch, B.; Paltiel, A. D.; Kuntz, K. M.; Neumann, P. J.; Dockery, D. W.; Weiss, S. T. (2001) FEVj is
associated with risk of asthma attacks in a pediatric population. J. Allergy Clin. Immunol. 107: 61-67.
Fung, K. Y.; Luginaah, I.; Gorey, K. M.; Webster, G. (2005) Air pollution and daily hospital admissions for
cardiovascular diseases in Windsor, Ontario. Can. J. Public Health 96: 29-33.
Galbraith, R. F. (1994) Some applications of radial plots. J. Am. Stat. Assoc. 89: 1232-1242.
Galizia, A.; Kinney, P. L. (1999) Long-term residence in areas of high ozone: associations with respiratory health in
a nationwide sample of nonsmoking young adults. Environ. Health Perspect. 107: 675-679.
Gamble, J. L. (1998) Effects of ambient air pollution on daily mortality: a time series analysis of Dallas, Texas,
1990-1994. Presented at: 91st annual meeting and exhibition of the Air & Waste Management Association;
June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 98-MP26.03.
Garcia-Aymerich, J.; Tobias, A.; Anto, J. M.; Sunyer, J. (2000) Air pollution and mortality in a cohort of patients
with chronic obstructive pulmonary disease: a time series analysis. J. Epidemiol. Community Health
54: 73-74.
7-182
-------
Garty, B. Z.; Kosman, E.; Ganor, E.; Berger, V.; Garty, L.; Wietzen, T.; Waisman, Y.; Mimouni, M.; Waisel, Y.
(1998) Emergency room visits of asthmatic children, relation to air pollution, weather, and airborne allergens.
Ann. Allergy Asthma Immunol. 81: 563-570.
Gauderman, W. I; McConnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and lung
function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
Gauderman, W. J.; Gilliland, G. F.; Vora, H.; Avol, E.; Stram, D.; McConnell, R.; Thomas, D.; Lurmann, F.;
Margolis, H. G.; Rappaport, E. B.; Berhane, K.; Peters, J. M. (2002) Association between air pollution and
lung function growth in southern California children: results from a second cohort. Am. J. Respir. Crit. Care
Med. 166: 76-84.
Gauderman, W. J.; Avol, E.; Gilliland, F.; Vora, H.; Thomas, D.; Berhane, K.; McConnell, R.; Kuenzli, N.;
Lurmann, F.; Rappaport, E.; Margolis, H.; Bates, D.; Peters, J. (2004a) The effect of air pollution on lung
development from 10 to 18 years of age. N. Engl. J. Med. 351: 1057-1067.
Gauderman, W. J.; Avol, E.; Gilliland, F. (2004b) Air pollution and lung function [author's reply]. N. Engl. J. Med.
351:2653.
Gent, J. F.; Triche, E. W.; Holford, T. R.; Belanger, K.; Bracken, M. B.; Beckett, W. S.; Leaderer, B. P. (2003)
Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA
J. Am. Med. Assoc. 290: 1859-1867.
George, E. I. (1999) Comment on "Hoeting, J. A.; Madigan, D.; Raftery, A. E.; Volinsky, C. T. 1999. Bayesian
model averaging: a tutorial. Stat. Sci. 14: 409-412."
Gielen, M. H.; VanDer Zee, S. C.; Van Wijnen, J. H.; Van Steen, C. J.; Brunekreef, B. (1997) Acute effects of
summer air pollution on respiratory health of asthmatic children. Am. J. Respir. Crit. Care Med.
155: 2105-2108.
Gilliland, F. D.; Berhane, K.; Rappaport, E. B.; Thomas, D. C.; Avol, E.; Gauderman, W. J.; London, S. J.;
Margolis, H. G.; McConnell, R.; Islam, K. T.; Peters, J. M. (2001) The effects of ambient air pollution on
school absenteeism due to respiratory illnesses. Epidemiology 12: 43-54.
Gold, D. R.; Damokosh, A. I.; Pope, C. A., Ill; Dockery, D. W.; McDonnell, W. F.; Serrano, P.; Retama, A.;
Castillejos, M. (1999) Paniculate and ozone pollutant effects on the respiratory function of children in
southwest Mexico City. Epidemiology 10: 8-16.
Gold, D. R.; Litonjua, A.; Schwartz, J.; Lovett, E.; Larson, A.; Nearing, B.; Allen, G.; Verrier, M.; Cherry, R.;
Verrier, R. (2000) Ambient pollution and heart rate variability. Circulation 101: 1267-1273.
Gold, D. R.; Schwartz, J.; Litonjua, A.; Verrier, R.; Zanobetti, A. (2003) Ambient pollution and reduced heart rate
variability. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
Health Effects Institute; pp. 107-112. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
[18 October, 2004].
Goldberg, M. S.; Burnett, R. T. (2003) Revised analysis of the Montreal time-series study. In: Revised analyses of
time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute;
pp. 113-132. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdftl3 August 2003).
Goldberg, M. S.; Burnett, R. T.; Brook, J.; Bailar, J. C., Ill; Valois, M.-F.; Vincent, R. (2001) Associations between
daily cause-specific mortality and concentrations of ground-level ozone in Montreal, Quebec. Am. J.
Epidemiol. 154: 817-826.
Goldberg, M. S.; Burnett, R. T.; Valois, M.-F.; Flegel, K.; Bailar, J. C., Ill; Brook, J.; Vincent, R.; Radon, K. (2003)
Associations between ambient air pollution and daily mortality among persons with congestive heart failure.
Environ. Res. 91: 8-20.
Gong, H., Jr.; Wong, R.; Sarma, R. J.; Linn, W. S.; Sullivan, E. D.; Shamoo, D. A.; Anderson, K. R.; Prasad, S. B.
(1998a) Cardiovascular effects of ozone exposure in human volunteers. Am. J. Respir. Crit. Care Med.
158: 538-546.
Gong, H., Jr.; Simmons, M. S.; Linn, W. S.; McDonnell, W. F.; Westerdahl, D. (1998b) Relationship between acute
ozone responsiveness and chronic loss of lung function in residents of a high-ozone community. Arch.
Environ. Health 53: 313-319.
Gonzales, M.; Ngo, L.; Hammond, S. K.; Tager, I. (2003) Validation of a questionnaire and microenvironmental
model for estimating past exposures to ozone. Int. J. Environ. Health Res. 13: 249-260.
Goodman, S. N. (2005) The methodologic ozone effect. Epidemiology 16: 430-435.
Goss, C. H.; Newsom, S. A.; Schildcrout, J. S.; Sheppard, L.; Kaufman, J. D. (2004) Effect of ambient air pollution
on pulmonary exacerbations and lung function in cystic fibrosis. Am. J. Respir. Crit. Care Med. 169: 816-821.
7-183
-------
Gouveia, N.; Fletcher, T. (2000a) Respiratory diseases in children and outdoor air pollution in Sao Paulo, Brazil:
a time series analysis. Occup. Environ. Med. 57: 477-483.
Gouveia, N.; Fletcher, T. (2000b) Time series analysis of air pollution and mortality: effects by cause, age and
socioeconomic status. J. Epidemiol. Community Health 54: 750-755.
Gouveia, N.; Bremner, S. A.; Novaes, H. M. D. (2004) Association between ambient air pollution and birth weight
in Sao Paulo, Brazil. J. Epidemiol. Community Health 58: 11-17.
Greer, J. R.; Abbey, D. E.; Burchette, R. J. (1993) Asthma related to occupational and ambient air pollutants in
nonsmokers. J. Occup. Med. 35: 909-915.
Gryparis, A.; Forsberg, B.; Katsouyanni, K.; Analitis, A.; Touloumi, G.; Schwartz, J.; Samoli, E.; Medina, S.;
Anderson, H. R.; Niciu, E. M.; Wichmann, H.-E.; Kriz, B.; Kosnik, M.; Skorkovsky, J.; Vonk, J. M.;
Dortbudak, Z. (2004) Acute effects of ozone on mortality from the "Air Pollution and Health: A European
Approach" project. Am. J. Respir. Crit. Care Med. 170: 1080-1087.
Gwynn, R. C.; Thurston, G. D. (2001) The burden of air pollution: impacts among racial minorities. Environ. Health
Perspect. Suppl. 109(4): 501-506.
Gwynn, R. C.; Burnett, R. T.; Thurston, G. D. (2000) A time-series analysis of acidic paniculate matter and daily
mortality and morbidity in the Buffalo, New York, region. Environ. Health Perspect. 108: 125-133.
Ha, E.-H.; Hong, Y.-C.; Lee, B.-E.; Woo, B.-H.; Schwartz, J.; Christiani, D. C. (2001) Is air pollution a risk factor
for low birth weight in Seoul? Epidemiology 12: 643-648.
Hagen, J. A.; Nafstad, P.; Skrondal, A.; Bjerkly, S.; Magnus, P. (2000) Associations between outdoor air pollutants
and hospitalization for respiratory diseases. Epidemiology 11: 136-140.
Hajat, S.; Haines, A.; Goubet, S. A.; Atkinson, R. W.; Anderson, H. R. (1999) Association of air pollution with daily
GP consultations for asthma and other lower respiratory conditions in London. Thorax 54: 597-605.
Hajat, S.; Anderson, H. R.; Atkinson, R. W.; Haines, A. (2002) Effects of air pollution on general practitioner
consultations for upper respiratory diseases in London. Occup. Environ. Med. 59: 294-299.
Hankinson, J. L.; Odencrantz, J. R.; Fedan, K. B. (1999) Spirometric reference values from a sample of the general
U.S. population. Am. J. Respir. Crit. Care Med. 159: 179-187.
Health Effects Institute. (2003) Revised analyses of time-series studies of air pollution and health. Boston, MA:
Health Effects Institute; special report. Available: Available:
http://www.healtheffects.org/Pubs/TimeSeries.pdf [27 June 2003].
Hedley, A. J.; Wong, C.-M.; Thach, T. Q.; Ma, S.; Lam, T.-H.; Anderson, H. R. (2002) Cardiorespiratory and
all-cause mortality after restrictions on sulphur content of fuel in Hong Kong: an intervention study. Lancet
360: 1646-1652.
Hernandez-Garduno, E.; Perez-Neria, J.; Paccagnella, A. M.; Pina-Garcia, M.; Munguia-Castro, M.;
Catalan-Vazquez, M.; Rojas-Ramos, M. (1997) Air pollution and respiratory health in Mexico City. J. Occup.
Environ. Med. 39: 299-307.
Higgins, I. T. T.; D'Arcy, J. B.; Gibbons, D. I.; Avol, E. L.; Gross, K. B. (1990) Effect of exposures to ambient
ozone on ventilatory lung function in children. Am. Rev. Respir. Dis. 141: 1136-1146.
Hill, A. B. (1965) The environment and disease: association or causation? Proc. R. Soc. Med. 58: 295-300.
Hiltermann, T. J. N.; Stolk, J.; Van derZee, S. C.; Brunekreef, B.; De Bruijne, C. R.; Fischer, P. H.; Ameling, C. B.;
Sterk, P. J.; Hiemstra, P. S.; VanBree, L. (1998) Asthma severity and susceptibility to air pollution. Eur.
Respir. J. 11:686-693.
Hoek, G. (2003) Daily mortality and air pollution in The Netherlands. In: Revised analyses of time-series studies of
air pollution and health. Special report. Boston, MA: Health Effects Institute; pp. 133-142. Available:
http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
Hoek, G.; Brunekreef, B. (1995) Effect of photochemical air pollution on acute respiratory symptoms in children.
Am. J. Respir. Crit. Care Med. 151: 27-32.
Hoek, G.; Brunekreef, B.; Verhoeff, A.; Van Wijnen, J.; Fischer, P. (2000) Daily mortality and air pollution in the
Netherlands. J. Air Waste Manage. Assoc. 50: 1380-1389.
Hoek, G.; Brunekreef, B.; Fischer, P.; Van Wijnen, J. (2001) The association between air pollution and heart failure,
arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time series study.
Epidemiology 12: 355-357.
Hoeting, J. A.; Madigan, D.; Raftery, A. E.; Volinsky, C. T. (1999) Bayesian model averaging: a tutorial. Stat. Sci.
14: 382-417.
Holguin, F.; Tellez-Rojo, M. M.; Hernandez, M.; Cortez, M.; Chow, J. C.; Watson, J. G.; Mannino, D.; Romieu, I.
(2003) Air pollution and heart rate variability among the elderly in Mexico City. Epidemiology 14: 521-527.
7-184
-------
Holmen, A.; Blomqvist, I; Frindberg, H.; Johnelius, Y.; Eriksson, N. E.; Henricson, K. A.; Herrstrom, P.;
Hogstedt, B. (1997) Frequency of patients with acute asthma in relation to ozone, nitrogen dioxide, other
pollutants of ambient air and meteorological observations. Int. Arch. Occup. Environ. Health 69: 317-322.
Hong, Y.-C.; Lee, J.-T.; Kim, H.; Ha, E.-H.; Schwartz, I; Christiani, D. C. (2002) Effects of air pollutants on acute
stroke mortality. Environ. Health Perspect. 110: 187-191.
Hoppe, P.; Praml, G.; Rabe, G.; Lindner, I; Fruhmann, G.; Kessel, R. (1995a) Environmental ozone field study on
pulmonary and subjective responses of assumed risk groups. Environ. Res. 71: 109-121.
Hoppe, P.; Lindner, I; Praml, G.; Bronner, N. (1995b) Effects of environmental ozone on the lung function of senior
citizens. Int. J. Biometeorol. 38: 122-125.
Hoppe, P.; Peters, A.; Rabe, G.; Praml, G.; Lindner, J.; Jakobi, G.; Fruhmann, G.; Nowak, D. (2003) Environmental
ozone effects in different population subgroups. Int. J. Hyg. Environ. Health 206: 505-516.
Horak, F., Jr.; Studnicka, M; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.; Frischer, T. (2002a)
Paniculate matter and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren.
Eur. Respir. J. 19: 838-845.
Horak, F., Jr.; Studnicka, M.; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.; Frischer, T. (2002b)
Paniculate matter and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren
[author response]. Eur. Respir. J. 20: 1355.
Huang, Y.; Dominici, F.; Bell, M. L. (2005) Bayesian hierarchical distributed lag models for summer ozone
exposure and cardio-respiratory mortality. Environmetrics 16: 547-562.
Hubbell, B. J.; Hallberg, A.; McCubbin, D. R.; Post, E. (2005) Health-related benefits of attaining the 8-hr ozone
standard. Environ. Health Perspect. 113: 73-82.
Hwang, J.-S.; Chan, C.-C. (2002) Effects of air pollution on daily clinic visits for lower respiratory tract illness.
Am. J. Epidemiol. 155: 1-10.
iHAPSS. (2005) iHAPSS internet-based health & air pollution surveillance system: city-specific descriptive
statistics. Baltimore, MD: Johns Hopkins Bloomberg School of Public Health, Department of Biostatistics.
Available: http://www.ihapss.jhsph.edu/data/NMMAPS/descriptives/frame.htm [25 January, 2006].
Ihorst, G.; Frischer, T.; Horak, F.; Schumacher, M.; Kopp, M.; Forster, J.; Mattes, J.; Kuehr, J. (2004) Long- and
medium-term ozone effects on lung growth including a broad spectrum of exposure. Eur. Respir. J.
23:292-299.
Ilabaca, M.; Olaeta, I.; Campos, E.; Villaire, J.; Tellez-Rojo, M. M.; Romieu, I. (1999) Association between levels
of fine paniculate and emergency visits for pneumonia and other respiratory illnesses among children in
Santiago, Chile. J. Air Waste Manage. Assoc. 49: 154-163.
Ito, K. (2003) Associations of paniculate matter components with daily mortality and morbidity in Detroit,
Michigan. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
Health Effects Institute; pp. 143-156. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
[12 May, 2004].
Ito, K. (2004) Revised ozone risk estimates for daily mortality and hospitalizations in Detroit, Michigan [personal
communication with attachments to Jee Young Kim]. New York, NY: New York University School of
Medicine, Nelson Institute of Environmental Medicine; October 31.
Ito, K.; Thurston, G. D. (1996) Daily PM10/mortality associations: an investigation of at-risk subpopulations.
J. Exposure Anal. Environ. Epidemiol. 6: 79-95.
Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and
meta-analysis. Epidemiology 16: 446-457.
Jaffe, D. H.; Singer, M. E.; Rimm, A. A. (2003) Air pollution and emergency department visits for asthma among
Ohio Medicaid recipients, 1991-1996. Environ. Res. 91: 21-28.
Jalaludin, B. B.; Chey, T.; O'Toole, B. I.; Smith, W. T.; Capon, A. G.; Leeder, S. R. (2000) Acute effects of low
levels of ambient ozone on peak expiratory flow rate in a cohort of Australian children. Int. J. Epidemiol.
29: 549-557.
Jalaludin, B. B.; O'Toole, B. I.; Leeder, S. R. (2004) Acute effects of urban ambient air pollution on respiratory
symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ
Res. 95: 32-42.
Janes, H.; Sheppard, L.; Lumley, T. (2005) Case-crossover analyses of air pollution exposure data: referent selection
strategies and their implications for bias. Epidemiology 16: 717-726.
Jedrychowski, W.; Maugeri, U.; Bianchi, I.; Flak, E. (2001) Transient or persistent asthma-like symptoms and lung
growth over 2-year follow-up in pre-adolescent children. J. Epidemiol. Biostat. 6: 229-233.
7-185
-------
Jones, M. C. (1983) The projection pursuit algorithm for exploratory data analysis [Ph.D. thesis]. Bath, England:
University of Bath.
Jones, G. N.; Sletten, C.; Mandry, C.; Brantley, P. J. (1995) Ozone level effect on respiratory illness: an
investigation of emergency department visits. South. Med. J. 88: 1049-1056.
Just, J.; Segala, C.; Sahraoui, F.; Priol, G.; Grimfeld, A.; Neukirch, F. (2002) Short-term health effects of paniculate
and photochemical air pollution in asthmatic children. Eur. Respir. J. 20: 899-906.
Kaiser, R.; Romieu, I.; Medina, S.; Schwartz, J.; Krzyzanowski, M.; Ktinzli, N. (2004) Air pollution attributable
postneonatal infant mortality in U.S. metropolitan areas: a risk assessment study. Environmental Health:
A Global Access Science Source 3: 4. Available: http;//www.ehjournal.net/content/3/l/4 [13 January, 2006].
Keatinge, W. R.; Donaldson, G. C. (2005) Heat acclimatization and sunshine cause false indications of mortality due
to ozone. Environ. Res.: in press.
Kim, S.-Y.; Lee, J.-T.; Hong, Y.-C.; Ahn, K.-J.; Kim, H. (2004) Determining the threshold effect of ozone on daily
mortality: an analysis of ozone and mortality in Seoul, Korea, 1995-1999. Environ. Res. 94: 113-119.
Kinney, P. L.; Lippmann, M. (2000) Respiratory effects of seasonal exposures to ozone and particles. Arch. Environ.
Health 55: 210-216.
Kinney, P. L.; Ozkaynak, H. (1991) Associations of daily mortality and air pollution in Los Angeles County.
Environ. Res. 54: 99-120.
Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
Kinney, P. L.; Thurston, G. D.; Raizenne, M. (1996a) The effects of ambient ozone on lung function in children:
a reanalysis of six summer camp studies. Environ. Health Perspect. 104: 170-174.
Kinney, P. L.; Nilsen, D. M.; Lippmann, M.; Brescia, M.; Gordon, T.; McGovern, T.; El Fawal, H.; Devlin, R. B.;
Rom, W. N. (1996b) Biomarkers of lung inflammation in recreational joggers exposed to ozone. Am. J.
Respir. Crit. Care Med. 154: 1430-1435.
Kinney, P. L.; Aggarwal, M.; Nikiforov, S. V.; Nadas, A. (1998) Methods development for epidemiologic
investigations of the health effects of prolonged ozone exposure. Part III: an approach to retrospective
estimation of lifetime ozone exposure using a questionnaire and ambient monitoring data (U.S. sites).
Cambridge, MA: Health Effects Institute; research report no. 81; pp. 79-108.
Klemm, R. J.; Mason, R. M., Jr. (2000) Aerosol Research and Inhalation Epidemiological Study (ARIES): air
quality and daily mortality statistical modeling—interim results. J. Air. Waste Manage. Assoc.
50: 1433-1439.
Klemm, R. J.; Lipfert, F. W.; Wyzga, R. E.; Gust, C. (2004) Daily mortality and air pollution in Atlanta: two years
of data from ARIES. Inhalation Toxicol. 16(suppl. 1): 131-141.
Klepeis, N. E. (1999) An introduction to the indirect exposure assessment approach: modeling human exposure
using microenvironmental measurements and the recent National Human Activity Pattern Survey. Environ.
Health Perspect. Suppl. 107: 365-374.
Kochi, I.; Hubbell, B.; Kramer, R. (2005) An empirical Bayes approach to combining and comparing estimates of
the value of a statistical life for environmental policy analysis. Available:
http://www.nicholas.duke.edu/people/faculty/kramer/VSLrevised%204-6-05.pdf [11 August, 2005].
Koken, P. J.; Piver, W. T.; Ye, F.; Elixhauser, A.; Olsen, L. M.; Portier, C. J. (2003) Temperature, air pollution, and
hospitalization for cardiovascular diseases among elderly people in Denver. Environ. Health Perspect.
Ill: 1312-1317.
Koop, G.; Tole, L. (2004) Measuring the health effects of air pollution: to what extent can we really say that people
are dying from bad air? J. Environ. Econ. Manage. 47: 30-54.
Kopp, M. V.; Ulmer, C.; Ihorst, G.; Seydewitz, H. H.; Frischer, T.; Forster, J.; Kuehr, J. (1999) Upper airway
inflammation in children exposed to ambient ozone and potential signs of adaptation. Eur. Respir. J.
14: 854-861.
Kopp, M. V.; Bohnet, W.; Frischer, T.; Ulmer, C.; Studnicka, M.; Ihorst, G.; Gardner, C.; Forster, J.; Urbanek, R.;
Kuehr, J. (2000) Effects of ambient ozone on lung function in children over a two-summer period. Eur.
Respir. J. 16: 893-900.
Koren, H. S.; Hatch, G. E.; Graham, D. E. (1990) Nasal lavage as a tool in assessing acute inflammation in response
to inhaled pollutants. Toxicology 60: 15-25.
Korrick, S. A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K. D.; Rosner, B. A.;
Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
Environ. Health Perspect. 106: 93-99.
7-186
-------
Krzyzanowski, M; Quackenboss, J. J.; Lebowitz, M. D. (1992) Relation of peak expiratory flow rates and symptoms
to ambient ozone. Arch. Environ. Health 47: 107-115.
Ktinzli, N.; Lurmann, F.; Segal, M.; Ngo, L.; Balmes, J.; Tager, I. B. (1997) Association between lifetime ambient
ozone exposure and pulmonary function in college freshmen—results of a pilot study. Environ. Res. 72: 8-23.
Kuo, H. W.; Lai, J. S.; Lee, M. C.; Tai, R. C.; Lee, M. C. (2002) Respiratory effects of air pollutants among
asthmatics in central Taiwan. Arch. Environ. Health 57: 194-200.
Kwon, H.-J.; Cho, S.-H.; Nyberg, F.; Pershagen, G. (2001) Effects of ambient air pollution on daily mortality in a
cohort of patients with congestive heart failure. Epidemiology 12: 413-419.
Lagerkvist, B. J.; Bernard, A.; Blomberg, A.; Bergstrom, E.; Forsberg, B.; Holmstrom, K.; Karp, K.; Lundstrom,
N.-G.; Segerstedt, B.; Svensson, M.; Nordberg, G. (2004) Pulmonary epithelial integrity in children:
relationship to ambient ozone exposure and swimming pool attendance. Environ. Health Perspect.
112: 1768-1771.
Langstaff, J. (2003) Percentiles of 1996-2000 ozone concentrations [memorandum to Joe Pinto]. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; September
17.
Lebowitz, M. D.; Camilli, A. E.; Bronnimann, D.; Quackenboss, J. (1987) The significance and meaningfulness of
intraindividual changes in objective test results as responses to air contaminants. Presented at: 80th annual
meeting of the Air Pollution Control Association; June; New York, NY. Pittsburgh, PA: Air Pollution Control
Association; paper no. 87-32.1.
Lebowitz, M. D.; Quackenboss, J. J.; Krzyzanowski, M. (1991) Acute respiratory effects of prolonged ambient
ozone. In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment:
papers from an international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste
Management Association; pp. 111-119. (A&WMA transactions series no. TR-19).
Lee, J.-T.; Schwartz, J. (1999) Reanalysis of the effects of air pollution on daily mortality in Seoul, Korea:
a case-crossover design. Environ. Health Perspect. 107: 633-636.
Lee, J.-T.; Shin, D.; Chung, Y. (1999) Air pollution and daily mortality in Seoul and Ulsan, Korea. Environ. Health
Perspect. 107: 149-154.
Lee, J.-T.; Kim, H.; Song, H.; Hong, Y.-C; Cho, Y.-S.; Shin, S.-Y.; Hyun, Y.-J.; Kim, Y.-S. (2002) Air pollution
and asthma among children in Seoul, Korea. Epidemiology 13: 481-484.
Lee, K.; Parkhurst, W. J.; Xue, J.; Ozkaynak, H.; Neuberg, D.; Spengler, J. D. (2004) Outdoor/indoor/presonal ozone
exposures of children in Nashville, Tennessee. J. Air Waste Manage. Assoc. 54: 352-359.
Le Tertre, A.; Quenel, P.; Eilstein, D.; Medina, S.; Prouvost, H.; Pascal, L.; Boumghar, A.; Saviuc, P.; Zeghnoun,
A.; Filleul, L.; Declercq, C.; Cassadou, S.; Le Coaster, C. (2002a) Short-term effects of air pollution on
mortality in nine French cities: a quantitative summary. Arch. Environ. Health 57: 311-319.
Le Tertre, A.; Medina, S.; Samoli, E.; Forsberg, B.; Michelozzi, P.; Boumghar, A.; Vonk, J. M.; Bellini, A.;
Atkinson, R.; Ayres, J. G.; Sunyer, J.; Schwartz, J.; Katsouyanni, K. (2002b) Short term effects of paniculate
air pollution on cardiovascular diseases in eight European cities. J. Epidemiol. Community Health
56: 773-779.
Levy, J. I.; Carrothers, T. J.; Tuomisto, J. T.; Hammitt, J. K.; Evans, J. S. (2001) Assessing the public health benefits
of reduced ozone concentrations. Environ. Health Perspect. 109: 1215-1226.
Levy, J. L; Chemerynski, S. M.; Sarnat, J. A. (2005) Ozone exposure and mortality, an empiric Bayes
metaregression analysis. Epidemiology 16: 458-468.
Liao, D.; Duan, Y.; Whitsel, E. A.; Zheng, Z.-J.; Heiss, G.; Chinchilli, V. M.; Lin, H.-M. (2004) Association of
higher levels of ambient criteria pollutants with impaired cardiac autonomic control: a population-based
study. Am. J. Epidemiol. 159: 768-777.
Lierl, M. B.; Hornung, R. W. (2003) Relationship of outdoor air quality to pediatric asthma exacerbations. Ann.
Allergy Asthma Immunol. 90: 28-33.
Lin, C. A.; Martins, M. A.; Farhat, S. C. L.; Pope, C. A., Ill; Conceicao, G. M. S.; Anastacio, V. M.; Hatanaka, M.;
Andrade, W. C.; Hamaue, W. R.; Bohm, G. M.; Saldiva, P. H. N. (1999) Air pollution and respiratory illness
of children in Sao Paulo, Brazil. Paediatr. Perinat. Epidemiol. 13: 475-488.
Lin, M.-C; Yu, H.-S.; Tsai, S.-S.; Cheng, B.-H.; Hsu, T.-Y.; Wu, T.-N.; Yang, C.-Y. (2001) Adverse pregnancy
outcome in a petrochemical polluted area in Taiwan. J. Toxicol. Environ. Health Part A 63: 565-574.
Lin, M.; Chen, Y.; Burnett, R. T.; Villeneuve, P. J.; Krewski, D. (2003) Effect of short-term exposure to gaseous
pollution on asthma hospitalisation in children: a bi-directional case-crossover analysis. J. Epidemiol.
Community Health 57: 50-55.
7-187
-------
Lin, M; Chen, Y.; Villeneuve, P. I; Burnett, R. T.; Lemyre, L.; Hertzman, C.; McGrail, K. M; Krewski, D. (2004)
Gaseous air pollutants and asthma hospitalization of children with low household income in Vancouver,
British Columbia, Canada. Am. J. Epidemiol. 159: 294-303.
Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996)
Short-term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal.
Environ. Epidemiol. 6: 449-472.
Linn, W. S.; Szlachcic, Y.; Gong, H., Jr.; Kinney, P. L.; Berhane, K. T. (2000) Air pollution and daily hospital
admissions in metropolitan Los Angeles. Environ. Health Perspect. 108: 427-434.
Lipfert, F. W.; Hammerstrom, T. (1992) Temporal patterns in air pollution and hospital admissions. Environ. Res.
59: 374-399.
Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000a) Daily mortality in the Philadelphia metropolitan area and
size-classified paniculate matter. J. Air Waste Manage. Assoc. 50: 1501-1513.
Lipfert, F. W.; Perry, H. M., Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2000b) The Washington
University-EPRI veterans' cohort mortality study: preliminary results. In: Grant, L. D., ed. PM2000:
paniculate matter and health. Inhalation Toxicol. 12(suppl. 4): 41-73.
Lipfert, F. W.; Perry, H. M., Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2003) Air pollution, blood
pressure, and their long-term associations with mortality. Inhalation Toxicol. 15: 493-512.
Lippmann, M. (1988) Health significance of pulmonary function responses to airborne irritants. JAPCA
38: 881-887.
Lippmann, M.; Spektor, D. M. (1998) Peak flow rate changes in O3 exposed children: spirometry vs. miniWright
flow meters. J. Exposure Anal. Environ. Epidemiol. 8: 101-107.
Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report
no. 95.
Lipsett, M.; Hurley, S.; Ostro, B. (1997) Air pollution and emergency room visits for asthma in Santa Clara County,
California. Environ. Health Perspect. 105: 216-222.
Liu, L.-J. S.; Koutrakis, P.; Leech, J.; Broder, I. (1995) Assessment of ozone exposures in the greater metropolitan
Toronto area. J. Air Waste Manage. Assoc. 45: 223-234.
Loomis, D. P.; Borja-Aburto, V. H.; Bangdiwala, S. I.; Shy, C. M. (1996) Ozone exposure and daily mortality in
Mexico City: a time-series analysis. Cambridge, MA: Health Effects Institute; research report no. 75.
Loomis, D.; Castillejos, M.; Gold, D. R.; McDonnell, W.; Borja-Aburto, V. H. (1999) Air pollution and infant
mortality in Mexico City. Epidemiology 10: 118-123.
Luginaah, I. N.; Fung, K. Y.; Gorey, K. M.; Webster, G.; Wills, C. (2005) Association of Ambient Air Pollution
with Respiratory Hospitalization in a Government Designated "Area of Concern": The Case of Windsor,
Ontario. Environ. Health Perspect. 113: 290-296.
Lumley, T.; Sheppard, L. (2000) Assessing seasonal confounding and model selection bias in air pollution
epidemiology using positive and negative control analyses. Environmetrics 11: 705-717.
Maisonet, M.; Bush, T. J.; Correa, A.; Jaakkola, J. J. K. (2001) Relation between ambient air pollution and low birth
weight in the northeastern United States. Environ. Health Perspect. Suppl. 109(3): 351-356.
Mann, J. K.; Tager, I. B.; Lurmann, F.; Segal, M.; Quesenberry, C. P., Jr.; Lugg, M. M.; Shan, J.; Van denEeden,
S. K. (2002) Air pollution and hospital admissions for ischemic heart disease in persons with congestive heart
failure or arrhythmia. Environ. Health Perspect. 110: 1247-1252.
Martins, L. C.; Latorre, M. R. D. O.; Saldiva, P. H. N.; Braga, A. L. F. (2002) Air pollution and emergency room
visits due to chronic lower respiratory diseases in the elderly: an ecological time-series study in Sao Paulo,
Brazil. J. Occup. Environ. Med. 44: 622-627.
McConnell, R.; Berhane, K.; Gilliland, F.; London, S. J.; Vora, H.; Avol, E.; Gauderman, W. J.; Margolis, H. G.;
Lurmann, F.; Thomas, D. C.; Peters, J. M. (1999) Air pollution and bronchitic symptoms in southern
California children with asthma. Environ. Health Perspect. 107: 757-760.
McConnell, R.; Berhane, K.; Gilliland, F.; London, S. J.; Islam, T.; Gauderman, W. J.; Avol, E.; Margolis, H. G.;
Peters, J. M. (2002) Asthma in exercising children exposed to ozone: a cohort study. Lancet 359: 386-391.
McConnell, R.; Berhane, K.; Gilliland, F.; Molitor, J.; Thomas, D.; Lurmann, F.; Avol, E.; Gauderman, W. J.;
Peters, J. M. (2003) Prospective study of air pollution and bronchitic symptoms in children with asthma. Am.
J. Respir. Crit. Care Med. 168:790-797.
McDonnell, W. F.; Abbey, D. E.; Nishino, N.; Lebowitz, M. D. (1999) Long-term ambient ozone concentration and
the incidence of asthma in nonsmoking adults: the ahsmog study. Environ. Res. 80: 110-121.
7-188
-------
Metzger, K. B.; Tolbert, P. E.; Klein, M.; Peel, J. L.; Flanders, W. D.; Todd, K. H.; Mulholland, J. A.; Ryan, P. B.;
Framkin, H. (2004) Ambient air pollution and cardiovascular emergency department visits. Epidemiology
15: 46-56.
Mitchell, H.; Senturia, Y.; Gergen, P.; Baker, D.; Joseph, C.; McNiff-Mortimer, K.; Wedner, H. J.; Grain, E.;
Eggleston, P.; Evans, R., Ill; Kattan, M.; Kercsmar, C.; Leickly, F.; Malveaux, F.; Smartt, E.; Weiss, K.
(1997) Design and methods of the National Cooperative Inner-City Asthma Study. Pediatr. Pulmonol.
24: 237-252.
Moolgavkar, S. H. (2003) Air pollution and daily mortality in two U.S. counties: season-specific analyses and
exposure-response relationships. Inhalation Toxicol. 15: 877-907.
Moolgavkar, S. H.; Luebeck, E. G. (1996) A critical review of the evidence on paniculate air pollution and
mortality. Epidemiology 7: 420-428.
Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A.; Anderson, E. L. (1995) Air pollution and daily mortality in
Philadelphia. Epidemiology 6: 476-484.
Moolgavkar, S. H.; Luebeck, E. G.; Anderson, E. L. (1997) Air pollution and hospital admissions for respiratory
causes in Minneapolis-St. Paul and Birmingham. Epidemiology 8: 364-370.
Morgan, G.; Corbett, S.; Wlodarczyk, J. (1998a) Air pollution and hospital admissions in Sydney, Australia, 1990 to
1994. Am. J. Public Health 88: 1761-1766.
Morgan, G.; Corbett, S.; Wlodarczyk, J.; Lewis, P. (1998b) Air pollution and daily mortality in Sydney, Australia,
1989 through 1993. Am. J. Public Health 88: 759-764.
Mortimer, K. M.; Tager, I. B.; Dockery, D. W.; Neas, L. M.; Redline, S. (2000) The effect of ozone on inner-city
children with asthma: identification of susceptible subgroups. Am. J. Respir. Crit. Care Med. 162: 1838-1845.
Mortimer, K. M.; Neas, L. M.; Dockery, D. W.; Redline, S.; Tager, I. B. (2002) The effect of air pollution on
inner-city children with asthma. Eur. Respir. J. 19: 699-705.
Naeher, L. P.; Holford, T. R.; Beckett, W. S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
Healthy women's PEF variations with ambient summer concentrations of PM10, PN2 5, SO42", H+, and O3. Am.
J. Respir. Crit. Care Med. 160: 117-125.
National Institutes of Health. (1997) Guidelines for the diagnosis and management of asthma: expert panel report 2.
Bethesda, MD: U.S. Department of Health and Human Services, National Heart, Lung, and Blood Institute;
publication no. 97-4051. Available: http://www.nhlbi.nih.gov/guidelines/asthma/asthgdln.pdf (11 April 2003).
National Toxicology Program. (1994) Toxicology and carcinogenesis: studies of ozone (CAS No. 10028-15-6) and
ozone/NNK (CAS No. 10028-15-6/64091-91-4) in F344/N rats and B6C3F! mice. Research Triangle Park,
NC: U.S. Department of Health and Human Services, National Institutes of Health; NIH publication no.
95-3371. (National Toxicology Program technical report no. 440).
Nauenberg, E.; Basu, K. (1999) Effect of insurance coverage on the relationship between asthma hospitalizations
and exposure to air pollution. Public Health Rep. 114: 135-148.
Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol.
141: 111-122.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Speizer, F. E. (1999) Fine particles and peak flow in children: acidity
versus mass. Epidemiology 10: 550-553.
Neidell, M. J. (2004) Air pollution, health, and socio-economic status: the effect of outdoor air quality on childhood
asthma. J. Health Econ. 23: 1209-1236.
Newhouse, C. P.; Levetin, B. S.; Levetin, E. (2004) Correlation of environmental factors with asthma and rhinitis
symptoms in Tulsa, OK. Ann. Allergy Asthma Immunol. 92: 356-366.
Nutman, A.; Solomon, Y.; Mendel, S.; Nutman, J.; Hines, E.; Topilsky, M.; Kivity, S. (1998) The use of a neural
network for studying the relationship between air pollution and asthma-related emergency room visits. Respir.
Med. 92: 1199-1202.
Oftedal, B.; Nafstad, P.; Magnus, P.; Bjerkly, S.; Skrondal, A. (2003) Traffic related air pollution and acute hospital
admission for respiratory diseases inDrammen, Norway 1995-2000. Eur. J. Epidemiol. 18: 671-675.
O'Neill, M. S.; Loomis, D.; Borja-Aburto, V. H. (2004) Ozone, area social conditions, and mortality in Mexico City.
Environ. Res. 94: 234-242.
Ostro, B. (1995) Fine paniculate air pollution and mortality in two Southern California counties. Environ. Res.
70: 98-104.
7-189
-------
Ostro, B.; Sanchez, J. M; Aranda, C.; Eskeland, G. S. (1996) Air pollution and mortality: results from a study of
Santiago, Chile. In: Lippmann, M., ed. Papers from the ISEA-ISEE annual meeting; September 1994;
Research Triangle Park, NC. J. Exposure Anal. Environ. Epidemiol. 6: 97-114.
Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coachella
Valley, California: a follow-up study. J. Exposure Anal. Environ. Epidemiol. 10: 412-419.
Ostro, B.; Lipsett, M.; Mann, J.; Braxton-Owens, H.; White, M. (2001) Air pollution and exacerbation of asthma in
African-American children in Los Angeles. Epidemiology 12: 200-208.
Palli, D.; Saieva, C.; Grechi, D.; Masala, G.; Zanna, I.; Barbara, A.; Decarli, A.; Munnia, A.; Peluso, M. (2004)
DNA bulky adducts in a Mediterranean population correlate with environmental ozone concentration, an
indicator of photochemical smog. Int. J. Cancer 109: 17-23.
Park, S. K.; O'Neill, M. S.; Vokonas, P. S.; Sparrow, D.; Schwartz, J. (2005) Effects of air pollution on heart rate
variability: the VA normative aging study. Environ. HealthPerspect. 113: 304-309.
Pearce, N.; Beasley, R.; Burgess, C.; Crane, J. (1998) Asthma epidemiology: principles and methods. New York,
NY: Oxford University Press.
Peel, J. L.; Tolbert, P. E.; Klein, M.; Metzger, K. B.; Flanders, W. D.; Knox, T.; Mulholland, J. A.; Ryan, P. B.;
Frumkin, H. (2005) Ambient air pollution and respiratory emergency department visits. Epidemiology
16: 164-174.
Pereira, L. A. A.; Loomis, D.; Conceicao, G. M. S.; Braga, A. L. F.; Areas, R. M.; Kishi, H. S.; Singer, J. M.; Bohm,
G. M.; Saldiva, P. H. N. (1998) Association between air pollution and intrauterine mortality in Sao Paulo,
Brazil. Environ. Health Perspect. 106: 325-329.
Pereira, F. A. C.; De Assuncao, J. V.; Saldiva, P. H. N.; Pereira, L. A. A.; Mirra, A. P.; Braga, A. L. F. (2005)
Influence of air pollution on the incidence of respiratory tract neoplasm. J. Air Waste Manage. Assoc.
55: 83-87.
Peters, J. M.; Avol, E.; Navidi, W.; London, S. J.; Gauderman, W. J.; Lurmann, F.; Linn, W. S.; Margolis, H.;
Rappaport, E.; Gong, H., Jr.; Thomas, D. C. (1999a) A study of twelve southern California communities with
differing levels and types of air pollution. I. Prevalence of respiratory morbidity. Am. J. Respir. Crit. Care
Med. 159: 760-767.
Peters, J. M.; Avol, E.; Gauderman, W. J.; Linn, W. S.; Navidi, W.; London, S. J.; Margolis, H.; Rappaport, E.;
Vora, H.; Gong, H., Jr.; Thomas, D. C. (1999b) A study of twelve southern California communities with
differing levels and types of air pollution. II. Effects on pulmonary function. Am. J. Respir. Crit. Care Med.
159: 768-775.
Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M.; Baliff, J.; Oh, J. A.; Allen, G.;
Monahan, K.; Dockery, D. W. (2000a) Air pollution and incidence of cardiac arrhythmia. Epidemiology
11: 11-17.
Peters, A.; Skorkovsky, J.; Kotesovec, F.; Brynda, J.; Spix, C.; Wichmann, H. E.; Heinrich, J. (2000b) Associations
between mortality and air pollution in central Europe. Environ. Health Perspect. 108: 283-287.
Peters, A.; Dockery, D. W.; Muller, J. E.; Mittleman, M. A. (2001) Increased paniculate air pollution and the
triggering of myocardial infarction. Circulation 103: 2810-2815.
Petroeschevsky, A.; Simpson, R. W.; Thalib, L.; Rutherford, S. (2001) Associations between outdoor air pollution
and hospital admissions in Brisbane, Australia. Arch. Environ. Health 56: 37-52.
Plunkett, L. M.; Turnbull, D.; Rodricks, J. V. (1992) Differences between adults and children affecting exposure
assessment. In: Guzelian, P. S.; Henry, D. J.; Olin, S. S., eds. Similarities and differences between children
and adults: implications for risk assessment. Washington, DC: ILSI Press, pp. 79-96.
Poloniecki, J. D.; Atkinson, R. W.; Ponce de Leon, A.; Anderson, H. R. (1997) Daily time series for cardiovascular
hospital admissions and previous day's air pollution in London, UK. Occup. Environ. Med. 54: 535-540.
Ponce de Leon, A.; Anderson, H. R.; Bland, J. M.; Strachan, D. P.; Bower, J. (1996) Effects of air pollution on daily
hospital admissions for respiratory disease in London between 1987-88 and 1991-92. In: St Leger, S., ed. The
APHEA project. Short term effects of air pollution on health: a European approach using epidemiological
time series data. J. Epidemiol. Commun. Health 50(suppl. 1): S63-S70.
Ponka, A.; Virtanen, M. (1996) Asthma and ambient air pollution in Helsinki. In: St Leger, S., ed. The APHEA
project. Short term effects of air pollution on health: a European approach using epidemiological time series
data. J. Epidemiol. Community Health 50(suppl. 1): S59-S62.
Ponka, A.; Savela, M.; Virtanen, M. (1998) Mortality and air pollution in Helsinki. Arch. Environ. Health
53:281-286.
Pope, C. A., Ill; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA J. Am. Med.
Assoc. 287: 1132-1141.
7-190
-------
Prescott, G. I; Cohen, G. R.; Elton, R. A.; Fowkes, F. G. R.; Agius, R. M. (1998) Urban air pollution and
cardiopulmonary ill health: a 14.5 year time series study. Occup. Environ. Med. 55: 697-704.
Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. (1987) Acute respiratory function and transported air pollutants:
observational studies. Presented at: 80th annual meeting of the Air Pollution Control Association; June;
New York, NY. Pittsburgh, PA: Air Pollution Control Association; paper no. 87-32.6.
Raizenne, M. E.; Burnett, R. T.; Stern, B.; Franklin, C. A.; Spengler, J. D. (1989) Acute lung function responses to
ambient acid aerosol exposures in children. Environ. Health Perspect. 79: 179-185.
Ramadour, M.; Burel, C.; Lanteaume, A.; Vervloet, D.; Charpin, D.; Brisse, F.; Dutau, H.; Charpin, D. (2000)
Prevalence of asthma and rhinitis in relation to long-term exposure to gaseous air pollutants. Allergy
(Copenhagen) 55: 1163-1169.
Ramsay, T. O.; Burnett, R. T.; Krewski, D. (2003) The effect of concurvity in generalized additive models linking
mortality to ambient paniculate matter. Epidemiology 14: 18-23.
Rich, K. E.; Petkau, J.; Vedal, S.; Brauer, M. (2004) A case-crossover analysis of paniculate air pollution and
cardiac arrhythmia in patients with implantable cardioverter defibrillators. Inhalation Toxicol. 16: 363-372.
Rich, D. Q.; Schwartz, J.; Mittleman, M. A.; Link, M.; Luttmann-Gibson, H.; Catalano, P. J.; Speizer, F. E.;
Dockery, D. W. (2005) Association of short-term ambient air pollution concentrations and ventricular
arrhythmias. Am. J. Epidemiol. 161: 1123-1132.
Ritz, B.; Yu, F. (1999) The effect of ambient carbon monoxide on low birth weight among children born in southern
California between 1989 and 1993. Environ. Health Perspect. 107: 17-25.
Ritz, B.; Yu, F.; Chapa, G.; Fruin, S. (2000) Effect of air pollution on preterm birth among children born in Southern
California between 1989 and 1993. Epidemiology 11: 502-511.
Ritz, B.; Yu, F.; Fruin, S.; Chapa, G.; Shaw, G. M.; Harris, J. A. (2002) Ambient air pollution and risk of birth
defects in Southern California. Am. J. Epidemiol. 155: 17-25.
Roemer, W. H.; Van Wijnen, J. H . (2001) Daily mortality and air pollution along busy streets in Amsterdam,
1987-1998. Epidemiology 12: 649-653.
Romieu, I.; Meneses, F.; Sienra-Monge, J. J. L.; Huerta, J.; Velasco, S. R.; White, M. C.; Etzel, R. A.;
Hernandez-Avila, M. (1995) Effects of urban air pollutants on emergency visits for childhood asthma in
Mexico City. Am. J. Epidemiol. 141: 546-553.
Romieu, L; Meneses, F.; Ruiz, S.; Sienra, J. J.; Huerta, J.; White, M. C.; Etzel, R. A. (1996) Effects of air pollution
on the respiratory health of asthmatic children living in Mexico City. Am. J. Respir. Crit. Care Med.
154: 300-307.
Romieu, L; Meneses, F.; Ruiz, S.; Huerta, J.; Sienra, J. J.; White, M.; Etzel, R.; Hernandez, M. (1997) Effects of
intermittent ozone exposure on peak expiratory flow and respiratory symptoms among asthmatic children in
Mexico City. Arch. Environ. Health 52: 368-376.
Romieu, L; Meneses, F.; Ramirez, M.; Ruiz, S.; Padilla, R. P.; Sienra, J. J.; Gerber, M.; Grievink, L.; Dekker, R.;
Walda, L; Brunekreef, B. (1998) Antioxidant supplementation and respiratory functions among workers
exposed to high levels of ozone. Am. J. Respir. Crit. Care Med. 158: 226-232.
Romieu, L; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Tellez-Rojo, M. M.; Moreno-Macias, H.; Reyes-Ruiz, N. L;
Del Rio-Navarro, B. E.; Ruiz-Navarro, M. X.; Hatch, G.; Slade, R.; Hernandez-Avila, M. (2002) Antioxidant
supplementation and lung functions among children with asthma exposed to high levels of air pollutants.
Am. J. Respir. Crit. Care Med. 166: 703-709.
Romieu, L; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. L; Estela del
Rio-Navarro, B.; Hernandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and
antioxidant supplementation influence lung function in relation to ozone exposure in asthmatic children in
Mexico City. Thorax 59: 8-10.
Rona, R. J.; Gulliford, M. C.; Chinn, S. (1993) Effects of prematurity and intrauterine growth on respiratory health
and lung function in childhood. Br. Med. J. 306: 817-820.
Rondeau, V.; Berhane, K.; Thomas, D. C. (2005) A three-level model for binary time-series data: the effects of air
pollution on school absences in the southern California Children's Health Study. Stat. Med. 24: 1103-1115.
Ross, M. A.; Persky, V. W.; Scheff, P. A.; Chung, J.; Curtis, L.; Ramakrishnan, V.; Wadden, R. A.; Hryhorczuk,
D. O. (2002) Effect of ozone and aeroallergens on the respiratory health of asthmatics. Arch. Environ. Health
57: 568-578.
Ruidavets, J.-B.; Cournot, M.; Cassadou, S.; Giroux, M.; Meybeck, M.; Ferrieres, J. (2005) Ozone air pollution is
associated with acute myocardial infarction. Circulation 111: 563-569.
Saez, M.; Tobias, A.; Munoz, P.; Campbell, M. J. (1999) A GEE moving average analysis of the relationship
between air pollution and mortality for asthma in Barcelona, Spain. Stat. Med. 18: 2077-2086.
7-191
-------
Saez, M; Ballester, F.; Barcelo, M. A.; Perez-Hoyos, S.; Bellido, I; Tenias, J. M.; Ocafia, R.; Figueiras, A.;
Arribas, F.; Aragones, N.; Tobias, A.; Cirera, L.; Canada, A.; on behalf of the EMECAM Group. (2002)
A combined analysis of the short-term effects of photochemical air pollutants on mortality within the
EMECAM project. Environ. Health Perspect. 110: 221-228.
Saldiva, P. H. N.; Lichtenfels, A. J. F. C.; Paiva, P. S. O.; Barone, I. A.; Martins, M. A.; Massad, E.; Pereira,
J. C. R.; Xavier, V. P.; Singer, J. M.; Bohm, G. M. (1994) Association between air pollution and mortality
due to respiratory diseases in children in Sao Paulo, Brazil: a preliminary report. Environ. Res. 65: 218-225.
Saldiva, P. H. N.; Pope, C. A., Ill; Schwartz, J.; Dockery, D. W.; Lichtenfels, A. J.; Salge, J. M.; Barone, L; Bohm,
G. M. (1995) Air pollution and mortality in elderly people: a time-series study in Sao Paulo, Brazil. Arch.
Environ. Health 50: 159-163.
Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, L; Dockery, D. W.; Schwartz, J.; Zanobetti, A.
(2000) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94, part II.
Sarnat, J. A.; Schwartz, J.; Catalano, P. J.; Suh, H. H. (2001) Gaseous pollutants in paniculate matter epidemiology:
confounders or surrogates? Environ. Health Perspect. 109: 1053-1061.
Sarnat, J. A.; Brown, K. W.; Schwartz, J.; Coull, B. A.; Koutrakis, P. (2005) Ambient gas concentrations and
personal paniculate matter exposures: implications for studying the health effects of particles. Epidemiology
16: 385-395.
Sartor, F.; Snacken, R.; Demuth, C.; Walckiers, D. (1995) Temperature, ambient ozone levels, and mortality during
summer, 1994, in Belgium. Environ. Res. 70: 105-113.
Scarlett, J. F.; Abbott, K. J.; Peacock, J. L.; Strachan, D. P.; Anderson, H. R. (1996) Acute effects of summer air
pollution on respiratory function in primary school children in southern England. Thorax 51:1109-1114.
Schindler, C.; Ktinzli, N.; Bongard, J.-P.; Leuenberger, P.; Karrer, W.; Rapp, R.; Monn, C.; Ackermann-Liebrich,
U.; Swiss Study on Air Pollution and Lung Diseases in Adults Investigators. (2001) Short-term variation in
air pollution and in average lung function among never-smokers. Am. J. Respir. Crit. Care Med.
163: 356-361.
Schouten, J. P.; Vonk, J. M.; de Graaf, A. (1996) Short term effects of air pollution on emergency hospital
admissions for respiratory disease: results of the APHEA project in two major cities in The Netherlands,
1977-89. In: St Leger, S., ed. The APHEA project. Short term effects of air pollution on health: a European
approach using epidemiological time series data. J. Epidemiol. Community Health 50(suppl. 1): S22-S29.
Schwartz, J. (1991) Paniculate air pollution and daily mortality in Detroit. Environ. Res. 56: 204-213.
Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease. Epidemiology 7: 20-28.
Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature? Am.
J. Respir. Crit. Care Med. 171: 627-631.
Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.; Ponce de
Leon, A.; Ponka, A.; Rossi, G.; Saez, M.; Schouten, J. P. (1996) Methodological issues in studies of air
pollution and daily counts of deaths or hospital admissions. In: St Leger, S., ed. The APHEA project. Short
term effects of air pollution on health: a European approach using epidemiological time series data.
J. Epidemiol. Commun. Health 50(suppl. 1): S3-S11.
Schwartz, J.; Litonjua, A.; Suh, H.; Verrier, M.; Zanobetti, A.; Syring, M.; Nearing, B.; Verrier, R.; Stone, P.;
MacCallum, G.; Speizer, F. E.; Gold, D. R. (2005) Traffic related pollution and heart rate variability in a
panel of elderly subjects. Thorax 60: 455-461.
Selwyn, B. J.; Stock, T. H.; Hardy, R. J.; Chan, F. A.; Jenkins, D. E.; Kotchmar, D. J.; Chapman, R. S. (1985)
Health effects of ambient ozone exposure in vigorously exercising adults. In: Lee, S. D., ed. Evaluation of the
scientific basis for ozone/oxidants standards: proceedings of an APCA international specialty conference;
November 1984; Houston, TX. Pittsburgh, PA: Air Pollution Control Association; pp. 281-296. (APCA
international specialty conference transactions: TR-4).
Sheppard, L. (2003) Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington,
1987-1994. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston,
MA: Health Effects Institute; pp. 227-230. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
[18 October, 2004].
Sheppard, L. (2005) Acute air pollution effects: consequences of exposure distribution and measurements.
J. Toxicol. Environ. Health Part A 68: 1127-1135.
Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air pollution on
nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
7-192
-------
Sheppard, L.; Slaughter, J. C.; Schildcrout, J.; Liu, L.-J. S.; Lumley, T. (2005) Exposure and measurement
contributions to estimates of acute air pollution effects. J. Exposure Anal. Environ. Epidemiol. 15: 366-376.
Shumway, R. H.; Azari, A. S.; Pawitan, Y. (1988) Modeling mortality fluctuations in Los Angeles as functions of
pollution and weather effects. Environ. Res. 45: 224-241.
Silverman, B. W. (1986) Density estimation for statistics and data analysis. New York, NY: Chapman & Hall;
pp. 137-141. [Monographs on Statistics and Applied Probability].
Simpson, R. W.; Williams, G.; Petroeschevsky, A.; Morgan, G.; Rutherford, S. (1997) Associations between outdoor
air pollution and daily mortality in Brisbane, Australia. Arch. Environ. Health 52: 442-454.
Spektor, D. M.; Lippmann, M.; Lioy, P. J.; Thurston, G. D.; Citak, K.; James, D. J.; Bock, N.; Speizer, F. E.;
Hayes, C. (1988a) Effects of ambient ozone on respiratory function in active, normal children. Am. Rev.
Respir. Dis. 137:313-320.
Spektor, D. M.; Lippmann, M.; Thurston, G. D.; Lioy, P. J.; Stecko, J.; O'Connor, G.; Garshick, E.; Speizer, F. E.;
Hayes, C. (1988b) Effects of ambient ozone on respiratory function in healthy adults exercising outdoors.
Am. Rev. Respir. Dis. 138: 821-828.
Spektor, D. M.; Lippmann, M. (1991) Health effects of ambient ozone on healthy children at a summer camp. In:
Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
Association; pp. 83-89. (A&WMA transactions series no. TR-19).
Stieb, D. M.; Burnett, R. T.; Beveridge, R. C.; Brook, J. R. (1996) Association between ozone and asthma
emergency department visits in Saint John, New Brunswick, Canada. Environ. Health Perspect.
104: 1354-1360.
Stieb, D. M.; Judek, S.; Burnett, R. T. (2002) Meta-analysis of time-series studies of air pollution and mortality:
effects of gases and particles and the influence of cause of death, age, and season. J. Air Waste Manage.
Assoc. 52: 470-484.
Stieb, D. M.; Judek, S.; Burnett, R. T. (2003) Meta-analysis of time-series studies of air pollution and mortality:
update in relation to the use of generalized additive models. J. Air Waste Manage. 53: 258-261.
Sunyer, J.; Basagana, X. (2001) Particles, and not gases, are associated with the risk of death in patients with chronic
obstructive pulmonary disease. Int. J. Epidemiol. 30: 1138-1140.
Sunyer, J.; Castellsague, J.; Saez, M.; Tobias, A.; Anto, J. M. (1996) Air pollution and mortality in Barcelona. In:
St Leger, S., ed. The APHEA project. Short term effects of air pollution on health: a European approach using
epidemiological time series data. J. Epidemiol. Community Health 50(suppl. 1): S76-S80.
Sunyer, J.; Spix, C.; Quenel, P.; Ponce-de-Leon, A.; Ponka, A.; Barumandzadeh, T.; Touloumi, G.; Bacharova, L.;
Wojtyniak, B.; Vonk, J.; Bisanti, L.; Schwartz, J.; Katsouyanni, K. (1997) Urban air pollution and emergency
admissions for asthma in four European cities: the APHEA project. Thorax 52: 760-765.
Sunyer, J.; Basagana, X.; Belmonte, J.; Anto, J. M. (2002) Effect of nitrogen dioxide and ozone on the risk of dying
in patients with severe asthma. Thorax 57: 687-693.
Tager, I. B. (1999) Air pollution and lung function growth. Is it ozone? [editorial]. Am. J. Respir. Crit. Care Med.
160: 387-389.
Tager, I. B.; Ktinzli, N.; Lurmann, F.; Ngo, L.; Segal, M.; Balmes, J. (1998) Methods development for
epidemiologic investigations of the health effects of prolonged ozone exposure. Part II: an approach to
retrospective estimation of lifetime ozone exposure using a questionnaire and ambient monitoring data
(California sites). Cambridge, MA: Health Effects Institute; research report no. 81; pp. 27-78.
Tager, I. B.; Balmes, J.; Lurmann, F.; Ngo, L.; Alcorn, S.; Ktinzli, N. (2005) Chronic exposure to ambient ozone and
lung function in young adults. Epidemiology 16: 751-759.
Taggart, S. C. O.; Custovic, A.; Francis, H. C.; Faragher, E. B.; Yates, C. J.; Higgins, B. G.; Woodcock, A. (1996)
Asthmatic bronchial hyperresponsiveness varies with ambient levels of summertime air pollution. Eur.
Respir. J. 9: 1146-1154.
Tellez-Rojo, M. M.; Romieu, I.; Ruiz-Velasco, S.; Lezana, M.-A.; Hernandez-Avila, M. M. (2000) Daily respiratory
mortality and PMio pollution in Mexico City: importance of considering place of death. Eur. Respir. J.
16: 391-396.
Tenias, J. M.; Ballester, F.; Rivera, M. L. (1998) Association between hospital emergency visits for asthma and air
pollution in Valencia, Spain. Occup. Environ. Med. 55: 541-547.
Tenias, J. M.; Ballester, F.; Perez-Hoyos, S.; Rivera, M. L. (2002) Air pollution and hospital emergency room
admissions for chronic obstructive pulmonary disease in Valencia, Spain. Arch. Environ. Health 57: 41-47.
Thompson, A. J.; Shields, M. D.; Patterson, C. C. (2001) Acute asthma exacerbations and air pollutants in children
living in Belfast, Northern Ireland. Arch. Environ. Health 56: 234-241.
7-193
-------
Thurston, G. D.; Ito, K. (2001) Epidemiological studies of acute ozone exposures and mortality. J. Exposure Anal.
Environ. Epidemiol. 11: 286-294.
Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers.
J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
Thurston, G. D.; Ito, K.; Hayes, C. G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital admissions and
summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ. Res.
65: 271-290.
Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. (1997) Summertime haze air pollution and children with
asthma. Am. J. Respir. Crit. Care Med. 155: 654-660.
Tobias, A.; Campbell, M. J.; Saez, M. (1999) Modelling asthma epidemics on the relationship between air pollution
and asthma emergency visits in Barcelona, Spain. Eur. J. Epidemiol. 15: 799-803.
Tolbert, P. E.; Mulholland, J. A.; Macintosh, D. L.; Xu, F.; Daniels, D.; Devine, O. J.; Carlin, B. P.; Klein, M.;
Dorley, J.; Butler, A. J.; Nordenberg, D. F.; Frumkin, H.; Ryan, P. B.; White, M. C. (2000) Air quality and
pediatric emergency room visits for asthma in Atlanta, Georgia. Am. J. Epidemiol. 151: 798-810.
Touloumi, G.; Katsouyanni, K.; Zmirou, D.; Schwartz, J.; Spix, C.; Ponce de Leon, A.; Tobias, A.; Quennel, P.;
Rabczenko, D.; Bacharova, L.; Bisanti, L.; Vonk, J. M.; Ponka, A. (1997) Short-term effects of ambient
oxidant exposure on mortality: a combined analysis within the APHEA project. Am. J. Epidemiol.
146: 177-185.
Tsai, S.-S.; Goggins, W. B.; Chiu, H.-F.; Yang, C.-Y. (2003a) Evidence for an association between air pollution and
daily stroke admissions inKaohsiung, Taiwan. Stroke 34: 2612-2616.
Tsai, S.-S.; Huang, C.-H.; Goggins, W. B.; Wu, T.-N.; Yang, C.-Y. (2003b) Relationship between air pollution and
daily mortality in a tropical city: Kaohsiung, Taiwan. J. Toxicol. Environ. Health Part A 66: 1341-1349.
Tyler, W. S.; Tyler, N. K.; Last, J. A.; Gillespie, M. J.; Barstow, T. J. (1988) Comparison of daily and seasonal
exposures of young monkeys to ozone. Toxicology 50: 131-144.
U.S. Environmental Protection Agency. (1996a) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
www.epa.gov/ncea/ozone.htm.
U.S. Environmental Protection Agency. (1996b) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
U.S. Environmental Protection Agency. (2004) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
http://cfpub.epa.gov/ncea/ [9 November, 2004].
Ulmer, C.; Kopp, M.; Ihorst, G.; Frischer, T.; Forster, J.; Kuehr, J. (1997) Effects of ambient ozone exposures during
the spring and summer of 1994 on pulmonary function of schoolchildren. Pediatr. Pulmonol. 23: 344-353.
Vaughan, T. R.; Weber, R. W.; Tipton, W. R.; Nelson, H. S. (1989) Comparison of PEFR and FEVj in patients with
varying degrees of airway obstruction: effect of modest altitude. Chest 95: 558-562.
Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
pollution. Environ. HealthPerspect. Ill: 45-51.
Vedal, S.; Rich, K.; Brauer, M.; White, R.; Petkau, J. (2004) Air pollution and cardiac arrhythmias in patients with
implantable cardiovascular defibrillators. Inhalation Toxicol. 16: 353-362.
Verhoeff, A. P.; Hoek, G.; Schwartz, J.; Van Wijnen, J. H. (1996) Air pollution and daily mortality in Amsterdam.
Epidemiology 7: 225-230.
Virnig, B. A.; McBean, M. (2001) Administrative data for public health surveillance and planning. Annu. Rev.
Public Health 22: 213-230.
Wang, X.; Ding, H.; Ryan, L.; Xu, X. (1997) Association between air pollution and low birth weight: a
community-based study. Environ. Health Perspect. 105: 514-520.
Ward, D. J.; Roberts, K. T.; Jones, N.; Harrison, R. M.; Ayres, J. G.; Hussain, S.; Walters, S. (2002) Effects of daily
variation in outdoor particulates and ambient acid species in normal and asthmatic children. Thorax
57: 489-502.
Weisel, C. P.; Cody, R. P.; Lioy, P. J. (1995) Relationship between summertime ambient ozone levels and
emergency department visits for asthma in central New Jersey. Environ. Health Perspect.
103(suppl. 2): 97-102.
7-194
-------
Weisel, C. P.; Cody, R. P.; Georgopoulos, P. G.; Purushothaman, V.; Weiss, S. H.; Bielory, L.; Gregory, P.; Stern,
A. H. (2002) Concepts in developing health-based indicators for ozone. Int. Arch. Occup. Environ. Health
75: 415-422.
White, M. C.; Etzel, R. A.; Wilcox, W. D.; Lloyd, C. (1994) Exacerbations of childhood asthma and ozone pollution
in Atlanta. Environ. Res. 65: 56-68.
Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. (199la) Activity
patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
ARB/R93/487. Available from: NTIS, Springfield, VA.; PB94-108719.
Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b) Study of children's activity
patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
Wilson, A. M.; Wake, C. P.; Kelly, T.; Salloway, J. C. (2005) Air pollution, weather, and respiratory emergency
room visits in two northern New England cities: an ecological time-series study. Environ. Res. 97: 312-321.
Wong, T. W.; Lau, T. S.; Yu, T. S.; Neller, A.; Wong, S. L.; Tarn, W.; Pang, S. W. (1999a) Air pollution and
hospital admissions for respiratory and cardiovascular diseases in Hong Kong. Occup. Environ. Med.
56: 679-683.
Wong, C.-M; Ma, S.; Hedley, A. J. Lam, T.-H. (1999b) Does ozone have any effect on daily hospital admissions for
circulatory diseases? J. Epidemiol. Community Health 53: 580-581.
Wong, C.-M.; Atkinson, R. W.; Anderson, H. R.; Hedley, A. J.; Ma, S.; Chau, P. Y.-K.; Lam, T.-H. (2002) A tale of
two cities: effects of air pollution on hospital admissions in Hong Kong and London compared. Environ.
Health Perspect. 110: 67-77.
World Health Organization. (2004) Meta-analysis of time-series studies and panel studies of paniculate matter (PM)
and ozone (O3): report of a WHO task group. Copenhagen, Denmark: WHO Regional Office for Europe;
document no. EUR/04/5042688. Available: http://www.euro.who.int/document/E82792.pdf [18 November,
2004].
Woodruff, T. J.; Grille, J.; Schoendorf, K. C. (1997) The relationship between selected causes of postneonatal infant
mortality and paniculate air pollution in the United States. Environ. Health Perspect. 105: 608-612.
Xu, X.; Ding, H.; Wang, X. (1995) Acute effects of total suspended particles and sulfur dioxides on preterm
delivery: a community-based cohort study. Arch. Environ. Health 50: 407-415.
Yang, Q.; Chen, Y.; Shi, Y.; Burnett, R. T.; McGrail, K. M.; Krewski, D. (2003) Association between ozone and
respiratory admissions among children and the elderly in Vancouver, Canada. Inhalation Toxicol.
15: 1297-1308.
Yang, C.-Y.; Chen, Y.-S.; Yang, C.-H.; Ho, S.-C. (2004a) Relationship between ambient air pollution and hospital
admissions for cardiovascular diseases inKaohsiung, Taiwan. J. Toxicol. Environ. Health Part A 67: 483-493.
Yang, C.-Y.; Chang, C.-C.; Chuang, H.-Y.; Tsai, S.-S.; Wu, T.-N.; Ho, C.-K. (2004b) Relationship between air
pollution and daily mortality in a subtropical city: Taipei, Taiwan. Environ. Int. 30: 519-523.
Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A. (2000) Exposure
measurement error in time-series studies of air pollution: concepts and consequences. Environ. Health
Perspect. 108: 419-426.
Zhu, L.; Carlin, B. P.; Gelfand, A. E. (2003) Hierarchical regression with misaligned spatial data: relating ambient
ozone and pediatric asthma ER visits in Atlanta. Environmetrics 14: 537-557.
Zidek, J. V. (1997) Interpolating air pollution for health impact assessment. In: Barnett, E. V.; Turkman, K. F., eds.
Pollution Assessment and Control. New York, NY: John Wiley & Sons. (Statistics for the Environment, no.
3).
Zidek, J. V.; Wong, H.; Le, N. D.; Burnett, R. (1996) Causality, measurement error and multicollinearity in
epidemiology. Environmetrics 7: 441-451.
Zidek, J. V.; White, R.; Le, N. D.; Sun, W.; Burnett, R. T. (1998) Imputing unmeasured explanatory variables in
environmental epidemiology with application to health impact analysis of air pollution. Environ. Ecol. Stat.
5:99-115.
Zmirou, D.; Barumandzadeh, T.; Balducci, F.; Ritter, P.; Laham, G.; Ghilardi, J.-P. (1996) Short term effects of air
pollution on mortality in the city of Lyon, France, 1985-90. In: St Leger, S., ed. The APHEA project. Short
term effects of air pollution on health: a European approach using epidemiological time series data. J.
Epidemiol. Community Health 50(suppl. 1): S30-S35.
Zmirou, D.; Schwartz, J.; Saez, M.; Zanobetti, A.; Wojtyniak, B.; Touloumi, G.; Spix, C.; Ponce de Leon, A.;
Le Moullec, Y.; Bacharova, L.; Schouten, J.; Ponka, A.; Katsouyanni, K. (1998) Time-series analysis of air
pollution and cause-specific mortality. Epidemiology 9: 495-503.
7-195
-------
8. INTEGRATIVE SYNTHESIS: OZONE EXPOSURE
AND HEALTH EFFECTS
8.1 INTRODUCTION
This integrative synthesis is structured to provide a coherent framework for the assessment
of health risks associated with human exposures to ambient surface-level (tropospheric) ozone
(O3) in the United States. The main goal of the chapter is to integrate newly available scientific
information with key findings and conclusions from the 1996 O3 AQCD (U.S. Environmental
Protection Agency, 1996a), so as to address issues central to the EPA's assessment of evidence
needed to support the current review of the primary O3 NAAQS. The integrated assessment of
key findings and conclusions provided here and elsewhere in this document with regard to O3
exposure and health effects will be drawn upon and their policy implications considered in an
Ozone Staff Paper prepared by EPA's Office of Air Quality Planning and Standards (OAQPS).
The analyses provided in that Staff Paper aim to "bridge the gap" between scientific assessments
in this criteria document and judgments required of the EPA administrator in evaluating whether
to retain or, possibly, to revise the current primary O3 NAAQS. Other types of scientific
information concerning ambient O3 welfare effects (i.e., tropospheric O3 effects on vegetation
and ecosystems, relationships to surface-level solar UV flux/climate changes, and effects on
man-made materials) are assessed in ensuing Chapters 9, 10, and 11. That information will also
be considered in the OAQPS staff paper in posing options regarding the secondary O3 NAAQS.
As discussed in Chapter 2 of this document, O3 found in the earth's troposphere generally
originates from photochemical reactions that involve the interaction of sunlight with precursor
pollutants, especially nitrogen oxides (NOX), carbon monoxide (CO), and volatile organic
compounds (VOCs) such as hydrocarbons emitted by surface-level mobile and stationary
sources and natural sources. Other photochemical oxidants, such as peroxyacetyl nitrate (PAN)
and hydrogen peroxide (H2O2), are also generated along with O3 by such atmospheric
interactions. In addition to the tropospheric O3 generated by these interactions, some O3 is found
near the earth's surface as the result of its downward transport from the stratosphere. However,
in contrast to stratospheric O3, which plays an important role in maintaining the habitability of
the planet by shielding the Earth's surface from harmful solar ultraviolet (UV) radiation,
8-1
-------
tropospheric O3 at the surface can also exert adverse effects on humans, nonhuman animal
species, and vegetation. As was the case for previous O3-related NAAQS criteria revisions, the
present criteria document focuses mainly on the assessment of health and welfare effects
resulting from exposures to surface-level concentrations of tropospheric O3, whereas less
attention is accorded to the distinctly much more limited available information on other
photochemical oxidants, e.g., PAN or H2O2.
Based on the criteria review completed in 1978, the original primary and secondary
NAAQS set in 1971 for total photochemical oxidants were revised in 1979 to focus on O3 as the
indicator for new primary and secondary standards that were attained when the expected number
of days per calender year with daily maximum 1-h average O3 concentrations (1-h max O3)
>0.12 ppm did not exceed one. The NAAQS for ambient O3 were revised in 1997 by replacing
the 1-h standards with an 8-h primary standard that is met when the 3-year average of the annual
fourth highest daily maximum 8-h average concentration (8-h max O3) is <0.08 ppm. The new
1997 primary standard was based on various scientific supportive data from experimental human
exposure studies, animal toxicologic studies, and epidemiological studies, as assessed in the
1996 O3 AQCD and in the 1996 O3 Staff Paper (U.S. Environmental Protection Agency, 1996b).
8.1.1 Chapter Organization
In addition to providing the above brief background information regarding prior O3
NAAQS reviews, this first section (8.1 Introduction) of the integrative synthesis chapter aims to
orient the reader to the organization and content of the chapter. The next section (Section 8.2)
focuses on air quality trends and current ambient O3 levels to help provide context for the
ensuing discussions of O3 exposures and associated health effects. Subsequent sections
then integrate newly available key scientific information assessed in Chapters 4 through 7 of
this document, including integration of information on O3 dosimetry, toxicological information
derived from controlled human exposure and laboratory animal studies, and epidemiologic
evidence.
These sections collectively address the following topics: (1) ambient O3 exposures,
personal exposures, and dosimetric considerations; (2) experimental studies on toxicological
responses to acute O3 exposures in humans (clinical studies) and both acute and chronic effects
in animals; (3) epidemiological evidence for associations between ambient O3 exposure of
8-2
-------
human populations and various health effects, as well as the strength and robustness of these
associations; (4) integration of the experimental and epidemiological evidence; (5) biological
mechanisms and other evidence useful in judging the plausibility of adverse health effects being
associated with human exposures to ambient O3 levels encountered in the United States; and
(6) delineation of susceptible and vulnerable populations likely at increased risk for O3-related
health effects and numbers of people potentially falling in such categories in the United States.
8.2 AMBIENT OZONE AIR QUALITY IN THE UNITED STATES
8.2.1 Current Ozone Concentrations and Spatial Patterns
The EPA has established "O3 seasons," during which ambient O3 concentrations must be
monitored in the United States and its territories. These seasons vary in length depending on
location. The O3 season extends all year in the Southwest. In most other areas of the country, O3
is monitored typically from April to October. However, O3 is monitored throughout the year in
many urban areas, because O3 is present year round not only in polluted areas but in clean areas
as well. The median of the daily 8-h max O3 in the United States, averaged over May to
September from 2000 to 2004 for all U.S. counties, was 0.049 ppm. In 95% of all counties, the
median of the daily 8-h max O3 was less than 0.057 ppm. However, it should be noted that most
monitors are located in the East and O3 data are sparse throughout large areas of the West.
Median values of daily 1-h max O3 were typically much higher in large urban areas or in areas
downwind of them. For example, in Houston, TX they approached 0.20 ppm during the same
2000-2004 period. Daily 1-h max O3 concentrations were lower in the rest of the country, but
were still above 0.12 ppm in many locations. Eight-hour daily maximum concentrations were
not as high, but tend to be highly correlated with 1-h daily maximums.
Within individual Metropolitan Statistical Areas (MSAs), O3 concentrations tend to be well
correlated across monitoring sites, although spatial variations in concentrations can be
substantial. In many city centers, O3 concentrations tend to be lower than in either upwind or
downwind areas, largely due to reaction of O3 with NO emitted by motor vehicles. For example,
much lower O3 concentrations overall are found in downtown Los Angeles (e.g., in Lynwood)
than at sites located further downwind (e.g., in San Bernadino). The much higher downwind
levels are formed from photochemical reactions involving the urban emissions, including
-------
products formed as the result of reactions titrating O3 in the urban core. Thus, O3 concentrations
tend to be higher downwind of urban centers, and they decrease again in areas that are more
remote from precursor sources. Likewise, surface-level O3 can be depleted in rural areas close to
NO emission sources, such as highways and powerplants.
8.2.2 Diurnal and Seasonal Variations
Ozone concentrations typically tend to peak in early to mid-afternoon in areas where there
is strong photochemical activity and to peak later in the afternoon or during early evening in
areas where transport is more important in determining the O3 abundance. Summertime maxima
in O3 concentrations occur in those U.S. areas where substantial photochemical activity acts
on O3 precursors emitted as the result of human activities. Monthly maxima can occur anytime
from June through August. However, springtime maxima are observed in some National Parks,
mainly in the western United States, and at a number of other relatively unpolluted monitoring
sites throughout the Northern Hemisphere. For example, the highest O3 concentrations at
Yellowstone National Park tend to occur during April and May. Typically, monthly minima
tend to occur from November through February at polluted sites and during the fall at relatively
remote sites.
8.2.3 Long-Term Trends
National attention started to be focused in the 1940s on O3 and associated photochemical
smog in the Los Angeles area. Prior to the adoption of stringent emissions controls to reduce O3
precursors, peak O3 levels were consistently higher in the Los Angeles area than are currently
observed. For example, in 1958, peak O3 concentrations measured in Los Angeles were about
0.6 ppm but have declined since then, although not at a steady rate. Peak O3 levels of 0.2 to
0.5 ppm were still found at some locations in the Los Angeles basin during the 1970s.
For example, on two days (October 13 and 14) during a 1978 episode, Tuazon et al. (1981)
observed peak 1-h averaged O3 values of nearly 0.4 ppm and nearly 0.5 ppm. Currently, peak
1-h and 8-h average O3 concentrations are about 0.17 and 0.15 ppm in the Los Angeles basin
(cf, Figures 3-10 and 3-11). High O3 levels were also earlier found throughout the rest of the
United States as well, but peak O3 levels have also gradually declined across the country during
the 1980s and 1990s. However, during one particularly hot summer (of 1988) in the East, peak
8-4
-------
1-h O3 concentrations of about 0.2 ppm were observed in many eastern U.S. cities (U.S.
Environmental Protection Agency, 1990).
Nationwide, 2nd highest 1-h ozone concentrations in the United States have decreased
dramatically during the past several decades, i.e., by -29 percent from 1980 to 2003 and
16 percent from 1990 to 2003. Also, 4th highest 8-h O3 concentrations decreased by -21 percent
since 1980 and 9 percent since 1990 (U.S. Environmental Protection Agency, 2003). Trends in
metrics for evaluating compliance with the O3 NAAQS (i.e., changes in the 4th highest O3
concentration) can be found in EPA's "National Air Quality and Emissions Trends Reports."
These reports indicate that the 4th highest O3 concentrations are still decreasing nationwide, but
the rate of decrease has slowed since 1990. However, such trends have not been uniform across
the United States. In general, reductions in the O3 metrics given above have been largest in New
England and in states along the West Coast and smallest in midwestern states. Downward trends
in California O3 concentrations have been driven mainly by notable decreases in Southern
California, with smaller reductions occurring in other areas. Trends in peak O3 metrics do not
necessarily reflect changes in O3 concentrations across the middle of the distribution of ambient
O3 values. Of note, O3 concentrations towards the center of its nationwide distribution have not
shown much change (cf, Figures 3-17, 3-18, and Table AX3-9), and there are some indications
that O3 concentrations at the lower end of the distribution may even be increasing.
8.2.4 Interrelationships Between Ozone and Other Ambient Pollutants
Data on ambient concentrations of other oxidants (e.g., H2O2, PAN) and oxidation products
(e.g., HNO3, H2SO4) in the atmosphere are not nearly as abundant as they are for O3. Because
data for such species are usually obtained only as part of specialized field studies, it is difficult to
relate observed ambient O3 concentrations to ambient levels of other oxidant species or oxidation
products. In general, such secondary species are expected to be at least moderately positively
correlated with O3, whereas, primary species are expected to be more highly correlated with each
other than with secondary species (provided that the primary species originate from common
sources in given areas). Measurements of gas phase oxidants obtained as part of the Southern
Oxidants Study (SOS) found combined hydroperoxide (H2O2, CH3OOH, and HOCH2OOH)
concentrations typically in the range of several ppb. Other SOS measurements likewise
indicated combined concentrations of PAN, PPN, and MPAN in the range of several ppb.
3-5
-------
Oxidants are also present in airborne cloud droplets, rain drops, and particulate matter (PM).
A few measurements of reactive oxygen species (ROS), expressed as equivalent H2O2,
in ambient fine PM have indicated levels of less than 1% of those for ambient O3 on a molar
basis. However, it should be noted that these measurements are potentially subject to both
positive and negative artifacts.
Relationships between ambient air O3 and PM levels can be quite complex because PM
is not a single distinct chemical species, but rather a mix of primary and secondary species.
As an example of this complexity, O3 concentrations were found to be positively correlated with
PM2 5 during the summer, but negatively correlated during the winter at Ft. Meade, MD (cf,
Figure 3-21). Also, Ito et al. (2005) examined relationships between ambient O3 and PM10 on a
seasonal basis in several U.S. urban areas (cf, Figure 7-24). Seasonal O3-PM relationships
similar to those observed at Ft. Meade were found, reflecting the dominant contribution of PM2 5
to PM10 in the urban areas studied. Possibly contributing to higher correlations seen between O3
and fine PM in the summer is the fact that O3 can contribute to formation of submicron particles
via interactions with various other atmospheric constituents present, such as terpenes and other
biogenically derived hydrocarbons from trees, other vegetation, and wood products. Formation
of ultrafine particles by this mechanism mostly occurs during summer days when temperatures
and O3 concentrations are sufficiently elevated to facilitate O3 reactions with increased amounts
of biogenic hydrocarbons emitted from vegetation. Bursts of ultrafine particle formation have
been observed repeatedly in both urban and rural air. Woo et al. (2001), for example, reported
rapid formation of ultrafine particles in the ambient air of Atlanta typically around noon in both
summer and winter. The mechanisms underlying such ultrafine particle formation events may
also involve other atmospheric reactions related to O3 formation, such as nucleation of H2SO4
(produced by oxidation of SO2) and, probably, NH3.
8.2.5 Policy Relevant Background (PRB) Ozone Concentrations
Policy relevant background (PRB) O3 concentrations, i.e., background O3 concentrations
used for NAAQS-setting purposes, are those that would occur in the United States in the absence
of anthropogenic emissions in continental North America (defined here as the United States,
Canada, and Mexico). Such PRB O3 concentrations include contributions from natural sources
everywhere in the world and from anthropogenic sources outside these three countries. For the
8-6
-------
purpose of informing O3 NAAQS decisions, EPA focuses on assessing risks to human health and
environmental effects from O3 levels in excess of PRB concentrations. Issues concerning the
methodology for estimating PRB O3 concentrations are discussed in detail in Annex AX3,
Section AX3.9.
Contributions to PRB O3 include photochemical reactions involving natural emissions of
VOCs, NOX, and CO, as well as the long-range transport of O3 and its precursors from outside
North America and the stratospheric-tropospheric exchange (STE) of O3 (see Annex AX2,
Section AX2.3 for details regarding O3-related STE processes). Natural sources of O3 precursors
include biogenic emissions, wildfires, and lightning. Biogenic emissions from agricultural
activities are not considered in the formation of PRB O3. However, estimates of PRB O3
concentrations cannot be derived solely from measurements of O3 at relatively unpolluted sites
because of long-range transport from anthropogenic source regions within North America. It is
impossible to determine sources of O3 at a particular location without ancillary data that could be
used as tracers of sources or to calculate photochemical production and loss rates for O3. Thus,
estimates of PRB O3 concentrations are currently based on predictions generated by the global
scale, three dimensional, chemical transport model GEOS-CHEM (Fiore et al., 2003).
Policy relevant background O3 concentrations vary as a function of season, altitude, and
total surface O3 concentration, with PRB O3 concentrations at the surface generally falling in the
range of 0.015 to 0.035 ppm from 1300 to 1700 local time and tending to decline under weather
conditions conducive to O3 episodes. The PRB concentrations are highest during spring and
decline into summer; and higher values also tend to occur at higher elevations during the spring
due to contributions from hemispheric pollution and stratospheric intrusions. The contribution to
surface O3 by stratospheric intrusions is typically well below 0.020 ppm. Stohl (2001) and
Sprenger et al. (2003) found that the maximum probability of stratospheric intrusions reaching
the 800 hPa level (-1800 m) was less than 1% and that higher probabilities (1 to 2%, and 10%)
applied for stratospheric intrusions penetrating to the 600 hPa level (-4100 m) and 500 hPa level
(-5400 m), respectively. Thus, stratospheric O3 intrusions only rarely contribute to elevated
surface-level O3 concentrations at low altitude sites but have a higher (albeit still low)
probability of elevating them at high-altitude sites.
3-7
-------
8.3 FACTORS AFFECTING HUMAN EXPOSURE TO AMBIENT OZONE
Human exposure to O3 and related photochemical oxidants varies over time due to changes
in their ambient concentrations and because people move between locations having notably
different O3 concentrations. Also, the amount of O3 delivered to the lung is not only influenced
by the ambient concentration but also by the individual's breathing route and rate. Thus, activity
level is an important consideration in determining potential O3 exposure and dose received.
The use of data from ambient air monitoring stations is still the most common surrogate for
assigning exposure estimates in epidemiologic studies. Since the primary source of O3 exposure
is the ambient air, O3 concentration data from outdoor community monitoring sites should
provide a relative assignment of exposure with time, if O3 concentrations are relatively uniform
across the region; time-activity patterns are roughly the same across the study population; and
housing characteristics (such as ventilation rates and O3 sinks contributing to indoor O3 decay
rates) are relatively constant for the study area. However, because these types of factors often do
vary across populations and locations, some error tends to be associated with estimates of the
magnitude of O3 exposure of large populations. Nevertheless, although substantial variability
may exist among personal measurements, human exposure studies have observed that daily
average personal O3 exposures for the general population tend to be reasonably well correlated
with monitored ambient O3 concentrations. Therefore, ambient O3 monitoring data appear to
provide the most useful index of human O3 exposure currently available to help characterize
health outcomes associated with O3 exposures of large population groups.
8.3.1 Personal Exposure
Personal O3 concentrations have been measured for children, outdoor workers, and
individuals with pulmonary diseases (populations potentially vulnerable to increased ambient O3
exposure and/or susceptible to O3 or other respiratory irritants). Outdoor workers can be
expected to have somewhat higher O3 exposures than other individuals, because they typically
spend more time outdoors and often engage in prolonged moderate and heavy exertion activities.
Children also tend to be more active outside and, therefore, often manifest a higher breathing
rate than most adults. Given the higher percentages of time spent outdoors than most other
population groups, personal exposures of both outdoor workers and children tend to be more
highly correlated with ambient O3 concentrations measured at community monitoring sites.
-------
8.3.2 Indoor Concentrations
Apart from only a few specific indoor sources such as photo-copying machines, O3 indoors
is derived from the infiltration of ambient air from outdoors. Generally, O3 enters indoor
environments through infiltration from the outdoors and through building components, such as
windows, doors, and ventilation systems. Ozone concentrations in indoor environments depend
primarily on the outdoor O3 concentration, outdoor/indoor infiltration and the air exchange rate
(AER). Hence, indoor O3 concentrations tend to reflect outdoor concentrations and are higher
when outdoor O3 is higher.
Once indoors, O3 reacts on various surfaces and with airborne components of either indoor
or outdoor origin. Because O3 reacts indoors with surfaces and other contaminants, O3
concentrations are typically lower indoors than outdoors. Gas phase reactions occurring indoors
also produce other oxidants analogous to the production of photochemical smog. The extent and
rate of production of these other species indoors is a function of indoor O3 concentrations and the
presence of other necessary precursors (i.e., VOCs), along with an optimal AER.
Several studies of O3 measured in residences, schools, office buildings and museums found
that typical O3 concentrations varied across all such locations. Indoor O3 levels generally varied
in relationship to the AER in the indoor environment (increasing with higher AER) and were
generally notably lower than outdoor ambient O3 levels. For example, one study of O3 levels
indoors and outside of a school in New England found average O3 concentrations of 40 ppb
(0.040 ppm) outdoors and 20 ppb (0.020 ppm) indoors. With regard to O3 levels in mobile
source microenvironments, as is the case for other enclosed environments, O3 concentrations
depend on the extent of mixing of outdoor air into the vehicle cabin. Thus, if windows are kept
open, O3 concentrations inside the vehicle may approach outdoor values. But, if windows are
kept closed and there is air conditioning, then interior values can be much lower than those
outside, especially if recirculated air is used. For example, in one N.C. study of police cars with
air conditioning and recirculated air, vehicle cabin O3 levels (11.7 ppb average) were less than
half those outside (28.3 ppb average at area outdoor monitoring sites).
Although O3 concentrations may be reduced to lower levels once ambient O3 enters
indoor environments, the indoor O3 may interact with other airborne substances of indoor or
outdoor origin present indoors. For example, Wainman et al. (2000) found that O3 reacts with
d-limonene, a common component of air fresheners to produce submicron particles found mainly
8-9
-------
to range in size from 0.1 to 0.3 jim. Wainman et al. noted that terpenes such as limonene are
emitted by wood products; that they are used as solvents, as odorants in cleaning products, and
as air fresheners; and, even though they are produced by vegetation outdoors, their indoor
concentrations are often higher than outdoors because of their widespread uses. In addition to
particle formation, Weschler (2004) noted that gas phase products (i.e., aldehydes and
hydroperoxides) produced by O3 reactions with terpenes and other unsaturated carbon
compounds may also be of concern. During formation of such products, OH radicals are also
generated that can react with compounds that do not react with O3. To the extent that building
ventilation rates remain constant between days characterized by high and low ambient O3, the
concentrations of these other secondary pollutants formed indoors will tend to be correlated with
ambient O3. Thus, ambient O3 concentrations measured outdoors at community monitoring sites
and/or personal O3 exposure monitor measurements may serve not only as indices of direct
human exposure to O3 per se, but also as surrogate indices of exposures to broader O3-containing
ambient mixtures of photochemical oxidants and/or other pollutants.
8.4 SYNTHESIS OF AVAILABLE INFORMATION ON OZONE-
RELATED HEALTH EFFECTS
The integrated synthesis of the latest available information on O3-related health effects
poses an interesting challenge in view of the emergence of highly important new information
since the 1996 O3 AQCD. Such information includes new findings from:
• Dosimetry studies that further characterize factors potentially affecting regional
distribution of O3 in the respiratory tract of humans and laboratory animals and provide
improved bases for animal-to-human extrapolations of experimentally-observed
O3-induced health effects.
• Experimental toxicological studies using controlled human exposures and laboratory
animals aimed at delineating exposure-response relationships and understanding potential
biochemical mechanisms underlying toxic effects, pathology, and susceptibility.
• Epidemiological studies, reflecting progress in addressing many research needs identified
during the last review, as well as raising new issues and reevaluating previously addressed
issues that remain important in interpreting the body of epidemiological evidence and
characterization of its strengths and limitations.
8-10
-------
Previous criteria assessments, including the 1996 O3 AQCD, found that experimental
studies of controlled human and laboratory animal exposures to O3 provided the most clear cut
and compelling evidence with regard to characterizing O3-related health effects. Based on
extensive dosimetric and experimental data, as well as growing epidemiologic evidence
available at the time, the 1996 O3 AQCD arrived at a set of findings and conclusions regarding
potential health effects of ambient O3 exposure. In general, the existing evidence was such to
warrant a high degree of confidence in those conclusions derived from experimental (controlled
exposure) studies of acute O3 exposure effects. Considerable confidence could also be placed in
the emerging field/panel studies providing observational study results substantiating and
extending the controlled exposure study findings. The controlled exposure and field/panel
studies clearly demonstrated respiratory effects (e.g., reduced pulmonary function) among
healthy and asthmatic children and adults acutely exposed for 1 to 8 h to 0.08 ppm O3 while
physically active. Other epidemiologic studies provided highly suggestive, although less
conclusive, indications of increased morbidity (e.g., as indexed by emergency department [ED]
visits, hospital admissions, etc.) and, possibly, mortality being associated with acute exposure of
human populations to ambient O3. Also, data from laboratory animal and controlled human
exposure studies supported the hypothesis that coexposure to O3 and other pollutants at low-
effect levels can result in enhanced effects. However, the issue of exposure to copollutants was
poorly understood, especially with regard to potential chronic effects.
Since the 1996 O3 AQCD evaluations, further controlled human exposure studies have
extended earlier findings of respiratory effects of acute exposures in exercising adults to O3
concentrations ranging below 0.08 ppm for some sensitive subjects. Also, a more extensive
database of air pollution epidemiologic studies has become available; and a subset of these new
observational studies have reported a variety of O3-related health effects associations, with
newly reported evidence of ambient O3 mortality relationships being of special interest. Based
on physiological, biochemical, and molecular changes observed in controlled human exposure
studies and animal toxicological studies, new evidence is now available by which to more fully
evaluate the biological plausibility and extent of coherence for various health outcomes (such as
respiratory and cardiovascular effects and mortality) reported in epidemiologic studies.
8-11
-------
8.4.1 Integration of Experimental and Epidemiologic Evidence
8.4.1.1 Cross-Cutting Issues Relevant to Assessment/Interpretation of Ozone
Health Effects
Three chapters in the current document provide detailed discussion of various experimental
approaches utilized to evaluate O3-related health effects. Chapter 4 discusses dosimetry issues
pertinent to both animal and human exposure scenarios. Chapter 5 discusses the experimental
studies of biochemical (cellular and molecular changes), physiological, and pathological
observations in laboratory animals (including nonhuman primates, dogs, and rodent species)
and in vitro studies using cell culture systems, in certain cases, on humans cells recovered from
bronchoalveolar lavage fluid (BALF) after exposure to O3. Chapter 6 evaluates studies on
human volunteers exposed to O3 which have investigated a variety of physiological and
biochemical endpoints. In interpreting the results from the experimental approaches, one must
consider the following issues: (1) exposure/dose considerations; (2) interpretation of results for
animals and humans from relatively high dose O3 exposures; (3) animal-to-human
extrapolations, and (4) experimental results derived from single pollutant exposures in
comparison to much more complex ambient exposures.
Earlier animal toxicology studies were carried out using relatively high O3 exposure
concentrations/doses that do not necessarily reflect "real-world" exposure scenarios. Those
experiments were primarily aimed at understanding the pathophysiology associated with O3
exposure in healthy animals, to help identify potential mechanisms(s) of action and to help
validate health outcomes reported in epidemiologic studies. Since the 1996 O3 AQCD, some of
the human and animal studies have used ambient and/or near ambient doses. The majority of
controlled chamber exposure studies on human volunteers mainly limited exposures to O3 alone
in comparison to sham (clean air) exposures, thus providing evidence concerning direct effects
of O3 per se versus more closely mimicking real-world atmospheric exposures to multipollutant
mixes. Some limited data from laboratory animal and controlled human exposure studies
reviewed in the 1996 O3 AQCD pointed towards enhancement of some observed health effects
by coexposures to O3 and other air pollutants. Since then a few newer human clinical and
laboratory animal air pollution studies have utilized various coexposure regimens to simulate
more closely ambient exposure to air pollution mixtures; and the results from these studies are
highly useful in developing better models to interpret toxicologic effects associated with
O3-containing ambient air pollutant mixes.
8-12
-------
Interpretations of experimental studies of air pollution effects in animals, as in the case of
environmental comparative toxicology studies, are affected by limitations associated with
extrapolation models. The differences between humans and rodents with regard to O3 absorption
and distribution profiles based on breathing pattern, exposure dose, and differences in lung
structure and anatomy (see Chapter 4 and 5 for details) all have to be taken into consideration.
Also, in spite of a high degree of homology and the existence of a high percentage of
orthologous genes across human and rodents (particularly mice), extrapolation of molecular
alterations at the gene level is complicated by species-specific differences in transcriptional
regulation. Given these molecular differences, there are large uncertainties associated with
quantitative extrapolations at this time between laboratory animals and humans of observed O3-
induced pathophysiological alterations under the control of widely varying biochemical,
endocrine, and neuronal factors.
Epidemiologic studies provide important additional observational information to our
knowledge base regarding O3-induced health effects. Of particular importance, such studies help
(1) to ascertain the extent to which the types of health effects found in experimental studies
occur in response to ambient exposures of human populations to O3 and/or other copollutants;
and (2) to delineate exposure-response relationships for more serious health effect outcomes
(e.g., hospital admissions, mortality) that help to elucidate more fully the likely public health
consequences of ambient O3 exposures.
The ensuing sections both briefly summarize key dosimetry and health related findings
derived from the 1996 O3 AQCD and integrate those findings with new information obtained
since 1996 from (a) human and animal experimental studies and (b) epidemiologic studies.
Physiological, pathological, cellular and biochemical alterations induced by experimental O3
exposures are evaluated and compared to human health effects associated with ambient O3
exposures. Also, the influence of O3-induced changes at cellular and molecular levels are
integrated to help elucidate mechanistic bases for the observed physiological and pathological
alterations. These research results are evaluated both to help assess the biological plausibility of
health outcome associations observed in epidemiologic studies and to assess the coherence of the
overall body of evidence relevant to O3-related health outcome conclusions, to aid in drawing
conclusions about the likelihood of causal relationships existing between ambient exposure to O3
and various types of reported O3-related health effects.
8-13
-------
8.4.1.2 Dosimetry
The state-of-the-art of O3 dosimetry, as described in 1996 O3 AQCD, indicated consistency
across data and models derived from in vivo human and animal studies, thus increasing the level
of confidence in the development of dosimetric extrapolation models. Earlier dosimetry models
predicted that the tissue dose of inhaled O3 was greatest at the bronchoalveolar junction, the lung
region experimentally shown to be most impacted by O3. Ozone bolus inhalation studies in
humans had indicated that inspired O3 reaches the distal airways and alveoli of resting humans;
and, with increased inspiratory flow rates due to exercise, O3 penetrates deeper and in greater
quantity to the distal regions of the lung. These findings were corroborated by observations of
18O3 (oxygen-18-labeled ozone) in the BALF of humans and rats (Hatch et al., 1994). Based on
BALF parameters monitored in this study, an exposure of 0.4 ppm O3 in exercising humans
appears to be approximately equivalent to an exposure of 2 ppm in resting rats.
Some O3-induced acute responses compared well across species when controlled for dose,
indicating that animals and humans (a) respond to O3 in a dose-dependent manner, i.e., they
exhibit increasing breathing frequency with an accompanying decrease in tidal volume
(tachypnea), and (b) show similar changes in alveolar permeability as measured by protein in the
BALF. These parallel changes in humans and animals were sufficiently homologous to suggest
a common mode of action. The majority of lung function decrements were seen to subside with
repeated exposures in both humans and animals, with analogous attenuation of certain (but not
all) parameters measured in the BALF. The mechanisms associated with attenuation are unclear
but may involve endogenous antioxidants. The significance of non-attenuated markers in BALF
has been interpreted to relate to potential chronicity of O3 effects.
During the past decade, no further reports have been published on O3 uptake studies in
animals, although several controlled human bolus and/or general O3 uptake studies have
provided refined data. The bolus uptake studies suggest that prior exposure to O3 diminishes
bolus uptake. In the 1996 O3 AQCD, the effect of mode of breathing (oral or nasal) on O3 uptake
was thought to be minimal, with approximately equal uptake via the nose or mouth. Newer
bolus dose studies have demonstrated that the uptake and regional respiratory tract distribution
of O3 is sensitive to mode of breathing (nasal uptake greater than oral) and to air flow rate
(uptake decreases with increasing flow). Studies (Rigas et al., 1997, Bush et al., 2001) that
evaluated the role of co-exposure with other oxidants such as NO2 and SO2 recognized the
8-14
-------
importance in mediating O3 toxicity of other oxidative environment components in the lung,
e.g., radicals generated by O3 interaction with the lipids in the airway and epithelial lining fluid
(ELF).
New uptake studies (Ultman et al., 2004) carried out in controlled human clinical studies
have observed gender-specific differences in the uptake of O3, but these differences do not
correlate well with spirometric responses. Rather, they appear to be related to breathing pattern
and lung size, due to females having smaller lungs than males. Thus, a number of variables
seem to affect O3 uptake, notably including route of breathing, breathing pattern, gender,
copollutants, and certain pre-exposure conditions. These differences are important in order to
interrelate experimentally-demonstrated pathophysiological effects and epidemiologically-
observed associations between ambient O3 concentrations and health risks among human
population groups.
8.4.2 Experimental Evidence for Ozone-Related Health Effects
Pulmonary Function
Numerous controlled human exposure, animal, and epidemiological studies assessed in the
1996 O3 AQCD demonstrated alterations in various pulmonary function measures. Ozone
inhalation for several hours (1 to 3 h) while physically active was shown to elicit both acute
pathophysiologic changes and subjective respiratory tract symptoms. The pulmonary responses
observed in healthy human subjects exposed to ambient O3 concentrations included decreased
inspiratory capacity; mild bronchoconstriction; rapid, shallow breathing pattern during exercise;
and subjective symptoms of tracheobronchial airway irritation, including cough and pain during
inspiration. Acute O3 exposures were also found to cause decreases in forced vital capacity
(FVC), forced expiratory volume in 1 s (FEVj), peak expiratory flow (PEF), and increased
airways resistance (SRaw). The severity of symptoms and the magnitude of pulmonary response
were reported to vary as a function of exposure concentration, duration, and level of exercise
(which help to determine inhaled O3 dose), as well as the individual sensitivity of exposed
subjects and the extent of tolerance resulting from previous exposures. With regard to the latter,
the 1996 O3 AQCD noted that during repeated short-term exposures, some of the O3-induced
responses (inflammation and lung function decrements) are partially or completely attenuated
with a differential attenuation profile. Over a 5-day exposure, both pulmonary and inflammatory
8-15
-------
markers were found to attenuate by the fifth day of exposure, but markers of cell damage
(discussed below under Lung Injury) do not attenuate and continue to increase.
Group mean data from numerous controlled human exposure (in healthy subjects 8 to
45 years of age, after 1 to 3 h exposure) and field studies indicated that, in general, statistically
significant pulmonary function decrements beyond the range of normal measurement variability
(e.g., 3 to 5% for FEVj) occur (a) at >0.50 ppm O3 when at rest; (b) at >0.37 ppm O3 with light
exercise (slow walking); (c) at >0.30 ppm O3 with moderate exercise (brisk walking); (d) at
>0.18 ppm O3 with heavy exercise (easy jogging) and (e) at >0.16 ppm O3with very heavy
exercise (running). Smaller group mean changes (e.g., <5%) in FEVj occurred at lower
O3 concentrations with very heavy exercise in healthy adults at 0.15 to 0.16 ppm O3 and in
healthy young adults at levels as low as 0.12 ppm. Also, pulmonary function decrements were
seen in children and adolescents at O3 concentrations of 0.12 and 0.14 ppm with heavy exercise.
Human studies reviewed in the 1996 O3 AQCD that used longer duration (6- to 8-h) acute
exposures with exercise (which better mimic multihour exposures to ambient O3 that typify more
prolonged elevated ambient O3 levels often observed in U.S. urban areas) provided some of the
strongest and most quantifiable concentration-response data on pulmonary function effects of
acute O3 exposure. Overall, the 1996 O3 AQCD found that for healthy subjects performing
moderate exercise during longer duration (6 to 8 h) acute O3 exposures, group mean 5%
decrements in FEVj were seen at relatively low O3 levels, such as: 0.08 ppm O3 after 5.6 h,
0.10 ppm O3 after 4.6 h, and 0.12 ppm O3 after 3 h. Some subjects in the earlier studies
experienced FEVj decrements in excess of 15% with short-term (1- to 3-h) or longer-duration
(6.6-h) acute O3 exposures, suggesting potential wide interindividual variability in pulmonary
function responses.
The few newly available controlled human exposure studies (since those assessed in the
1996 O3 AQCD) that used near-ambient O3 concentrations (<0.12 ppm) are summarized in
Appendix Table 8A-1; and these, like most controlled exposure studies to date, continue to
indicate that considerable interindividual differences exist in the magnitude of responses to O3.
However, a given individual's lung function and, to a lesser extent, respiratory symptom
responses to O3 are reproducible over a period of time, indicating that some individuals are
consistently more responsive than others to O3. Figure 8-1A illustrates well the variability in
FEVj responses in young healthy adults following a prolonged (6.6 h) exposure to O3, as
8-16
-------
A.
43 ou
u
•£. 40-
.Q
W 30-
| 20-
| 10-
<5
n Q
0 ppm
n n = 87
n
~
0%
n
.
-
.
.
Jl
~i
10 0 -10 -20 -30 -40 10 0
B.
vi 60
& 50-
3 40-
W
*- 30-
O
•£ 20-
0 10-
0)
n n-
n
0 ppm
n = 30
0%
,_,
.
-
I
0.08 ppm
n = 60
26%
n
.
•
.
'
-10 -20 -30 -40 10 0
1
0.04 ppm
n = 30
0%
.
.
-
'
r
_
n
-10
i
[— |
0.10 ppm
n = 32
31%
n n
-
-
_
"
r
-20 -30 -40 10
0.06 ppm
n = 30
3%
.
.
.
.
-
n
—
0.12 ppm
n = 49
46%
-i
In
n
nilm n n
0 -10 -20 -30 -40
0.08 ppm
n = 30
17%
n
10 0 -10 -20 -30 -40 10 0 -10 -20 -30 -40 10 0 -10 -20 -30 -40 10 0 -10 -20 -30 -40
FEV., (%change)
Figure 8-1 A,B. Frequency distributions of FEVt changes following 6.6-h exposures to a
constant concentration of O3 or filtered air. During each hour of the
exposures, subjects were engaged in moderate exercise for 50 minutes.
With increasing O3 concentration, the distribution of responses becomes
asymmetric, with a few individuals exhibiting large FEVt decrements.
Note that the percentage in each panel indicates the portion of subjects
having a FEVt decrement in excess of 10%.
Source: Panel A, McDonnell (1996); Panel B, Adams (2002, 2006), pre- and post-FEV; data for each subject
provided by author.
summarized in the 1996 O3 AQCD. Referring to this figure, the group mean FEVj response
following continuous (square-wave) exposure to 0.08 ppm O3 was relatively small (between a
5 and 10% decrement). However, 18% of the exposed subjects had moderate FEVj decrements
of 10 to 20%, and 8% experienced large FEVj decrements of greater than 20%. This serves to
emphasize that, while group mean responses may be small and seem physiologically
insignificant, some individuals can experience distinctly larger effects under similar O3 exposure
conditions. Newer data from Adams (2002, 2006), as illustrated in Figure 8-1B, demonstrate
notable interindividual variability for O3 exposure concentrations at and even below 0.08 ppm.
8-17
-------
Following 6.6-h continuous exposure to 0.08 ppm O3 under intermittent, moderate exercise
conditions, the group mean FEVj decrement was 5%, but 17% of the subjects had greater than a
10% decrease in FEVj. Following exposure to 0.06 ppm O3, the group mean FEVj decrement
was less than 2%. However, five subjects still had greater than a 5% decrease in FEVj whereas
only one experienced this magnitude of effect following exposure to filtered air.
It should be noted that spirometry typically improves in healthy young adults with exercise
exposures to filtered air (FA). The term "O3-induced" was used in Chapter 6 and Annex 6 to
designate effects that have been corrected for these filtered air responses. For healthy adults,
an O3-induced change in lung function is the difference between the decrement experienced
with O3 exposure and the improvement observed with filtered air exposure. The FEVj responses
illustrated in Figure 8-1 were not corrected for the responses following filtered air exposures.
For comparison to Figure 8-IB, O3-induced FEVj responses from the Adams (2002, 2006)
studies are illustrated in Figure 8-2. For the exposures to 0.04, 0.06, and 0.08 ppm O3,
an O3-induced FEVj decrement of > 10% was experienced by 7, 7, and 23% of subjects,
respectively. Effects of a 0.04 ppm O3 exposure were not apparent when simply comparing
pre-post FEVj (Figure 8-IB), whereas they are evident when considering O3-induced FEVj data
(Figure 8-2). The distinction between an O3-induced change and a post- versus preexposure
change is particularly important in individuals with respiratory disease who may experience
exercise-induced decrements in pulmonary function during both filtered air and O3 exposures.
Other new studies (assessed in Chapter 6 and Annex 6 of this document) that evaluated
responses in hundreds of subjects clearly indicate that O3-related FEVj decrements and symptom
responses decrease with age beyond young adulthood (18 to 20 years). For example, Hazucha
et al. (2003) studied possible gender and age differences in O3 responsiveness and found that
young females lose O3 sensitivity faster than young males, but the rate is about the same for both
genders by middle age.
A few controlled human exposure studies that explored a triangular exposure profile of
average O3 concentrations between 0.08 to 0.12 ppm over 6.6 to 8 h (to more closely mimic the
typical ambient O3 exposure pattern) observed greater overall FEVj decrements with triangular
exposures compared to constant or square-wave exposures (see Appendix Table 8A-1 and Annex
Figure AX6-3). Furthermore, the peak FEVj decrements observed during triangular exposures
exceed those observed during square-wave exposures. At a lower average O3 concentration of
8-18
-------
t« ou
.«[ SO-
ls 40-
-------
earlier research on repeated exposure of rats to an episodic profile of O3 found small (but
significant) decrements in lung function that were consistent with early indicators of focal
fibrinogenesis in the proximal alveolar region.
In the 1996 O3 AQCD, O3-induced decreases in inspiratory capacity were hypothesized to
result from neurogenic inhibition of maximal inspiration due to stimulation of C-fiber afferents,
either directly or from O3-induced inflammatory mediators. Earlier human studies (Coleridge
et al., 1993; Hazucha and Sanf Ambrogio, 1993) reported a role for bronchial C-fibers and
rapidly adapting receptors as primary vagal afferents responsible for O3-induced changes in
ventilatory rate and depth. As discussed in Chapter 6, the newer results of Passannante et al.
(1998) also support C-fiber stimulation as a primary mechanism of the O3-induced reduction in
inspiratory capacity and suggest a role for nociceptive mechanisms. This neurogenic mechanism
is also likely related to effects such as increased airway responsiveness and lung inflammation.
Only two studies evaluated lung function changes due to O3 exposure in patients with
preexisting respiratory diseases under experimental controlled exposure regimens. Studies using
COPD patients found that O3 exposure-induced minimal effects (nonsignificant decreases in lung
function) in this population (Gong et al., 1997a). However, newer studies of asthmatics by
Alexis et al. (2000) still continue to indicate that asthmatics are at least as sensitive, if not more,
than healthy subjects, based on pulmonary function deficiencies detected by spirometric
analyses.
In addition to effects of O3 exposure on the large airways as indicated by spirometric
responses, O3 exposure also affects the function of the small airways and parenchymal lung
tissue. Studies reported by Foster et al. (1993, 1997) that examined the effect of O3 on
ventilation distribution in healthy adult males suggest a prolonged O3 effect on the small airways
and ventilation distribution in some individuals. Animal toxicology studies have shown the
centriacinar region (CAR) of the lung (the segment between the last conducting airway and the
gas exchange region) to be a region highly susceptible to O3-induced damage (epithelial cell
necrosis and remodeling of respiratory bronchioles). Common pulmonary function tests do not
measure acute changes in the small airways of the CAR. Identification of acute effects of O3 in
small airways, if any, would lend additional support for concerns about long-term effects of
repeated O3 exposures.
8-20
-------
Airway Responsiveness
The 1996 O3 AQCD recognized the induction of airway hyperresponsiveness (AHR) in
humans on exposure to O3, which usually resolves in 18 to 24 h after exposure in a majority of
subjects, but may persist in some individuals for longer periods. Ozone-induced nonspecific
AHR could have clinical implications in asthmatics, possibly putting them at potential increased
risk for more prolonged bouts of bronchoconstriction in response to various triggering stimuli in
the ambient air. Three new studies (discussed in Chapter 6) suggest that (1) subjects with
asthma developed tolerance to repeated O3 exposures in a manner similar to normal subjects; but
there were more persistent effects of O3 on airway responsiveness, which only partially
attenuated when compared to filtered air control exposures (Gong et al., 1997b); (2) the
enhancement of allergen responsiveness after O3 exposure appears to be time-dependent,
suggesting that the timing of allergen challenge in O3-exposed subjects with allergic asthma is
important (Torres et al., 1996); and (3) subjects with rhinitis exhibit significant, clinically
relevant decreases in pulmonary function in the early phase allergen response (Holz et al., 2002).
These observations suggest that O3 exposure may be a clinically important factor that can
exacerbate responses to other ambiently-encountered bronchoconstrictor substances in
individuals with preexisting allergic asthma and that its influence may be both immediate and
persist for relatively long periods of time (a few days).
An extensive laboratory animal study database (see Chapter 5 Annex Tables) exploring
the effects of acute, long-term, and repeated exposures to O3, clearly indicates that induction of
AHR occurs at relatively high O3 concentrations. Only one study that utilized very low O3
(0.05 ppm) concentrations found that acute exposure induced AHR in a few inbred strains of rat,
suggesting a role for genetic susceptibility in this process (Depudyt et al., 1999). As seen with
humans, acute changes in AHR do not persist upon long-term exposure of laboratory animals to
near-ambient concentrations of O3; and attenuation has been observed. Both human and animal
studies indicate that airway responses are not associated with inflammation, but they do suggest
a likely role for neuronal involvement.
Lung Inflammation, Permeability, and Biochemical Alterations
Respiratory tract inflammation and increased cellular permeability are the two most
important biological markers of O3-induced injury response mechanisms in both humans and
8-21
-------
animals. These distinct, independent biological effects have been observed in all species studied
in response to acute O3 exposure.
Several controlled human exposure studies reviewed in the 1996 O3 AQCD clearly showed
that a single acute exposure of humans to moderate O3 concentrations (>0.08 to 0.1 ppm) while
exercising at moderate to heavy levels disrupts the barrier function of the lung, initiating a
cascade of responses indicative of increased lung permeability and pulmonary inflammation,
as indexed by a variety of cellular and biochemical changes, e.g., increases in levels of
polymorphic neutrophils (PMNs) and protein in lung fluid. Both the inflammatory response and
increased lung permeability have been observed as early as 1 h and persisted for at least 18 h.
The newer studies reviewed in this document (see Chapter 6) provide additional information on
three different aspects of O3-induced inflammatory responses, such as (1) intersubject variability;
(2) differential attenuation profiles for different inflammatory markers; and (3) effects of
repeated exposures.
Soluble mediators of inflammation (e.g., the cytokines IL-6 and IL-8), as well as
arachidonic acid metabolites (e.g., PGE2, PGF2a, thromboxane, and leukotrienes [LTs] such
as LTB4), have been measured in the BAL fluid of humans exposed to O3. In addition to their
role in inflammation, many of these compounds have bronchoconstrictive properties and may be
involved in increased airway responsiveness following O3 exposure. The time course for the
inflammatory responses (including recruitment of neutrophils and other soluble mediators) is not
clearly established, but differential attenuation profiles for many of these parameters are evident
from the meta-analysis by Mudway and Kelly (2004) of 21 controlled human exposure studies.
Recent studies continue to support the observation made in the 1996 O3 AQCD that
repeated O3 exposures in humans also induce ongoing cellular damage irrespective of attenuation
of the inflammatory responses and lung function decrements (Devlin et al., 1997; Torres et al.,
2000). Devlin et al. (1997) studied inflammatory responses of humans repeatedly exposed to
0.4 ppm O3 for 5 consecutive days. Several indicators of inflammation (e.g., PMN influx, IL-6,
PGE2, fibronectin) were attenuated after 5 days of exposure (i.e., values were not different from
FA). Several other markers (LDH, IL-8, total protein, epithelial cells) did not show attenuation,
indicating that tissue damage probably continues to occur during repeated exposure. Recovery
of the inflammatory response occurred for some markers after 10 days, but some did not return
to normal even after 20 days. When re-exposed 2 weeks later, changes in BALF indicated that
8-22
-------
epithelial cells appeared to be fully repaired (Devlin et al., 1997). Kopp et al. (1999) observed
inflammatory responses only after the first O3 peak in summer; and their absence late in summer
(even after exposure to higher levels of O3) may be due to attenuation of such responses upon
repeated O3 exposures.
Numerous studies reported acute O3-induced changes in lung epithelial permeability
assessed by indirect assay (increased levels of albumin and protein in BALF). A few other
studies demonstrated O3-induced epithelial cell permeability through direct assessment of
clearance of 99mTc-DTPA (technetium-99m labeled diethylene triamine pentaacetic acid).
For example, Kehrl et al. (1987) showed increased 99mTc-DTPA clearance in healthy young
adults at 75 minutes postexposure to 0.4 ppm O3 for 2 h. More recently, Foster and Stetkiewicz
(1996) have shown that increased 99mTc-DTPA clearance persists for at least 18 to 20 h post-O3
exposure (130 min to average O3 concentration of 0.24 ppm), and the effect is greater at the lung
apices than at the base.
Laboratory animals, like humans, exhibit varying degrees of sensitivity to O3 exposure
(see Chapter 5 for detailed discussion); and this is evident even for the induction of pulmonary
inflammation and permeability. New animal toxicology studies of O3-induced inflammation
assessed in Chapter 5 indicate that the lowest acute O3 exposures that had an effect on mouse
lung inflammation was 0.12 ppm for 24 hours. Shorter durations (8 h) required greater O3
exposure (0.26 ppm) for effects on epithelial permeability but had no effect on inflammation.
The lowest acute O3 concentrations that had an effect on epithelial permeability or inflammation
in the rat were 0.5 ppm for 3 hours or 0.12 ppm for 6 hours. Also, increased lung permeability
and inflammation occurred in rabbits with O3 exposures as low as 0.10 ppm for 2 h/day for
6 days. Subchronic exposures in animals suggest that permeability changes are transient (and
species-dependent) and return to control levels even with continuing exposure. Chronic animal
O3 exposure studies suggest a role for persistent inflammation in O3-induced alterations in lung
structure and function. Significant remodeling of epithelium and underlying connective tissues
in distal airways have been reported in rats exposed to 0.25 ppm O3 (12 h/day for 6 wk) and in
monkeys exposed to 0.2 ppm O3 (8 h/day for 90 d). Various factors such as viral infection,
chemotactants and oxidized matrix fragments are also implicated in the establishment and
persistence of O3-induced inflammation.
-------
Animal toxicology and human in vitro studies that evaluated biochemical mediators
implicated in injury and inflammation found alterations in the expression of cytokines,
chemokines, and adhesion molecules, indicative of an ongoing active stress response as well as
injury repair and regeneration processes. Both animal and human studies indicate cellular and
biochemical changes associated with inflammation and increased permeability, but the
relationship between these changes and possible airway remodeling is not known.
As discussed in the earlier section on dosimetry, interaction of O3 with the lipid
constituents of pulmonary surfactant has been proposed as one of the key mechanisms by which
O3 exerts its toxic effects. Experimental evidence clearly indicates a role for the interaction of
O3 with lipid constituents of the ELF and cell membranes and the generation of lipid ozonation
products and secondary redox mediators in the initiation of site-specific cell-injury response
cascades. One such lipid ozonation product, 4-hydroxynonenal, has been found to bind to
proteins and increased protein adducts in human alveolar macrophages, suggesting a role for
4-hydroxynonenal in acute cell toxicity. Cholesterol, the most abundant neutral lipid in
pulmonary surfactant is susceptible to attack by O3, resulting in formation of multiple oxidized
cholesterol products, e.g., cholesterol epoxide. A 20-fold increase in cholesterol epoxide in the
BALF from mice exposed to 0.5 ppm O3 for 3 h suggests a potential role for this oxidation
product in O3 toxicity (Pulfer et al., 2005). Species- and region-specific increases in lung
xenobiotic metabolism have been observed in response to both short- and long-term O3
exposure. It has been well recognized that antioxidants in the ELF confer some protection
against O3 toxicity. But even with environmentally relevant exposures, O3 reactivity is not
quenched. Further, antioxidant reactivity with O3 is both species-specific and dose-dependent.
Lung Injury
Pulmonary histopathological observations reported in the 1996 O3 AQCD suggested that
similar types of alterations occurred in lung morphology in all laboratory animal species studied,
including primates, upon short-term O3 exposure. The cells in the CAR were recognized as a
primary target, possibly because the CAR receives the greatest dose of O3 delivered to the lower
respiratory tract. With chronic O3 exposure, structural changes were also observed in this region
of the respiratory tract (the region typically found affected in most chronic airway diseases of the
human lung). Simulated seasonal exposure studies in animals also suggested that seasonal O3
8-24
-------
exposures may have potential for cumulative impact over many years. Ciliated cells in the nasal
cavity and airways and Type I epithelial cells in the gas-exchange region have also been
identified as targets. Though acute O3 exposure induces structural changes such as fibrosis in the
CAR, these structural alterations appear to be partially transient, with recovery shortly post O3
exposure; but the time for recovery is dependent on the species and the dose of O3. Long-term or
prolonged exposure to O3 has been found to cause chronic lesions similar to early lesions of
respiratory bronchiolitis, which have the potential to progress to fibrotic lung disease. Some of
the morphological changes associated with long-term O3 exposures, e.g., increases in
hyperplastic epithelial cells, appear to reverse following cessation of O3 exposure. However,
in the underlying interstitium of the CAR, proliferation of fibroblasts creates excess noncellular
matrices. These processes are only partially reversible and may progress following cessation of
O3 exposure. This suggests initiation of focal interstitial fibrosis, which can progress to
nonreversible structural damage to lung tissue.
Reports of morphological changes following chronic O3 exposures in animal studies
(rodents and primates) published since the 1996 AQCD allude to the earlier findings assessed in
that document. In rats, the effects of chronic -0.5 ppm O3 exposure included mucous cell
metaplasia, hyperplasia of the nasal epithelium, increased mucosubstances, and increased Bel-2
protein levels. In mice, lifetime exposures of 0.5 ppm O3 were linked to similar outcomes.
Taken together, the rodent studies suggest that O3 exposure may have the potential to induce
similar long-lasting alterations in human airways. A series of new studies that utilized infant
rhesus monkeys and simulated seasonal ambient exposure (0.5 ppm 8 h/day for 5 days, every
14 days for 11 episodes) reported remodeling in the distal airways; abnormalities in tracheal
basement membrane; eosinophil accumulation in conducting airways; and decrements in airway
innervation, again confirming the potential greater injury due to seasonal exposure compared to
continuous exposure alluded to in the 1996 O3 AQCD.
Host Defense
Based on a small number of studies available at the time, the 1996 O3 AQCD concluded
that short-term O3 exposure of laboratory animals and humans impairs alveolar macrophage
(AM) clearance of viable and nonviable particles from the lungs and decreases the effectiveness
of host defenses against bacterial lung infections in animals and perhaps humans. A single
8-25
-------
controlled human exposure study reviewed in the 1996 O3 AQCD found decrements in the
ability of alveolar macrophages to phagocytose microorganisms upon exposure to 0.08 to
0.1 ppm O3 for 6.6 h during moderate exercise (Devlin et al., 1991).
Other evidence for O3-induced dysfunction of host defense components and for subsequent
enhanced susceptibility to bacterial lung infection has been derived from laboratory animal
studies. Acute exposures of 0.08 ppm (3 h) O3 were found to result in the mortality of mice due
to Streptococcal bacterial infection. Changes in antibacterial defenses appear to be dependent on
exposure regimens, species and strain of test animal, species of bacteria, and age of animal (with
young mice being more susceptible to the effects of O3, for example). Animal toxicology studies
indicated that acute O3-induced suppression of alveolar phagocytosis and immune functions
observed in animals appeared to be transient and attenuated with continuous or repeated
exposures.
It has also been reported that O3 exposures can interfere with AM-mediated clearance in
the respiratory region of the lung and with mucociliary clearance of the tracheobronchial
airways. Ozone-induced perturbations in the clearance process have been found to be dose-
dependent, with low dose exposures accelerating clearance and high doses slowing the clearance
process. Some respiratory tract regional- and species-specific differences have also been
observed.
In vitro cultures of epithelial cells obtained from nonatopic and mild atopic asthmatics
exposed to 0.01 to 0.1 ppm O3, exhibited significantly increased permeability compared to cells
from normal subjects, thus indicating a potential inherent susceptibility of cells from asthmatics
for O3-induced permeability. Newer in vitro cell culture studies of human bronchial epithelial
cells indicate O3-induced exacerbation of human rhinovirus type 16 infection (Spannhake et al.,
2002); and new animal toxicology studies have shown O3-induced modulation of cell-mediated
immune responses affecting the onset and persistence of infection in rats (Cohen et al., 2001,
2002).
The available data at this time indicate that acute O3 exposure has a potential to impair host
defense capability, primarily by interfering with the functions of alveolar macrophages. Any
impairment in macrophage function may lead to decreased clearance of microorganisms or
nonviable particles. Compromised alveolar macrophage functions in asthmatics may increase
their susceptibility to other O3 effects, the effects of particles, and respiratory infections.
8-26
-------
Cardiovascular Effects
Based on the paucity of then-available information, the 1996 O3 AQCD did not accord
much, if any, attention to possible O3-related cardiovascular effects. However, since then,
an emerging body of animal toxicology evidence is beginning to suggest that hematological and
thermoregulatory alterations (in heart rate and/or core body temperature) may mediate acute O3
cardiovascular effects. For example, it is beginning to be recognized that O3-induced lung injury
and permeability changes, as well as O3-induced alterations in hemodynamics, may lead to
cardiovascular system effects. Also, interactions of O3 with ELF lipids and surfactants result in
lipid ozonation products and reactive oxidant species (ROS) that have the potential to penetrate
the epithelial barrier and to initiate toxic effects on the cardiovascular system.
Earlier studies in rats indicated a potential role for platelet activating factor (PAF) in the
O3-induced inflammatory response. Recent observations of O3-induced generation of oxysterols
and p-epoxides from cholesterol in surfactant also suggest that these lipid ozonation products
(like lysophospholipids) might exhibit PAF-like activity and contribute to clotting and
thrombolytic effects in the cardiovascular system.
Other studies carried out using isolated perfused rat lung model (Delaunois et al., 1998)
indicate inhibition of pulmonary mechanical reactivity to bronchoconstrictors and persistent
vasoreactivity of the vascular bed upon exposure to O3 (0.4 ppm for 4 h). Newer studies have
also now found that acute (less than 5 h) 0.1 ppm O3 exposure caused decreased heart rate in
young, but not old, rats (Arito et al., 1997).
Only one human experimental O3 exposure study (Gong et al., 1998) evaluated potential
cardiovascular effects in normal and hypertensive adult males. Various cardiovascular and
hemodynamic parameters were monitored upon exposure to O3 (0.3 ppm for 3 h) with
intermittent exercise. No significant O3-induced differences were observed in ECG, heart rate,
or blood pressure in either normal or hypertensive subjects. An overall increase in myocardial
work and impairment in pulmonary gas exchange was observed, however, that might be
clinically important in patients with preexisting cardiovascular impairment with or without
concomitant lung disease.
8-27
-------
8.4.3 Biological Basis for O3 Health Effects Assessment
The scientific knowledge base gained from animal toxicological studies and experimental
studies of human volunteers in clinical settings (discussed in Chapters 4, 5, and 6 of this
document) has notably expanded the knowledge base beyond that available at the time of 1996
O3 AQCD. This section provides an interpretive integration of key findings derived from the
experimental knowledge base that can aid in evaluating the biological plausibility for health
effects observed in O3-related epidemiological studies discussed later in this chapter.
Animal-to-Human Extrapolation Issues
The physiological and biochemical observations summarized in Table 8-1 provide an
overview of the main types of acute O3-induced health effects as demonstrated by toxicological
studies of humans and animals. This table was generated from those experimental studies (see
Annexes for Chapters 5 and 6 for experimental details) that utilized exposure regimens of varied
concentration and duration that are environmentally relevant. As noted above in Section 8.4.1,
observed acute O3 effects are mostly transient (generally persisting <24 to 48 h) and attenuate
over time. However, the time-line for resolution of many of these physiological and biochemical
parameters in normal and human subjects with underlying cardiopulmonary diseases follow
different profiles, as presented in Figure 8-3. Alterations in the cellular and molecular profiles
observed in human airway epithelium upon acute exposure to O3 evolve over time (Figure 8-3),
and the knowledge of this profile is valuable in assessing biological plausibility to integrate
across evidence for various health endpoints.
Basic similarities in physiological, biochemical, and pathological processes that exist
between human and other mammalian species are derived from the high degree of genome
sequence homology that exists across various species. This homology reinforces the
significance of knowledge gained on the initiation, progression and treatment regimes for
various disease processes across animal species. This homology is also apparent in acute O3-
induced effects, especially on the respiratory tract of human and animal species as presented in
Table 8-1 and Figures 8-3 and 8-4. The commonality of phenomenon observed in humans and
rats with regard to respiratory system effects (in terms of spirometry, ventilatory response, host
defense and inflammation) and their attenuation adds strength to animal-human extrapolations.
Such similarities observed at higher levels of cellular organization (neutrophilic inflammation,
8-28
-------
Table 8-1. Acute O3-Induced Physiological and Biochemical
Changes in Human and Animals
Physiological/Biochemical
Alterations
Human Exposure Studies1>2
Animal Toxicology Studies 3'4
Pulmonary Function:
Airway Responsiveness:
Inflammation:
Reactive Oxygen Species:
Host Defense:
Lung Injury:
Morphology:
Susceptibility:
Cardiovascular Changes:
1 FEVj
T Frequency of breathing
(rapid, shallow)
1 FVC
(cough, breathing discomfort,
throat irritation, wheezing)
Mild bronchoconstriction
T (neuronal involvement)
Change in lung resistance
Yes
T inflammatory mediators
T
T particle clearance
T permeability
i AM phagocytosis
Yes
Age,
Interindividual variability
Disease status
Polymorphism in certain genes
being recognized
Impairment in arterial O2 transfer
Ventilation-perfusion mismatch
(suggesting potential arterial
vasoconstriction)
T rate pressure product5
T myocardial work5
T Frequency of breathing
(rapid, shallow)
1 FVC
T (vagal mediation)
Change in lung resistance
Yes
T inflammatory mediators
T
T particle clearance
T permeability
I clearance of bacteria
T severity of infection
T mortality & morbidity
Yes
Species-specific differences
Genetic basis for susceptibility
indicated
Heart rate
1 core body temperature
T atrial natriuretic factor
Role for platelet activity factor
(PAF) indicated
Increased pulmonary vascular
resistance
1 Controlled chamber exposure studies in human volunteers were carried out for a duration of 1 to 6.6 h with O3
concentration in the range of 0.04-0.40 ppm with intermittent exercise. See text for discussion of O3 levels
shown to cause different types of effects listed.
2 Data on some biochemical parameters obtained from in vitro studies of cells recovered from B ALF.
3 Responses were observed in animal toxicology studies with exposure for a duration of 2 to 72 h with O3
concentration in the range of 0.1 to 2.0 ppm.
4 Various species (mice, rat, guinea pigs and rabbit) and strains.
5 In hypertensive subjects.
8-29
-------
Resolution Time-Line for Acute Ozone-Induced Physiological and Biochemical Responses in Humans
oo
OJ
o
c
Pulmonary Function:
Spirometric Changes (FEV^)
airways uncions( 25-75)
Mediators
irway esponsiveness.
Host Defense:
Injury/permeability:
) 2 4 6 12 18 20 24 36 48 72 Hours
I I I // \ ss I yy I I // 1 // I // \
* Hyperresponsive Subjects
** Asthmatics
•« »> Partial
< ^ Complete
Figure 8-3. Resolution time-line for the respiratory, physiological, and biochemical parameters are derived
from studies reported in Chapter 6 and Chapter 6 Annex.
-------
Postulated Cellular and Molecular Changes In Human Airway
Cells In Response to Acute Exposure to Ozone
Response Time
I ff t ft 7 h ^^^^k. Chemical reaction with ELF and epithelial cell membrane,
^^^^r Generation of ozortation products, Bpid peroxides
Early 2-24 h
(Neutrophil infiltration)
Lipid ozonation products
Pro-inflammatory mediators (neutrophil chemotaxins,
Anti-inflammatory mediators (prostanoWs)
Cytokines. proteases
Late (12-24 h)
(Eosinophii/monocyte
infiltration)
Lipid ozonation products
Increase in pro-inflammatory mediators (monocyte chemotaxirts,
Decrease in anti-inflammatory mediators (prostanoids)
Release of cytokinas
increased expression of intraceltular adhesion molecules
Increased synthesis of collagen, fibronaetin
Release of leukocyte prateinase inhibitors
increased synthesis of antioxidants (SOD, GSH, catalase)
Figure 8-4. Acute (1-8 h) O3 exposure-induced cellular and molecular changes and
timelines for their resolution depicted here are derived from the data
reported in Leikauf et al. (1995) and Mudway and Kelly (2000).
macrophage phagocytosis processes) have increased the value and importance of animal studies
in generating important data that are impossible to collect in human studies but which may
corroborate both human clinical and epidemiologic studies.
Extrapolation of results derived from laboratory animal studies to humans involves a
combination of dosimetry, end point homology and species sensitivity, particularly in the case of
exposure and health outcome analyses. However, existing extrapolation models have not yet
been sufficiently validated to allow for highly confident quantitative animal-to-human
extrapolation for O3 effects. Still, qualitative extrapolation appears to be reasonable for some
endpoints. For example, based on inflammatory markers in BALF, a 2 ppm O3 exposure in
nonexercising rats approximates to a 0.4 ppm exposure in exercising humans (Hatch et al.,
1994). This observation lends support to the use of some of the animal toxicology data derived
from relatively high O3 concentration exposure regimens in understanding putative molecular
changes likely to be associated with acute O3 exposure in humans.
8-31
-------
The time courses for induction and resolution of various types of O3 effects experimentally
induced in laboratory settings (see Figures 8-3 and 8-4) may provide a basis for reasonable
projection of likely lag times for observing associations between ambient O3 concentrations/
exposures and analogous effects in epidemiologic studies of human populations. For example,
it would not be unreasonable to expect that pulmonary function decrements could be detected
epidemiologically within lags of 0 (same day) or 1 or 2 days following O3 exposure, given the
rapid onset of spirometric changes and their persistence for 24 to 48 hours among
hyperresponsive human subjects in clinical studies. On the other hand, although asthmatic
individuals may begin to experience symptoms soon after O3 exposure, it may take a day or so,
coupled with breathing difficulties secondary to O3-induced hyperresponsiveness that may
persist for 2 to 3 days, for members of that sensitive population group to seek medical attention.
This could possibly be reflected by epidemiologic observations of significantly increased risk for
asthma-related ED visits or hospital admissions with 1- to 3-day lags, or, perhaps, enhanced
distributed lag risks (combined across 3 days) for such morbidity indicators. Analogously, one
might project increased mortality within 0 to 3 day lags as a possible consequence of O3-induced
increases in prothrombotic agents, such as platelet aggregating factor (PAF) or lipid ozonation
products (e.g., cholesterol epoxide) arising from inflammation cascades occurring within 12 to
24 hours of O3 exposure (see Figure 8-4).
Similarly, the presence of apparent O3-induced lesions in animals from chronic O3
exposure studies (12 to 24 months) indicate morphological alterations that may analogously
occur in humans with long-term (months, years) chronic exposure to relatively high O3 levels,
but specific O3 concentrations and exposure patterns that may produce analogous alterations in
human lungs remain to be substantiated.
8.4.4 Epidemiologic Evidence
Epidemiologic evidence available at the time of the 1996 O3 AQCD indicated potential
health effects associated with acute ambient O3 exposures. The 1996 O3 AQCD further stated
that only suggestive epidemiologic evidence existed for health effects of chronic ambient O3
exposure in the population, and this was partly due to an inability to isolate potential effects
related to O3 from those of other pollutants, especially PM (U.S. Environmental Protection
Agency, 1996a). The availability of numerous recent epidemiologic studies, with some studies
8-32
-------
designed explicitly to evaluate ambient O3 exposure effects (as discussed in detail in Chapter 7
of this document), makes it possible to better assess the health effects of both acute and chronic
ambient O3 exposure. There exist a number of issues and uncertainties associated with the
interpretation of O3 health effects evidence in epidemiologic studies, which are discussed in
Sections 7.1.3 and 7.6 of Chapter 7 in this document. These include the use of various indices to
represent O3 exposure, exposure measurement errors (i.e., differences between ambient
concentrations and personal exposure), lag period between exposure and effect, potential
confounding by temporal and meteorological factors, potential confounding by copollutants, the
concentration-response function, and heterogeneity of O3 health effects.
The recent epidemiological studies that have been conducted in areas across the United
States and Canada, as well as in Europe, Latin America, Australia and Asia, are reviewed in
Chapter 7 (for details see Annex to Chapter 7). Present discussions in this integrative synthesis
chapter focus on findings from studies carried out in the United States and Canada. Important
features of these epidemiologic studies (e.g., study location, study population, mean O3
concentrations, health risk estimates) are summarized in Chapter 8 Appendix Tables 8A-2 to
8A-5. The studies are ordered in those tables by the 98th and 99th percentile values of the
calculated 8-h max O3 data for the full study period. The 98th and 99th percentile values were
selected as they represent a high concentration that roughly approximates a 4th maximum
concentration, depending on the study period length. For studies that did not have this data
available, their ordering was approximated based on the mean O3 concentrations observed in
the study.
8.4.4.1 Acute Ozone Exposure Studies
Numerous epidemiological studies carried out over the past decade have added evidence to
the knowledge base assessed in the 1996 O3 AQCD, which included both (a) individual-level
summer camp and exercise studies that established a relationship between ambient O3 exposure
and lung function decline and (b) aggregate population studies that suggested positive
relationships for O3-related respiratory morbidity endpoints (i.e., respiratory ED visits and
hospitalizations). The new studies reviewed in Chapter 7 in this document include numerous
field/panel studies and population studies from various regions in the United States and abroad.
In field/panel studies on the effects of air pollution exposure, the most common health outcomes
8-33
-------
measured were lung function and respiratory symptoms. The population studies examined daily
ED visits, hospital admissions, and mortality data.
Field/Panel Studies of Acute Exposure Effects
Pulmonary Function and Respiratory Symptoms
Many of the new field/panel studies reviewed in Chapter 7 and the controlled human
exposure studies reviewed in Chapter 6 of this document provide additional data supporting two
key findings reported in the 1996 O3 AQCD. One key finding was that acute O3 exposure was
associated with a significant decline in lung function parameters. Ozone-related lung function
decrements were most notable in children, as indicated by the results from the meta-analysis of
summer camp studies by Kinney et al. (1996). Similar responses were reported for children and
adolescents exposed to O3 in ambient air in the summer camp studies as were found with
exposure to O3 in clean air for 1 to 2 h while exercising in clinical studies. A second key finding
was that adults who work or exercise outdoors were also found to be vulnerable to O3-associated
declines in lung function due to their greater exposure to O3 during periods of increased physical
activity.
In a number of newly available field/panel studies, FEVj was measured in panels of
exercising children, outdoor workers, and adult hikers exposed to ambient O3 while experiencing
elevated exertion levels. Collectively, the results of the new studies (discussed in
Section 7.2.3.1) confirm and extend the findings from analogous field/panel studies and
experimental controlled human exposure studies assessed in the 1996 O3 AQCD. Acute O3
exposures prolonged over several hours and combined with elevated levels of exertion or
exercise were found to have magnified effects on lung function, as evaluated in terms of FEVj.
Only studies carried out in the U.S. and Canada that evaluated lung function changes are
discussed here (see Appendix Table 8A-2 for details).
Brauer et al. (1996) measured lung function in berry pickers in British Columbia, Canada.
The mean ambient 1-h max O3 was 40.3 ppb (SD 15.2). Significant O3-related decreases in FEVj
were observed in the morning and afternoon, but not across the day. Another field study on
adult hikers by Korrick et al. (1998) observed that the mean FEVj decline in healthy hikers was
marginal, but a 4-fold greater decline was observed in hikers with asthma and wheeze. Both
studies, however, did not evaluate the possible role of confounding by copollutants. Other
8-34
-------
studies from Germany and Mexico City provided additional supportive evidence that healthy
adults with prolonged periods of exposure and exertion experienced declines in lung function.
Several panel studies that examined lung function changes in children, including a few
U.S. studies (Neas et al., 1995, 1999; Linn et al., 1996), observed small declines in FEVj or in
peak expiratory flow (PEF). It should be noted, however, that the group mean responses
reported in panel or population studies may mask interindividual differences, e.g., larger lung
function changes in sensitive populations (as discussed in Section 8.4.2 and in more detail in
Chapter 6).
Panel studies that evaluated lung function decrements in asthmatic children collectively
indicated O3-induced decrements, although most of the individual-study estimates were not
statistically significant. Results from a multicity study of asthmatic children from eight urban
areas in the U.S. (Mortimer et al., 2002) suggest a small, but significant decline in PEF based on
a multiday distributed lag model. Of all the pollutants examined, including O3, PM10, NO2, and
SO2, only O3 was found to be associated with morning PEF. The small change in lung function
observed likely reflects the low mean O3 levels across the eight cities studied (8-h avg O3 of
48 ppb). Overall, acute ambient O3 exposure was observed to be associated with lung function
decrements at O3 concentrations ranging down to a 98th percentile 8-h max of 55 to 60 ppb for
potentially vulnerable and susceptible populations, as shown in Appendix Table 8A-2.
In addition to pulmonary function, the majority of asthma panel studies also evaluated the
relationship between O3 and respiratory symptoms in asthmatic children. The results obtained
from these studies show some inconsistencies, with some indicating significant positive
associations and other well-conducted, albeit relatively smaller, studies not finding such effects.
Overall, however, the multicity study by Mortimer et al. (2002) and several credible single-city
studies (e.g., by Gent et al., 2003) indicate a fairly robust positive association between ambient
O3 concentrations and increased respiratory symptoms in asthmatics.
Airway Inflammation
A few epidemiologic studies, conducted in Germany and Mexico City, examined the effect
of acute ambient O3 exposure on airway inflammation in children. Results indicated that
associations were found largely on days or in study locations having the higher O3
concentrations (1-h max O3 of approximately 100 ppb) among those present. At locations with
8-35
-------
lower O3 levels, no association was detected between upper airway inflammation and ambient O3
concentrations across the entire study period (Kopp et al., 1999). However, a more detailed
analysis showed that the first significant O3 episode of the summer was associated with increases
in inflammatory markers, but subsequent and even higher O3 episodes had no effect, suggesting
an adaptive response.
Cardiophysiological Endpoints
A few recent air pollution epidemiology studies have also evaluated the potential effects of
ambient O3 on Cardiophysiological endpoints. Based on the discussions presented in Chapter 7
(see Section 7.2.7), the limited data available at this time are suggestive of altered heart rate
variability (HRV), ventricular arrhythmias, and MI incidence possibly being associated with
acute O3 exposures; but, overall, these findings must currently be considered inconclusive.
Population Studies of Acute Exposure Effects
Emergency Department Visits and Hospital Admissions
Many time-series studies reviewed in the 1996 O3 AQCD indicated positive associations
between O3 air pollution and increased respiratory hospital admissions and ED visits, providing
strong evidence for a relationship between O3 exposure and increased exacerbations of
preexisting respiratory disease in the general public at 1 h-max O3 concentrations <0.12 ppm
(120 ppb). Analyses of data for the northeastern United States suggested that O3 air pollution
was associated with a substantial portion (on the order of 10 to 20%) of all summertime
respiratory hospital visits and admissions. Several new studies have been published in the past
decade examining temporal associations between O3 exposures and emergency department visits
and hospital admissions for respiratory diseases.
Studies conducted in the United States and Canada examining the association between
acute O3 exposure and asthma ED visits are summarized in Appendix Table 8A-3. The risk
estimates presented in Appendix Table 8A-3 include those using all-year and warm-season only
O3 data. A majority of the studies conducted in the U.S. and Canada examined the effects of O3
during the warm season only. An analysis of respiratory disease ED visits in Atlanta, GA over
an 8-year period indicated a significant positive association between ambient O3 concentrations
and asthma visits during the warm season (Peel et al., 2005). The mean 8-h max O3 from March
8-36
-------
to November was 55.6 ppb (SD 23.8) in this study. A study conducted in Saint John, Canada
observed a positive association between O3 and asthma emergency visits during the warm season
in individuals 15 year or older at even lower O3 concentrations (mean 1-h max O3 of 41.6 ppb)
(Stieb et al., 1996). However, several other studies, mostly using all-year data, reported no
association between O3 and ED visits for respiratory causes. In general, ambient O3
concentrations were associated with ED visits for asthma in warm-season only analyses.
Relationships between ambient O3 and respiratory-related hospital admissions have also
been evaluated in various epidemiologic studies carried out in the United States and abroad.
Results from studies in the U.S. and Canada are summarized in Appendix Table 8A-4. Some of
these studies evaluated potential confounding by season or copollutants of the O3 effects on
respiratory hospitalizations. A multicity study by Burnett et al. (1997a) evaluated the
relationship between ambient O3 and respiratory hospitalizations in 16 Canadian cities over an
11-year period. Seasonal analyses indicated that O3 effects were only observed during the warm
season. As shown in Appendix Table 8A-4, most studies in the United States and Canada
indicated a consistent positive association between ambient O3 concentrations and respiratory
hospital admissions in the warm season, including studies with 98th percentile 8-h max O3 levels
as low as about 50 ppb. Analyses of all-year data were mostly positive, but nonsignificant for
either total respiratory or asthma hospital admissions. The studies that found positive
associations based on analyses using various lag periods found the association to be strong
with short lag periods (0 to 2 day), suggesting likely short-term O3 effects on asthma hospital
admissions.
Some new epidemiologic studies have also evaluated the associations between ambient O3
concentrations and cardiovascular-related hospitalizations (see Section 7.3.4 for details). The
epidemiologic evidence for cardiovascular morbidity is much more mixed than for respiratory
morbidity, with only one of several U.S./Canadian studies showing statistically significant
positive associations of cardiovascular hospitalizations with warm-season O3 concentrations.
Studies from Taiwan and Hong Kong also provided suggestive evidence of an association, but
most of the available European and Australian studies (all of which conducted all-year O3
analyses) did not find an association between short-term O3 concentrations and cardiovascular
hospitalizations.
8-37
-------
Acute Effects of Ozone on Mortality
The 1996 O3 AQCD noted that an association between O3 concentrations and daily
mortality for areas with high O3 levels (e.g., Los Angeles) was suggested by a few studies, but
the magnitude of such an effect was unclear. However, due to the limited number of available
studies and uncertainties regarding weather model specifications, no conclusive evidence of O3-
mortality associations was found in the 1996 O3 AQCD. In contrast, newly available large
multicity studies designed specifically to examine the effect of O3 on mortality have since
provided much more credible information. Numerous recent air pollution epidemiologic studies
have been published that examined acute O3 exposure effects on all-cause mortality in the U.S.
and elsewhere. Studies from the U.S. and Canada are presented in Appendix Table 8A-5.
Five new multicity studies, three conducted in the United States and two in Europe, found
small, but very precise (extremely narrow 95% CIs) positive associations for increased mortality
risk using all-year ambient O3 data and warm-season only data. Most notably, the largest of the
multicity studies conducted in the U.S. that evaluated the association between ambient O3
concentrations and excess mortality risk, the NMMAPS 95 U.S. communities study by Bell et al.
(2004), observed a significant positive association between O3 and all-cause mortality using all
available data (55 communities with all-year data and 45 communities with warm-season only
data). When the data were restricted to the warm season only, a slightly smaller, but still
significant, O3 effect was observed. Analyses suggested that the O3-related excess mortality risk
estimate was not confounded by exposure to PM. Another multicity study by Schwartz (2005)
examining the association between O3 and all-cause mortality observed a slightly greater excess
risk with warm-season only data compared to all-year data using a case-crossover study design.
Mortality risk estimates from single-city studies were also positive, but many were not
statistically significant. However, as in the case of ED visits and hospitalizations, restricting
data to the warm season resulted in mortality risk estimates that were more positive and stronger.
Meta-analyses of single-city mortality studies (Bell et al., 2005; Ito et al., 2005; Levy et al.,
2005) observed that O3-mortality risk estimates from warm-season only analyses were
approximately 2-fold greater than those from all-year analyses. As shown in Appendix Table
8A-5, the recent multicity and single-city studies generally show consistent positive and
significant associations between acute O3 exposure and all-cause mortality in studies with 98th
percentile 8-h max O3 values of 80 to 85 ppb and above.
8-38
-------
Some studies examined specific subcategories of mortality; however, most of these studies
had limited statistical power to detect associations due to the small daily counts for the specific
causes of death. In studies using broad categories of death, such as cardiopulmonary mortality,
significant associations with acute O3 exposure have been observed. For example, an analysis of
the NMMAPS data by Huang et al. (2005) examined the association between ambient O3
concentrations and cardiopulmonary mortality in 19 large U.S. cities during the warm season.
Ozone was positively associated with excess risk in cardiopulmonary mortality, even after
adjusting for PM and heat waves. Also, the meta-analysis by Bell et al. (2005) examined the
excess risk estimates for cardiovascular and respiratory causes and observed a slightly greater
risk of cardiovascular mortality compared to all-cause and respiratory disease mortality. It is
notable that only one single-city study, from among approximately 30 studies (including
2 multicity ones), did not find a positive association between ambient O3 concentrations and
increased cause-specific cardiovascular mortality, with both the multicity and several single-city
associations being statistically significant. It should be noted that these analyses of cause-
specific mortality do not provide direct evidence on possible causal mechanisms as they often do
not incorporate information on the contributing causes of death. For example, if an individual
has been suffering from a major cardiovascular disease, his/her death may be misclassified as
cardiovascular, even if a respiratory condition causes the death.
8.4.4.2 Chronic Ozone Exposure Studies
There were only a limited number of studies reported in the 1996 O3 AQCD that addressed
potential health effects of long-term ambient O3 exposures. The few available epidemiological
studies that had attempted to associate chronic health effects in humans with long-term O3
exposure provided very limited suggestive evidence that such a linkage may exist. Several
longitudinal epidemiologic studies carried out during the past decade have further evaluated the
potential effects of chronic (several weeks to many years) O3 exposure on lung function,
respiratory symptoms, lung inflammation, asthma prevalence, and birth defects. The strongest
evidence is for seasonal effects of extended O3 exposures on lung function in children, i.e.,
reduced lung function growth being associated with higher ambient O3 levels. Based on the
available data at this time, however, no clear conclusions can be drawn regarding the
relationship between chronic O3 exposure and the other health outcomes.
8-39
-------
Few studies have evaluated effects of long-term O3 exposure on the incidence of cancer
and/or mortality. Uncertainties regarding possible exposure periods of relevance and, also,
inconsistencies across mortality outcomes and gender raise concerns regarding plausibility.
The largest and most representative U.S. study, by Pope et al. (2002), observed positive but
nonsignificant associations between long-term O3 concentrations and all-cause,
cardiopulmonary, and lung cancer mortality. Thus, there is currently little evidence for potential
relationships between chronic O3 exposure and increased incidence of cancer and mortality.
8.4.4.3 Summary of the Epidemiologic Evidence
Assessment of the evidence from epidemiologic studies for various health outcomes has
focused on several considerations that are important in forming judgments as to the likely causal
significance of the observed associations. As discussed in Section 7.1.2, these include the
strength of the epidemiologic evidence, including the magnitude and precision of reported O3
effect estimates and their statistical significance, and the robustness of the effects associations,
or stability in the effect estimates after considering a number of factors, including alternative
models and model specifications, potential confounding by copollutants, as well as issues related
to the consequences of measurement error. Consistency of the effects associations involves
looking across the results of multiple- and single-city studies conducted by different
investigators in different places and times.
In general, when associations are strong in terms of yielding large relative risk estimates,
it is less likely that the association could be completely accounted for by a potential confounder
or some other source of bias. With associations that yield small relative risk estimates, it is
especially important to consider potential confounding and other factors in assessing causality.
Across the range of different health outcomes, effect estimates for many health associations
reported with O3 are generally small in size and could thus be characterized as weak. For
example, effect estimates for associations with mortality generally range from 0.5 to 5%
increases per 40 ppb increase in 1-h max O3 or equivalent, whereas associations for
hospitalization range up to about 50% increases per standardized O3 increment. Of particular
note are several multicity studies that have yielded relative risk estimates for associations
between short-term O3 exposure and mortality or morbidity that, although small in size, have
great precision due to the statistical power of the studies. For example, corresponding summary
8-40
-------
estimates for mortality in large U.S. multicity studies ranged from 0.5 to 1.0% per standardized
24-h increment, with some notable heterogeneity across cities. Thus, such associations are
strong relative to the precision of the studies; that is, the associations were strong enough to have
been reliably measured by the studies such that many of the associations can be distinguished
from the null hypothesis with statistical confidence (e.g., at p < 0.05).
As discussed in Section 7.6.4.2, the associations reported between short-term O3 exposure
and various health outcomes were generally robust to adjustment for copollutants, including
PM2 5 and sulfates, which are correlated with O3 concentrations in many areas in the United
States. More limited evidence is available on the effects of different modeling strategies on
associations with O3; however, these studies indicate that O3 associations may be robust to
various model specifications for temporal trend adjustment. Risk estimates for O3 were
generally more sensitive to the use of alternative weather models than to adjustment for temporal
trends, due to the temperature-dependent nature of O3 formation in the atmosphere.
In considering results from the multicity studies and single-city studies in different
geographic areas, there was general consistency in effects of short-term O3 exposure on
mortality, respiratory hospitalization and other respiratory health outcomes. Some variation in
effects was observed and that may be attributable to differences in relative personal exposure to
O3, which is affected by factors such as air conditioning prevalence and activity patterns, as well
as varying concentrations and composition of copollutants present in different regions. Thus,
consideration of consistency or heterogeneity of effects is appropriately understood as an
evaluation of the similarity or general concordance of results, rather than an expectation of
finding quantitative results within a very narrow range.
Taken together, the epidemiologic evidence shows strength, robustness, and consistency in
associations between short-term O3 exposure and a range of respiratory morbidity health
outcomes, including increased respiratory-related hospital admissions and asthma-related ED
visits. There is also relatively strong epidemiologic evidence for associations between short-
term O3 exposure and all-cause mortality. There is less strong, but nevertheless highly
suggestive evidence for effects related to cardiovascular morbidity. Moreover, consistently
positive associations were found for O3-related cardiovascular mortality across approximately
30 studies, with two well-conducted multicity studies in the United States and Europe yielding
small but statistically significant positive associations. In contrast, generally inconclusive
8-41
-------
evidence currently exists for associations between long-term chronic O3 exposure and either
morbidity or mortality.
8.5 ASSESSMENT OF POTENTIAL THRESHOLDS
A key issue in assessment of air pollution health effects is whether any thresholds can be
discerned for various types of health effects. Animal toxicology studies reviewed in Chapter 5
appear to indicate a linear concentration-response relationship for O3 exposure effects. Lowest
concentrations found to have caused statistically significant health effect changes, e.g., altered
pulmonary function, airway hyperresponsiveness, lung inflammation, increased lung tissue
permeability, or altered host defense, in various laboratory animal species are listed in Appendix
Table 8A-6. Some of these concentrations are rather high, but others appear to be more
environmentally relevant, including at least one study where the induction of airway
hyperresponsiveness was seen at O3 levels distinctly below current ambient O3 standards.
One recent controlled human exposure study, reviewed in Chapter 6, included exposures to
low, clearly environmentally-relevant, O3 concentrations for a group of healthy young adults.
Adams (2006) reported that following a 6.6 h exposure to 0.08 ppm O3, group mean O3-induced
FEVj decreases (-6.1%, square-wave; -7.0%, triangular) and symptom responses in healthy
adults were statistically significantly greater than after 0.04 and 0.06 ppm O3 exposures. During
a 6.6 h exposure to 0.06 ppm O3, group mean FEVj responses diverged from responses for
filtered-air and 0.04 ppm O3 by 5.6 h, but did not reach statistical significance at p < 0.05 by the
end of the exposure period (Adams, 2006). Exposures to 0.04 ppm O3 for 6.6 h produced group
mean FEVj responses quantitatively similar to those observed for filtered-air exposures (Adams,
2002, 2006). These results suggest that acute exposure to O3 can have effects in healthy young
adults with exposure levels below 0.08 ppm O3 (6.6 hr average) although no statistically
significant effects were observed at the lowest exposure level of 0.04 ppm O3. It is important to
note, however, that there is considerable interindividual variability in responses, as was
illustrated in Figures 8-IB and 8-2, and some evidence of FEVj decrements >10% for some
subjects at 0.06 ppm. It is also important to note that this type of study generally includes
healthy individuals and more subtle health measures, in contrast to population-based
epidemiologic studies.
8-42
-------
The potential for the existence of threshold levels has been explored in several recent
epidemiologic studies, as described in more detail in Section 7.6.5. A U.S. multicity panel study
by Mortimer et al. (2002) examined associations of ambient O3 concentrations with changes
in lung function and respiratory symptoms in asthmatic children. The mean 8-h avg O3
(10 a.m.-6 p.m.) was 48 ppb, with less than 5% of study days exceeding 80 ppb. Analyses
indicated that ambient O3 was significantly associated with decrements in PEF, both when using
all data available and after restricting data to days when 8-h avg O3 concentrations were
<80 ppb. A slightly larger effect estimate was observed using the restricted data set. Ozone-
related increased incidence of respiratory symptoms also persisted after eliminating data when
8-h avg O3 exceeded 80 ppb.
Examination of the shape of the concentration-response function in several time-series
studies of O3-related ED visits and hospitalizations has provided some indications of an effect
threshold. In a study of ED visits for asthma in St. John, Canada, the O3 effect observed in the
>15 years age group was apparent only when 1-h max O3 data above the 95th percentile value of
75 ppb were included (Stieb et al., 1996). In another study conducted in Toronto, Canada, the
association between ambient O3 and all-age respiratory hospital admissions only became
apparent above a daily 1-h max O3 level of approximately 30 ppb (Burnett et al., 1997b). In
London, England, possible thresholds for O3 effects on respiratory hospitalizations were
observed at 40 to 50 ppb for 8-h max O3 and 50 to 60 ppb for 1-h max O3 (Ponce de Leon et al.,
1996). However, other studies have reported a monotonic increase in the concentration-response
function throughout the range of ambient O3 concentrations, suggesting a lack of any evident
threshold for O3-related effects on respiratory hospitalizations and asthma ED visits (Burnett
et al., 1997a; Jaffe et al., 2003; Petroeschevsky et al., 2001; Tenias et al., 1998).
In the 95 U.S. communities study, Bell et al. (2004) reported that the risk estimate for
excess mortality associated with short-term O3 exposure was slightly smaller but remained
statistically significant when only using data from days with 24-h avg O3 concentrations below
60 ppb. A more formal threshold analyses recently reported by Bell et al. (2006) for 98 U.S.
communities, including the same 95 communities in Bell et al. (2004), indicated that if a
population threshold existed for mortality, it would likely fall below a 24-h avg O3 concentration
of 15 ppb. Other analyses by Kim et al. (2004) investigating the presence of a threshold in O3-
mortality effects in Seoul, Korea estimated threshold values of 28 ppb in 1-h max O3 when using
8-43
-------
all-year data and 45 ppb for summer-only data. None of the other pollutants examined,
including PM10, SO2, NO2, and CO, had a nonlinear association with mortality. Results from the
Kim et al. study suggested that if a threshold truly exists, the use of log-linear models may
underestimate the O3 effect on mortality at levels above the threshold.
Taken together, the available evidence from toxicologic, clinical and epidemiologic studies
suggests that no clear conclusion can now be reached regarding possible threshold levels for O3-
induced effects. The controlled human exposure studies demonstrate notable variability in
responsiveness among healthy subjects, even at low (<0.08 ppm) O3 exposure levels, making it
infeasible to suggest any clear threshold for the heterogeneous general population from these
results. As discussed in Section 7.6.5, there are limitations in epidemiologic studies that make
discerning thresholds in populations difficult, including low data density in the lower
concentration ranges, the possible influence of measurement error, and interindividual
differences in susceptibility to O3-related effects in populations. Nevertheless, the limited
clinical and epidemiologic evidence suggests that if a population threshold level does exist, it is
likely near the lower limit of ambient O3 concentrations in the United States.
8.6 BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE
FOR OZONE-RELATED HEALTH EFFECTS
This section integrates findings from epidemiologic studies with toxicologic and
mechanistic information obtained from controlled human exposure studies and animal
toxicology studies for major health endpoints reported to be associated with either short- or long-
term exposure to ambient O3. The section focuses on evidence related to two key considerations
for drawing conclusions about causality - biological plausibility and coherence of the evidence.
Evaluation of the biological plausibility of the O3-health effects associations observed in
epidemiologic studies reflects consideration of both exposure-related factors and
dosimetric/toxicologic evidence relevant to identification of potential biological mechanisms.
Similarly, coherence of health effects associations reported in the epidemiologic literature
reflects consideration of information pertaining to the nature of the various respiratory- and
cardiovascular-related morbidity and mortality effects and physiological endpoints evaluated in
human and animal toxicologic studies. The discussion in each subsection summarizes pertinent
8-44
-------
key information and then discusses plausibility based on toxicological studies of effects
attributed to O3 exposures in epidemiologic studies.
For this assessment, the ensuing discussion on biological plausibility and coherence
considers (a) the extent to which available epidemiological evidence logically ties together a
range of relevant health endpoints (from cardiopulmonary physiological changes to morbidity to
mortality) and (b) the extent to which available toxicological and biochemical evidence supports
plausible causal relationships for observed epidemiological associations for specific types of
health outcomes, i.e., how well do key epidemiologic findings compare with reasonably
hypothesized or experimentally demonstrated biological mechanisms of action.
8.6.1 Acute Ozone Exposure-Induced Health Effects
Respiratory Health Effects
As noted in Section 8.4.3, several new epidemiologic (field/panel) studies show positive
associations between short-term exposure to ambient O3 and human respiratory effects. These
health effects, as evaluated, include reductions in lung function, increased use of asthma
medication, and increased hospitalization, especially among individuals with asthma or certain
known cardiopulmonary diseases (see Chapter 7). The patterns of experimentally demonstrated
physiological and biochemical alterations noted earlier (see Table 8-1 and Figures 8-3 and 8-4)
support certain hypotheses regarding underlying pathological mechanisms in the development of
respiratory effects reported in the epidemiologic studies. Some of these mechanisms (see Table
8-1) include (a) decrements in lung function (capacities and volume), (b) bronchoconstriction,
(c) increased airway responsiveness, (d) airway inflammation, (e) epithelial injury, (f) immune
system activation, (g) host defense and (h) individual sensitivity factors such as age, genetic
susceptibility and the extent of tolerance resulting from previous exposures. The time sequence,
magnitude, and overlap of these complex events, both in terms of development and recovery (see
Figures 8-3 and 8-4), indicate the difficulties associated with the interpretation of biological
plausibility associated with the cardiopulmonary health effects.
Controlled human exposure studies have clearly demonstrated the following three types of
respiratory responses to acute O3 exposures: (1) irritative cough and substernal chest pain upon
inspiration; (2) decrements in FVC and FEVj due to decreased inspiratory capacity rather than
airways obstruction and (3) neutrophilic inflammation of the respiratory tract. Increased
8-45
-------
sensitivity (susceptibility) to these effects was observed even among a carefully selected
homogeneous study population. The sources of this heterogeneity are not fully understood.
As discussed in the earlier section, changes in baseline levels of various responses, the lag in the
recovery phase and the role of residual defects in these mechanisms in hyperresponsive
individuals suggest the potential for increased health effects in compromised individuals with
preexisting cardiopulmonary diseases. Recent research has emphasized further characterization
of the mechanisms and consequences of O3-induced pulmonary function and inflammatory
responses. In addition, animal studies indicate morphological changes associated with acute O3
exposures.
Ozone-induced altered breathing patterns (rapid shallow breathing) observed in controlled
human exposure studies and animals occur without significantly affecting minute ventilation,
suggesting compensatory changes in breathing pattern. Such a shift in breathing pattern
diminishes deep lung penetration of O3. Breathing pattern is modulated by changes in peripheral
mechanisms, such as direct or indirect stimulation of lung receptors and bronchial C-fibers. The
activity of these afferents is integrated with input from sensory pathways and thus determines the
depth and frequency of breathing. Stimulation of bronchial C-fibers along with inhibition of
inspiration through local axon reflexes can induce neurogenic inflammation via tachykinins and
other proinflammatory neuropetides. Ozone-induced increases in the levels of the neuropeptide
substance P observed in the BALF of human subjects suggests potential neurogenic involvement
in O3-induced increased vascular permeability, plasma protein extravasation, mucus secretion,
and bronchoconstriction (Solway and Leff, 1991). Similar neurogenic involvement due to
vagally mediated stimulation of C-fibers seen in animal toxicology studies also supports O3-
induced airway hyperresponsiveness observed in humans.
An extensive database of animal, human, and in vitro studies supports the conclusion
that O3 interacts with airway epithelial cell membranes and lining fluid to form lipid ozonation
products and reactive oxygen species (ROS). These reactive products initiate a cascade of
events leading to oxidative stress, injury, inflammation, airway epithelial damage and increased
alveolar permeability to vascular fluids. Inflammation is the outcome of host response to injury
and usually resolves completely. Continued irritant challenge may evolve into a chronic
inflammatory state with simultaneous alterations in lung structure and function, leading to
diseases such as fibrosis and emphysema, although neither fibrosis nor emphysema have yet
8-46
-------
been observed with O3 exposure. Continued inflammation can also alter the lung's ability to
respond to infectious agents, allergens, and toxins. Acute inflammatory responses to O3
exposure have been well documented experimentally in both humans and animals. As presented
in Figure 8-4, the early inflammatory response to O3-induced lung injury is apparent in human
subjects within 3 h postexposure. This initial neutrophilic inflammatory response phase is
characterized by increases in PMNs in BALF along with increased levels of inflammatory
mediators such as interleukins, prostaglandins and complement component C3a. In vitro studies
using human and animal lung cell culture systems have further examined the involvement of
various inflammatory mediators and in some instances their downstream signaling pathways.
The late inflammatory phase in the lung is characterized by increased levels of monocytes
and eosinophils, as well as respective mediators such as cytokines, leukotrines, proteinases,
and ROS.
Disruption of the lung's blood barrier by O3 results in vascular permeability changes
and plasma protein extravasation. Analysis of BALF plasma influx markers such as albumin,
other proteins, immunoglobulins, and epithelial cell damage markers such as LDH indicate
O3-induced lung epithelial injury. Ozone-induced lung injury and subsequent disruption of the
airway epithelial barrier has been implicated in altered mucociliary clearance of particles
observed in controlled human studies. Analogously, animal toxicology studies (see Chapter 5)
have reported increased mortality to bacterial and viral infections subsequent to O3 exposure and
also increased clearance of particles with low O3 exposure levels.
Controlled O3 exposure studies of healthy humans have indicated a large degree of
intersubject variability. The spirometric and symptomatic responses are highly reproducible
within a given subject; but, within a group, pulmonary function can vary widely (see Figure 8-1)
across different subjects. This interindividual variability is likely to be due to factors such as
age, genetic background and antioxidant defenses, but as yet other factors remain to be
characterized. Epidemiologic studies indicate a positive association between exposure to
ambient O3 and declines in lung function in children and those with preexisting respiratory
diseases such as asthma. Increased incidence of ED visits and hospitalizations due to respiratory
causes that have been reported in various epidemiologic studies (discussed in Chapter 7), as also
supported by animal toxicology data, suggest a causal association with O3. The epidemiologic
associations become more apparent when the data are analyzed for the influence of seasonal
8-47
-------
differences in ambient O3 levels, with stronger evidence for associations in the warm season.
Together, the evidence from the animal and human studies suggests that acute O3 exposure is
causally associated with respiratory system effects, including O3-induced pulmonary function
decrements, respiratory symptoms, lung inflammation, and increased lung permeability, airway
hyperresponsiveness, increased uptake of nonviable and viable particles, and consequent
increased susceptibility to PM-related toxic effects and respiratory infections.
Cardiovascular Health Effects
Only a few experimental studies of animals and humans have evaluated possible
mechanisms or physiological pathways by which acute O3 exposures may induce cardiovascular
effects. Ozone induces lung injury, inflammation, and impaired mucociliary clearance, with a
host of associated biochemical changes all leading to increased lung epithelial permeability.
As discussed in Section 5.2.1, the generation of lipid ozonation products and reactive oxygen
species in lung tissue can influence pulmonary hemodynamics and, ultimately, the
cardiovascular system.
Recent in vitro studies of O3 reactions with cholesterol in lung surfactant found consequent
generation of highly reactive products such as oxysterols and p-epoxide in BALF isolated from
rats exposed to 2.0 ppm O3 for 4 h (Pulfer and Murphy, 2004). Additionally, both 5P,6P-
epoxycholesterol and its most abundant metabolite, cholestan-6-oxo-3p,5a-diol, were shown to
be cytotoxic to human lung epithelial (16-HBE) cells and to inhibit cholesterol synthesis.
Studies (Pulfer et al., 2005) of mice exposed to 0.5, 1.0, 2.0, or 3.0 ppm O3 for 3 h also
demonstrated that these oxysterols were produced in vivo. These results suggest that this may
be an additional mechanism of O3 toxicity, including a pathway by which O3 may play a possible
role in atherosclerosis and other cardiovascular effects.
The presence of oxysterols in human atherosclerotic lesions implicates the oxidation of
cholesterol in the pathogenesis of atherosclerosis, a well-known contributor to development of
cardiovascular disease. Oxysterols may arise from different cholesterol oxidation mechanisms,
(including free radical-mediated oxidations), and their unabated accumulation in macrophages
and smooth muscle cells of arterial walls lead to formation of fatty streaks in advanced lesions.
The presence of one of the O3-induced oxysterols, secosterol, in endogenously formed arterial
plaques (Wentworth et al., 2003) suggests that the oxysterols produced in the lung either due to
8-48
-------
direct O3 interaction with surfactant cholesterol or with oxidant radicals at the O3-induced
inflammation site may have potential involvement in the development of cardiovascular and
myocardial diseases. In addition, the recent in vitro observation (Sathishkumar et al. 2005) of
increased apoptosis (programmed cell death) induced by secosterol in H9c2 cardiomyocytes
(heart cells) supports possible involvement of such biologically active oxysterols in O3-induced
cardiovascular effects observed in the epidemiologic studies. Also, the detection of oxysterols in
the BALF of rats exposed to O3 suggests their potential to be used as biomarkers of O3 exposure.
Demonstration of relationships between oxysterols of the type generated in lung surfactant with
O3 exposure and cardiovascular disease outcomes in clinical settings or epidemiologic studies
would add considerable value to the experimental observations thus for reported in the animal
toxicology studies.
Ozone-induced changes in heart rate, edema of heart tissue, and increased tissue and serum
levels of atrial natriuretic factor (ANF) found with 8-h 0.5 ppm O3 exposure in animal
toxicology studies (Vesely et al., 1994a,b,c) also raise the possibility of potential cardiovascular
effects of acute O3 exposures. Such effects resulting from stimulation of airway irritant
receptors, c-fiber activation, may result from either local or central nervous system involvement.
Only one controlled human study (Gong et al., 1998) evaluated potential cardiovascular health
effects of O3 exposure and reported O3 -induced changes in alveolar-arterial oxygen transfer in
subjects with hypertension. However, several other cardiovascular parameters (e.g., heart rate)
evaluated in this study did not show any O3-induced effects.
Animal toxicology studies have found both transient and persistent ventilatory responses
with or without progressive decrease in heart rate (Arito et al., 1997). Observations of
O3-induced vasoconstriction in a controlled human exposure study by Brook et al. (2002)
suggests another possible mechanism for O3-related exacerbations of preexisting cardiovascular
disease.
A few new field/panel studies of human adults have reported associations between ambient
O3 concentrations and changes in electrophysiologic indicators of cardiac function, e.g., heart
rate variability (FtRV). Also, some population time-series studies have suggested positive
associations between acute O3 exposure and cardiovascular hospitalizations in the warm season.
These suggestive positive epidemiologic associations gain credibility and scientific support from
8-49
-------
results of experimental animal toxicology and human clinical studies, as discussed above, which
are indicative of plausible pathways by which O3 may exert cardiovascular effects.
8.6.2 Chronic O3 Exposure-Induced Health Effects
The effects of chronic O3 exposure in humans were earlier addressed primarily with cross-
sectional epidemiologic studies, as discussed in the 1996 O3 AQCD. Due to lack of precise
information on exposure, the possibility of selection bias, and the difficulty of controlling for
confounders, the earlier study findings were considered to be inconclusive. Since 1996, several
new cross-sectional epidemiological studies have evaluated potential associations between
chronic exposure to O3 and morbidity or mortality (see Chapter 7, Section 7.5). These studies
suggest that seasonal exposure to O3 may be related to changes in lung function in children.
However, little evidence is available to support a relationship between chronic O3 exposure and
mortality or lung cancer incidence. There are no data available from controlled human chamber
studies that evaluated chronic exposure regimens.
The lack of adequate data from epidemiologic and clinical studies in human has helped to
focus attention much more so on results from chronic O3 exposure studies in animals. Earlier
chronic animal studies employed traditional exposure designs using chronic stable O3 exposures
to one or another single O3 concentration over extended periods (weeks, months). More recent
studies have attempted to incorporate design features that more closely mimic diurnal and
seasonal patterns of O3 exposure and realistic exposure concentrations. Studies of monkeys that
compared these two designs reported greater airway pathology with the latter design. Persistent
and irreversible effects observed in chronic animal toxicology studies indicate the need for
further complementary human data from epidemiologic studies.
Animal toxicology data provide a clearer picture indicating that long-term O3 exposure
may have lasting effects. Chronic exposure studies in animals have reported biochemical and
morphological changes suggestive of irreversible long-term O3 impacts on the lung. Some of the
studies in rats (0.5-1.0 ppm O3 for 6 h/day) for 20 months and monkeys (0.61 ppm) for one year
noted increased deposition of collagen and thickening of the CAR. Differences in the degree of
this type of lung damage have been observed between continuous exposure and seasonal patterns
of variations in O3 exposure levels over time. A long-term study of infant rhesus monkeys
exposed to simulated seasonal O3 (0.5 ppm 8 h/day for 5 days every 14 days for 11 episodes)
8-50
-------
resulted in remodeling in the distal airways, abnormalities in tracheal basement membrane,
accumulation of eosinophils in conducting airways and decrements in airway innervation.
Earlier studies in rats following seasonal episodic profiles also showed small, but significant,
decrements in lung function that were consistent with focal fibrinogenesis in the proximal
alveolar region. On the other hand, chronic O3 exposures in a range of 0.5 to 1.0 ppm induce
epithelial hyperplasia that disappears in a few days. The weight of evidence from the new
experimental animal studies (using non-lifetime exposures) does not support ambient O3 as being
a pulmonary carcinogen.
Collectively, evidence from animal studies strongly suggests that chronic O3 exposure is
capable of damaging the distal airways and proximal alveoli, resulting in lung tissue remodeling
leading to apparent irreversible changes. Compromised pulmonary function and structural
changes due to persistent inflammation may exacerbate the progression and development of
chronic lung disease. These findings offer some insight into potential biological mechanisms for
the suggested association between seasonal O3 exposure and reduced lung function development
in children as observed in epidemiologic studies.
8.6.3 Mortality-Related Health Endpoints
An extensive analysis of population time-series studies that evaluated the air pollution
related mortality risk estimates presented in Section 7-4 utilized data from single and multicity
studies from around the world. Mortality risk estimates derived from single- and multicity
studies in U.S. and Canada coupled with meta-analyses generally indicate associations between
acute O3 exposure and elevated risk for all-cause mortality, even after adjustment for the
influence of season and PM. Several single-city studies that specifically evaluated the
relationship between O3 exposure and cardiopulmonary mortality also reported results suggestive
of a positive association.
The epidemiology results outlined above for mortality suggest a pattern of effects that may
be biologically germane to interpretation of its causality, but our knowledge about potential
underlying mechanisms remains relatively limited and suggests a need for further experimental
support. The majority of the physiological and biochemical parameters evaluated both in human
clinical and animal toxicology studies (Table 8-1; Figure 8-3) suggest a relatively transient
nature for O3-induced biochemical perturbations. Most effects attenuate over time, depending on
8-51
-------
the preexisting pathophysiology. However, one can hypothesize a generic pathway of O3-
induced lung damage, potentially involving oxidative lung damage with subsequent
inflammation and/or decline in lung function leading to respiratory distress in some sensitive
population groups (e.g., asthmatics) or, other plausible pathways noted below may lead to O3-
related contributions to cardiovascular effects that ultimately increase the risk of mortality.
Recent analysis of data from the third National Health and Nutrition Examination
Followup study indicated that about 20% of the adult population have reduced FEVj values
indicative of impaired lung function. The majority of these individuals have COPD, asthma or
fibrotic lung disease (Mannino et al., 2003). These cardiopulmonary disease conditions are
associated with persistent low-grade systemic inflammation. It has also been reported that
patients with COPD are at increased risk for cardiovascular disease. Lung disease with
underlying inflammation may also link to low-grade systemic inflammation associated with
atherosclerosis. These effects in disease are independent of cigarette smoking (Sin et al., 2005).
Lung function decrements in cardiopulmonary disease have also been associated with
inflammatory markers such as C-reactive protein (CRP) in blood. In fact, at the population
level, individuals with the lowest FEVj have the highest levels of CRP, while those with highest
FEVj have the lowest values for CRP (Mannino et al., 2003; Sin and Man, 2003). The complex,
physiological and biochemical perturbations that exist simultaneously (Figures 8-3 and 8-4)
subsequent to acute exposure to O3 may tilt the biological homeostasis mechanisms leading to
adverse health effects in people with compromised cardiopulmonary systems.
Of much interest are several other types of newly available experimental data that support
reasonable hypotheses that may help to explain findings of O3-related increases in cardiovascular
mortality observed in some epidemiological studies. These include the direct effects of O3 in
terms of its increasing platelet aggregating factor (PAF) in lung tissue that can then enter the
general circulation and possibly contribute to increased risk of blood clot formation (as
illustrated in Figure 5-3) and consequent increased risk of myocardial infarction (heart attack),
cerebrovascular events (stroke), and/or associated cardiovascular-related mortality. Other O3-
induced effects may also contribute to increased risk of cardiovascular impacts. Ozone reactions
with cholesterol in lung surfactant to produce epoxides and/or associated diol metabolites that
are cytotoxic to lung and heart muscle cells and that contribute to atherosclerotic plaque
formation in arterial walls represent another such pathway. Also, to the extent that O3-induced
8-52
-------
increases in lung permeability allow more ready entry of inhaled particulate matter (PM) into the
blood stream, then O3 exposure would enhance the risk for PM-related cardiovascular effects
(see U.S. Environmental Protection Agency, 2004). An associated factor to consider is O3's
ability to contribute to ultrafine PM formation in ambient air and indoor environments (as
discussed in Sections 8.2.4 and 8.3.2). Thus, O3 can both contribute to increased presence of
fine particles and human exposure to them, as well as enhance uptake of inhaled fine PM and
thereby presumably contribute to exacerbation of PM-induced cardiovascular effects in addition
to those more directly induced by O3 per se.
8.6.4 Health Effects of Ozone-Containing Pollutant Mixtures
The above-noted potential for O3-related enhancements of PM formation, particle uptake,
and exacerbation of PM-induced cardiovascular effects underscores the importance of
considering contributions of O3 interactions with other often co-occurring air pollutants to health
effects due to O3-containing pollutant mixes. Chapters 4, 5, and 6 provided a discussion of
experimental studies that evaluate interactions of O3 with other co-occurring pollutants. Some
examples of important pollutant mixture effects noted there are highlighted below.
First, Chapter 4 noted some important interactive effects of coexposures to O3 and NO2
and SO2, two other common gaseous copollutants found in ambient air mixes. That is, a study
by Rigas et al. (1997) showed that continuous exposure of healthy human adults to SO2 or
to NO2 increased inhaled bolus O3 absorption, while continuous exposure to O3 alone decreased
bolus absorption of O3. This suggests enhancement of O3 uptake by NO2 or SO2 coexposure in
ambient air mixes. Also, as noted in Chapter 6, another study by Jenkins et al. (1999) showed
that asthmatics exhibited enhanced airway responsiveness to house dust mite following
exposures to O3, NO2, and the combination of the two gases. Spirometric response, however,
was impaired only by O3 and O3+NO2 at higher concentrations. On the other hand, animal
toxicology studies discussed in Chapter 5 that evaluated exposures to O3 in mixture with NO2,
formaldehyde, and PM demonstrated additive, synergistic or antagonistic effects, depending on
the exposure regimen and the specific health endpoints evaluated.
The results of the Jenkins et al. (1999) study also help to illustrate a more general
phenomena of enhancement by O3 exposure of various respiratory responses of sensitive
individuals to allergens. Chapter 6 noted, for example, studies (a) by Peden et al. (1995)
8-53
-------
showing O3-induced increased response to nasal allergen challenge among allergic asthmatic
subjects, and (b) by Michelson et al. (1999) showing promotion by 0.4 ppm O3 exposure of
inflammatory cell influx in response to nasal allergen challenge in asymptomatic dust-mite
sensitive asthmatics. In addition, Torres et al. (1996) demonstrated enhancement by 0.25 ppm O3
exposure of airway responsiveness in mildly allergic asthmatics that was increased in response to
an individual's historical allergen (grass and birch pollen, house dust mite, animal dander).
These results were further extended by Holz et al. (2002) who showed that repeated daily
exposure to 0.125 ppm O3 for 4 days exacerbated lung function decrements (e.g., decreased
FEVj) in response to bronchial allergen challenges among subjects with preexisting allergic
airway disease, with or without asthma. This suggests that O3 exposure can place allergic
nonasthmatic persons, as well as asthmatics, at increased risk for allergic respiratory effects.
Consistent with and supporting the above findings are animal toxicology studies reviewed in
detail by Harkema and Wagner (2005), which indicate that (a) O3-induced epithelial and
inflammatory responses in laboratory rodents are markedly enhanced by coexposure to inhaled
biogenic substances (e.g., bacterial endotoxin or ovalbumin, an experimental aeroallergen) and
(b) adverse airway effects of biogenic substances can be exacerbated by coexposure to O3.
Also of much note is a newly emerging literature which indicates that O3 can modify the
biological potency of certain types of ambient PM, as shown by experimental tests. For
example, as described in Chapter 5 (Section 5.4.2) reaction of diesel PM with 0.1 ppm O3 for
48 h increased the potency (compared to non-exposed or air-exposed diesel PM) to induce
neutrophil inflex, total protein, and LDH in lung lavage fluid in response to intratracheal PM
instillation in rats (Madden et al., 2000). However, the potency of carbon black particles was not
enhanced by exposure to O3, suggesting that O3 reaction with organic components of the diesel
PM were responsible for the observed increased diesel PM effects.
Potential interaction of O3 with fine PM in aged rats was examined by Kleinman et al.
(2000). In this study the effects of fine PM containing two common toxic constituents,
ammonium bisulfate (ABS, 0.3 jim 70 jig/nT3) and elemental carbon (C, 0.3 |im 50 jig/nT3) and
a mixture (ABS + C) with 0.2 ppm O3 was evaluated on aged rat lung structure and macrophage
function. Exposures of O3, elemental carbon or ABS alone did not cause significant lung injury,
lung tissue collagen content or respiratory burst activity. On the other hand, mixtures (ABS +
C + O3) caused significant lung injury as assessed by increased cell proliferation response in
8-54
-------
lung epithelial and interstitial cells, loss of lung tissue collagen and increase in respiratory burst
and phagocytic activity.
The majority of toxicological studies discussed earlier in this chapter evaluated effects of
individual pollutants or simple mixtures of the constituents of urban smog mixtures, and these
toxicology studies do not fully explain epidemiologic findings that have increasingly shown
ambient O3, other gaseous pollutants, and/or PM to be associated with various health effects at
relatively low concentrations. In a recent report, Sexton et al (2004) utilized "smog chambers",
i.e., environmental irradiation chambers to generate synthetic photochemical oxidants mixtures
similar to urban smog, and studied the toxicity of such mixtures on the inflammatory response of
A549 cells in an in vitro exposure system. In this preliminary study, the authors found the
simulated urban photochemical oxidant mixture generated with the addition of O3 to have
enhanced toxicity (as assessed by the expression of IL-8 mRNA). Additional toxicology studies
using similar realistic air pollution smog mixtures in the future may provide more relevant
biological understanding for the potential interactions that occur in the ambient air among
various pollutants.
All of the above types of interactive effects of O3 with other co-occurring gaseous and
nongaseous viable and nonviable PM components of ambient air mixes argue for not only being
concerned about direct effects of O3 acting alone but also the need for viewing O3 as a surrogate
indicator for air pollution mixes which may enhance risk of adverse effects due to O3 acting in
combination with other pollutants. Viewed from this perspective, epidemiologic findings of
morbidity and mortality associations with ambient O3 concentrations extending to concentrations
below 0.08 ppm become more understandable and plausible.
8.7 SUSCEPTIBLE AND VULNERABLE POPULATIONS, AND
POTENTIAL PUBLIC HEALTH IMPACTS
Many factors such as age, gender, disease, nutritional status, smoking, and genetic
variability may contribute to the differential effects of environmental pollutants, including O3.
Genetic factors, such as single nucleotide polymorphisms (SNPs) and developmental defects,
can contribute to innate susceptibility, while acquired susceptibility may develop due to personal
habits (smoking, diet, exercise) and other risk factors such as age, gender, pregnancy, and
8-55
-------
copollutant exposures. In the 1996 O3 AQCD, children, outdoor workers, and people with
preexisting respiratory disease were identified as likely being more susceptible or vulnerable to
effects of ambient O3 exposure. However, the available toxicological and human data had not
shown that males and females respond differently to O3; nor were available data adequate to
suggest differences in responsiveness to O3 based on ethnic or racial background. Overall, then,
the available information from animal toxicology and epidemiologic studies provided only
relatively limited evidence by which to confidently identify likely susceptible groups and/or to
associate specific factors as contributing to increased risk of O3-related adverse health effects.
Advances in available research results since 1996 have improved our ability to delineate likely
susceptible or vulnerable populations at increased risk for O3-induced health effects and to
delineate factors contributing to such risk.
8.7.1 Preexisting Disease as a Potential Risk Factor
People with preexisting pulmonary disease are likely to be among those at increased risk
from O3 exposure. Altered physiological, morphological and biochemical states typical of
respiratory diseases like asthma, COPD and chronic bronchitis may render people sensitive to
additional oxidative burden induced by O3 exposure. Based on studies assessed in the 1996 O3
AQCD (U.S. Environmental Protection Agency, 1996a), asthmatics appear to be at least as, or
more, sensitive to acute O3 exposure as healthy nonasthmatic subjects. The new results reviewed
in Chapters 6 and 7 of this document from controlled exposure and epidemiologic studies also
indicate that asthmatics are a potentially sensitive subpopulation for O3 health effects.
A multicity study by Mortimer et al. (2002) support earlier observations that asthmatic
children are particularly susceptible to ambient O3. This association, based on decrements in
lung function and exacerbation of pulmonary disease symptoms, suggests that O3 exposures may
result in increased use of medication in asthmatic children. A number of time-series
epidemiologic studies have reported increased risk in study subsets of individuals with
preexisting lung diseases, especially asthma, as potentially susceptible individuals. The
epidemiologic studies of acute exposure to O3 discussed in Section 8.4.2 indicate increased risk
for exacerbation of respiratory disease symptoms during the warm season.
Several clinical studies reviewed in the 1996 O3 AQCD on atopic and asthmatic subjects
had suggested but not clearly demonstrated enhanced responsiveness to acute O3 exposure
8-56
-------
compared to healthy subjects. The majority of the newer studies reviewed in Chapter 6 indicate
that asthmatics are as sensitive as, if not more sensitive than, normal subjects in manifesting O3-
induced pulmonary function decrements.
Ozone-induced increases in neutrophils, protein, and IL-8 were found to be significantly
higher in the BALF from asthmatics compared to healthy subjects, suggesting mechanisms for
the increased sensitivity of asthmatics. Similarly, subjects with allergic asthma exhibited
increased airway responsiveness to inhaled allergens upon acute O3 exposure. Consistent with
these changes, it is suggested that asthmatics will be more sensitive to small airway effects of
ambient O3. Asthmatics present a differential response profile for cellular, molecular, and
biochemical parameters (Figure 8-1) that are altered in response to acute O3 exposure. Increases
in O3-induced nonspecific airway responsiveness incidence and duration could have important
clinical implications for asthmatics.
Bronchial constriction following provocation with allergens presents a two-phase response.
The early response is mediated by release of histamine and leukotrienes that leads to contraction
of smooth muscle cells in the bronchi, narrowing the lumen and decreasing the airflow.
In asthmatics, these mediators also cause accumulation of eosinophils, followed by production
of mucus and a late-phase bronchial constriction and reduced airflow. Holz et al. (2002)
reported an early phase response in subjects with rhinitis after a consecutive 4-day exposure to
0.125 ppm O3 that resulted in a clinically relevant (>20%) decrease in FEVj. Allergen challenge
in mild asthmatics 24 h postexposure to 0.27 ppm O3 for 2 h resulted in significantly increased
eosinophil counts in BALF compared to healthy subjects (Vagaggini et al., 2002). Epithelial
cells from mucosal biopsies of allergic asthmatics indicated significant increases in the
expression of IL-5, IL-8 and GM-CSF, suggesting increased neutrophilic inflammation
compared to healthy subjects (Bosson et al., 2003).
Several human exposure studies have shown differences between asthmatics and healthy
human subjects with regard to PMN influx in BALF . In vitro studies (Schierhorn et al., 1999)
of nasal mucosal biopsies from atopic and nonatopic subjects exposed to 0.1 ppm O3 found
significant differences in release of IL-4, IL-6, IL-8, and TNF-a. Another study by Schierhorn
et al. (2002) found significant differences in the O3-induced release of the neuropeptides
neurokinin A and substance P for allergic patients in comparison to nonallergic controls,
suggesting increased activation of sensory nerves by O3 in the allergic tissues. Another study by
8-57
-------
Bayram et al. (2002) using in vitro culture of bronchial epithelial cells recovered from atopic and
nonatopic asthmatics also found significant increases in epithelial permeability in response to O3
exposure. In addition, some controlled human O3 exposure studies in asthmatics (Hiltermann
et al., 1999; Scannell et al., 1996) reported increased secretion of IL-8, suggesting increased
neutrophilic inflammation. Two studies (Torres et al., 1996; Holz et al., 2002) observed
increased airway responsiveness to repeated daily O3 exposure to bronchial allergen challenge in
subjects with preexisting allergic airway disease.
Newly available reports from controlled human exposure studies (see Chapter 6) utilized
subjects with preexisting cardiopulmonary diseases such as COPD, asthma, allergic rhinitis, and
hypertension. The data generated from these studies that evaluated pulmonary function changes
in spirometry did not find clear differences between filtered air and O3 exposure in COPD and
asthmatic subjects. However, the new data on airway responsiveness, inflammation, and various
molecular markers of inflammation and bronchoconstriction indicate that people with atopic
asthma and allergic rhinitis comprise susceptible groups for O3-induced adverse health effects.
Collectively, these observations suggest that O3 exposure exacerbates pre-existing allergic
asthma. People with allergic asthma likely represent a sizable segment of the population
reported to have increased symptoms of respiratory illness exacerbations, ED visits, and hospital
admissions in epidemiologic studies.
Although controlled human exposure studies have not found evidence of larger spirometric
changes in people with COPD relative to healthy subjects, this may be due to the fact that most
people with COPD are older adults who would not be expected to have such changes based on
their age. However, new epidemiological evidence indicates that people with COPD may be
more likely to experience other effects, including emergency room visits, hospital admissions, or
premature mortality. A study by Tenias et al. (2002) observed a significant positive association
between O3 and for COPD. Results from an analysis of five European cities indicated strong and
consistent O3 effects on unscheduled respiratory hospital admissions, including COPD
(Anderson et al., 1997). Also, an analysis of a 9-year data set for the whole population of the
Netherlands provided risk estimates for more specific causes of mortality, including COPD
(Hoek et al., 2000, 2001; reanalysis Hoek, 2003); a positive, but nonsignificant excess risk of
COPD-related mortality was found to be associated with short-term O3 concentrations.
8-58
-------
8.7.2 Age-Related Variations in Susceptibility/Vulnerability
Growth and development of the human respiratory system is not complete until 18 to
20 years of age; 80% of alveoli are formed postnatally, and changes in the lung continue through
adolescence (Dietert et al., 2000). Because the developing lung is highly sensitive to damage
from exposure to environmental toxicants, children and infants are likely to be among the most
susceptible to detrimental effects from O3 exposure. In addition, children experience enhanced
O3 vulnerability because they generally spend more time outdoors and are highly active,
especially during the warm season when ambient O3 concentrations are expected to be high and
results in increased O3 dose delivered to their lungs (Plunkett et al., 1992; Wiley et al., 1991a,b).
Controlled human exposure studies assessed in the 1996 O3 AQCD indicated that children
(generally aged 8 to 11 years) and adolescents (generally aged 12 to 18 years) exhibited, on
average, greater spirometric responses to O3 than middle-aged or older adults when exposed to
comparable O3 doses. However, less evidence of respiratory symptoms has been observed
among healthy children in clinical studies of O3 exposure. Such diminished symptomatic
responses in children may put them at an increased risk for continued O3 exposure. As reported
in the 1996 O3 AQCD, several summer camp panel studies collectively provided strong evidence
for an association between acute O3 exposure and lung function declines in mostly healthy,
nonasthmatic children (aged 7 to 17 years) who spent long hours outdoors. Additional
epidemiologic field studies observed that respiratory symptoms (or exacerbation of asthma) and
decrements in PEF were associated with increased ambient O3 concentrations, with greater
responses found in asthmatic than in nonasthmatic children. No new human exposure studies
investigating O3 responses in children have been published since the last O3 AQCD; however,
the epidemiologic studies published during the last decade (see Section 7.2.3.1 for details)
continue to indicate that children are at increased risk for to O3-related respiratory health effects.
Many recent field studies published in the past decade have focused on the effects of
acute O3 exposure on the respiratory health of children. In general, children experienced
O3-related decrements in pulmonary function parameters, including PEF, FEVj, and FVC (e.g.,
Mortimer et al., 2002; Thurston et al., 1997). Declines in lung function were observed in both
healthy and asthmatic children following acute exposure to O3, but respiratory symptoms were
largely found only in asthmatic children. The O3-related changes in lung function in children
were often small, <5% change per 40 ppb increase in mean 1-h max O3 or equivalent. However,
8-59
-------
as stated earlier, the group mean responses reported may mask larger lung function changes in
sensitive individuals. Hoppe et al. (2003) examined several susceptible populations, including
children (aged 6 to 8 years) and juvenile asthmatics (aged 12 to 23 years), for changes in
pulmonary function attributable to O3 exposure in Munich, Germany. Using group mean values,
consistent O3 effects were not observed in both groups; but, when data was analyzed on an
individual basis, a potential pattern of increased O3 sensitivity was observed. About 20% of the
children and asthmatics were found to be particularly susceptible to O3 health effects (i.e.,
experienced greater than 10% change in FEVj).
Several epidemiologic field studies have examined effects of long-term O3 exposure on
lung function in children. The results collectively indicate that seasonal O3 exposure is
associated with smaller increases in lung function in children (mean age 7 to 8 years). These
results are supported by a long-term study of infant rhesus monkeys exposed to simulated
seasonal O3 patterns (0.5 ppm 8h/day for 5 days, every 14 days for 11 episodes) which found
remodeling in the distal conducting airways as a result of damage and repair processes occurring
with repeated O3 exposure (Evans et al., 2003; Schelegle et al., 2003). However, epidemiologic
evidence of an effect of longer-term (i.e., multiyear) O3 exposure on lung function development
in children has generally not yet been found, as most definitively indicated by the southern
California Children's Health Study that examined children from 10 to 18 years of age
(Gauderman et al., 2004).
In time-series studies that examined associations between ambient O3 concentrations and
ED visits, hospitalizations, and mortality in children, consistent results were not observed. This
is likely due to the small number of daily deaths and hospitalizations among children and, also,
the fact that the large number of competing causes of mortality and hospitalizations in children
(which are together more important than air pollution) make it difficult to detect any excess risk
attributable to ambient O3 exposure.
The elderly are also often classified as being particularly susceptible to air pollution. The
basis for increased O3 sensitivity among older adults is not known, but one hypothesis is that it
may be related to changes in the respiratory tract lining fluid antioxidant defense network (Kelly
et al., 2003). Increased susceptibility of older adults to O3 health effects is most clearly indicated
in the newer mortality studies. In the meta-analysis by Bell et al. (2005), a comparison of O3-
mortality risk estimates by age indicates that, in general, the older adults population (>65 years
8-60
-------
of age) is more susceptible to O3 effects, with older adults experiencing a more than 50% greater
excess risk in mortality compared to the all-age group. In the large U.S. 95 communities study
(Bell et al., 2004), the O3 effect on all-cause mortality was slightly larger for older adults
compared to younger individuals (<65 years of age). Bell et al. (2004) further noted that though
the risk estimates for older adults were only slightly higher, the absolute effect of O3 is
substantially greater due to the higher underlying mortality rates in the older adult population
(e.g., approximately 50% of deaths were in those >75 years), which leads to a larger number of
extra deaths for older adults compared to the general population.
Lung function responses to O3 exposure, however, have been found to be diminished in the
older adult population. The study by Hoppe et al. (2003) observed that ambient O3 exposure was
not associated with lung function declines in seniors (69 to 95 years of age). While a greater
than 10% change in FEVj was observed in 20% of the children and juvenile asthmatics, only 5%
of the older adult population experienced such a change. The results by Hoppe et al. are
consistent with the diminishing lung function responses to O3 exposure with increasing age
observed in clinical studies.
In summary, new epidemiologic research presented in Chapter 7 of this document
continues to indicate that children, asthmatics in particular, constitute a sensitive group in
epidemiologic studies of oxidant air pollution. Healthy children with prolonged exposure
periods, combined with elevated levels of exertion or exercise also appear to be more vulnerable
to O3-related respiratory health effects. Older adults (>65 years of age) also have been shown to
be susceptible to O3 health effects, particularly O3-related mortality. One epidemiologic study
also found that the elderly experience diminished lung function responses to O3 exposure, which
is supported by evidence from clinical studies.
8.7.3 Vulnerability of Outdoor Workers and Others Who Participate in
Outdoor Activities
The 1996 O3 AQCD indicated that one population group that was shown to have increased
responsiveness to ambient O3 exposure consisted of exercising healthy and asthmatic
individuals, including children, adolescents, and adults. The effects of O3 on the respiratory
health of outdoor workers and others who participate in outdoor activities have been investigated
in several recent epidemiologic studies. These individuals may experience increased
8-61
-------
vulnerability for O3 health effects, because they are typically exposed to high doses of O3 as they
spend long hours outdoors often at elevated exertion levels.
In a group of berry pickers in Fraser Valley, Canada, large decrements in lung function
(-5% decrease in FEVj per 40 ppb increase in 1-h max O3) were associated with acute exposure
to O3 (Brauer et al., 1996). The mean ambient 1-h max O3 was 40.3 ppb (SD 15.2) over the
study period of June to August 1993. The berry pickers worked outdoors for an average of 11 h
at elevated heart rates (on average, 36% higher than resting levels). These results indicate that
extended exposures to O3 at elevated exertion levels can produce marked effects on lung
function among outdoor workers.
Hoppe et al. (1995) examined forestry workers for O3-related changes in pulmonary
function in Munich, Germany. Ventilation rates, estimated from their average activity levels,
were elevated. When comparisons were made between high O3 days (mean lA>-h max O3 of 64
ppb) and low O3 days (mean !/2-h max O3 of 32 ppb), 59% of the forestry workers experienced a
notable decrement in lung function (i.e., at least a 20% increase in specific airway resistance or
at least a 10% decrease in FEVl5 FVC, or PEF) on high O3 days. None experienced improved
lung function. This study also examined athletes following a 2-h outdoor training period in the
afternoon yielding a ventilation rate double the estimate for the forestry workers. Though a
significant association between ambient O3 levels and decrements in FEVj was observed overall,
a smaller percentage of the athletes (14%) experienced a notable decrement in lung function on
high O3 days compared to the forestry workers; and 19% of the athletes actually showed an
improvement.
A large field study by Korrick et al. (1998) examined the effects of multihour O3 exposures
(on average, 8 h) on adults hiking outdoors in Mount Washington, NH. The mean of the hourly
O3 concentrations during the hike was 40 ppb (range 21-74). After the hike, all subjects
combined experienced a relatively small mean decline in FEVj (1.5% decrease per 30 ppb
increase in mean hourly O3 concentrations) during the hike. Ozone-related changes in lung
function parameters were estimated. Stratifying the data by hiking duration indicated that
individuals who hiked 8 to 12 h experienced a >2-fold decline in FEVj versus those only hiking
2 to 8 h.
Results from the above field studies are consistent with those from earlier summer camp
studies (Avol et al., 1990; Higgins et al., 1990; Raizenne et al., 1987, 1989; Spektor et al., 1988,
8-62
-------
1991), which also observed strong associations between acute O3 exposure and decrements in
lung function among children who spent long hours outdoors. In a recent analysis by the
southern California Children's Health Study, a total of 3,535 initially nonasthmatic children
(ages 9 to 16 years at enrollment) were followed for up to 5 years to identify new-onset asthma
cases associated with higher long-term ambient O3 concentrations (McConnell et al., 2002).
Communities were stratified by pollution levels, with six high-O3 communities (mean 1-h max
O3 of 75.4 ppb [SD 6.8] over four years) and six low-O3 communities (mean 50.1 ppb
[SD 11.0]). In the combined analysis using all children, asthma risk was not found to be higher
for residents of the six high-O3 communities versus those from the six low-O3 communities.
However, within the high-O3 communities, asthma risk was more than 3 times greater for
children who played three or more sports versus those who played no sports, an association not
observed in the low-O3 communities. Therefore, among children repeatedly exposed to higher
O3 levels, increased exertion outdoors (and resulting increased O3 dose) was associated with
excess asthma risk.
These field studies with subjects at elevated exertion levels support the extensive evidence
derived from controlled human exposure studies. The majority of human chamber studies have
examined the effects of O3 exposure in subjects performing continuous or intermittent exercise
for variable periods of time (see Chapter 6). Significant O3-induced respiratory responses have
been observed in clinical studies of exercising individuals. The epidemiologic studies discussed
above also indicate that prolonged exposure periods, combined with elevated levels of exertion
or exercise, may magnify O3 effects on lung function. Thus, outdoor workers and others who
participate in higher exertion activities outdoors during the time of day when high peak O3
concentrations occur appear to be particularly vulnerable to O3 effects on respiratory health.
8.7.4 Genetic Factors Affecting O3 Susceptibility
In the 1996 O3 AQCD, the potential influence of genetic factors on responses to O3
exposure could not be thoroughly evaluated due to the very limited data then available. The
human and toxicologic data were inadequate to suggest differences in O3 responsiveness based
on gender, ethnic, or racial background. New animal toxicology studies using various strains of
mice and rats that have identified O3-sensitive and resistant strains illustrate the importance of
genetic background in determining O3 susceptibility. Using subacute low exposure regimen
8-63
-------
(0.3 ppm O3, 48 h) studies on inbred strains that have been designated as inflammation prone or
resistant, Kleeberger et al., (1997) identified the pro-inflammatory cytokine gene, Tnf-a, as a
susceptibility gene. Further characterization of this model indicated a role for TNF receptors in
O3-induced pulmonary epithelial injury and inflammation (Cho et al., 2001). Studies of five
inbred strains of mouse with differing response to O3 exposure (acute high dose or low dose
continuous exposure for 3 days), found a protective role for clara cell secretory protein against
O3-induced oxidative damage (Broeckaert et al., 2003; Wattiez et al., 2003). The role for these
genes and/or their orthologs in human susceptibility to O3 exposure has yet to be examined.
Biochemical and molecular parameters evaluated in these toxicology experiments were
used to identify specific loci on chromosomes and, in some cases, to relate the differential
expression of specific genes to biochemical and physiological differences seen among these
species. Using O3-sensitive and O3-resistant species, it has been possible to identify the
involvement of AHR and inflammation processes in O3 susceptibility. However, most of these
studies were carried out using relatively high doses of O3, making the relevance of these studies
questionable for human health effects assessment. The molecular parameters identified in
animal and human experimental studies may serve as useful biomarkers with the availability
of suitable technologies and may, ultimately, be integrated with epidemiologic studies. The
interindividual differences in O3 responsiveness observed across a spectrum of symptoms and
lung function responses do not yet allow identification of important underlying factors, except a
significant role for age.
Apart from age at the time of exposure, controlled human exposure studies have also
indicated a high degree of interindividual variability in some pulmonary physiological
parameters. Genetic factors likely contribute to the substantial variability observed between
individuals. Several recent human clinical and epidemiologic studies (Bergamaschi et al., 2001;
Corradi et al., 2002; David et al., 2003; Romieu et al., 2004; Yang et al., 2005) have reported
that genetic polymorphisms for antioxidant enzymes and inflammatory genes may modulate
pulmonary function and inflammatory responses to O3 exposure. Glutathione S-transferases
(GSTs) play a major role in protecting cells against damage from reactive oxygen species by
conjugating them with glutathione so that they can be rapidly eliminated. A common
homozygous deletion polymorphism of the GSTM1 gene (GSTM1 null genotype) abolishes
enzyme activity and may increase susceptibility to O3-induced oxidative stress. Asthmatic
8-64
-------
children with a genetic deficiency of GSTM1 were found to be more responsive to ambient O3
exposure, as assessed by decrements in FEF25.75, in a Mexico City study by Romieu et al. (2004).
Antioxidant supplementation of vitamins C and E attenuated post-O3 lung function response in
these children. More specific genotyping has shown that O3 responsiveness of asthmatic
children may be related to the presence of variant Ser allele for a detoxifying enzyme (NQO1)
induced in response to oxidative stress (David et al., 2003). The presence of at least one NQO1
Ser allele in asthmatic children with GSTM1 null genotype was found to have a protective effect
against O3 exposure. Also, polymorphism in Tnf-a has been implicated in the risk of O3-induced
lung function changes in healthy, mild asthmatics and individuals with rhinitis (Yang et al.,
2005), which supports toxicologic evidence referred to earlier in this section.
The above observations suggest a potential role for these genetic markers in the innate
susceptibility to O3; however, the validity of these markers and their relevance in the context of
prediction to population studies need additional validation. The lack of correlation between lung
function and airway inflammatory responses to O3 in healthy subjects, combined with the
evidence of separate chromosomal loci for O3-induced AHR and airway inflammation in inbred
mice, suggest that these two responses are probably independently regulated.
8.7.5 Potential Public Health Impacts
Exposure to ambient O3 is associated with a variety of health outcomes, including
increased incidence of cough, reduction in lung function, increased inflammation, and increased
hospital admissions and mortality. In protecting public health, a distinction must be made
between health effects that are considered "adverse" and those that are not. What constitutes an
adverse health effect varies for different population groups, with some changes in healthy
individuals not being viewed as adverse but those of similar type and magnitude in other
susceptible individuals with preexisting disease being seen as adverse.
8.7.5.1 Concepts Related to Defining of Adverse Health Effects
The American Thoracic Society (ATS) published an official statement on "What
Constitutes an Adverse Health Effect of Air Pollution?" (ATS, 2000). This statement updated
guidance for defining adverse respiratory health effects published fifteen years earlier (ATS,
8-65
-------
1985), in order to take into account new investigative approaches used to identify the effects of
air pollution and to reflect concern for impacts of air pollution on specific susceptible groups.
In the 2000 update, there is an increased focus on quality of life measures as indicators of
adversity and, also, a more specific consideration of population risk. Exposure to air pollution
that increases the risk of an adverse effect to the entire population is viewed as adverse, even
though it may not increase the risk of any identifiable individual to an unacceptable level.
For example, a population of asthmatics could have a distribution of lung function such that no
identifiable single individual has a level associated with significant impairment; and exposure to
air pollution could shift the distribution to lower levels that still do not bring any identifiable
individual to a level that is associated with clinically relevant effects. However, this would be
considered to be adverse because individuals within the population would have diminished
reserve function and, therefore, would be at increased risk if affected by another agent.
Reflecting new investigative approaches, the ATS statement also describes the potential
usefulness of research into the genetic basis for disease, including responses to environmental
agents that provide insights into the mechanistic basis for susceptibility and provide markers of
risk status. Likewise, biomarkers that are indicators of exposure, effect or susceptibility may
someday be useful in defining the point at which one or an array of responses should be
considered an adverse effect.
The 1996 O3 AQCD provided information useful in helping to define adverse health effects
associated with ambient O3 exposure by describing the gradation of severity and adversity of
respiratory-related O3 effects, and those definitions are reproduced and presented here as
Tables 8-2 and 8-3. The severity of effects described in those tables and the approaches taken to
define their relative adversity still appear to be valid and reasonable even in the context of the
new ATS statement (ATS, 2000).
As assessed in detail in earlier chapters of this document and briefly recapitulated in
preceding sections of this chapter, exposures to a range of O3 concentrations have been reported
to be associated with increasing severity of several categories of health effects.
Respiratory effects associated with short-term O3 exposures have been by far the most
extensively studied and most clearly shown to be causally related to O3 exposure. Such effects
include the induction of pulmonary function decrements and respiratory symptoms demonstrated
in response to controlled acute (1- to 8-h) O3 exposures of human subjects and also observed
8-66
-------
Table 8-2. Gradation of Individual Responses to Short-Term Ozone Exposure in
Healthy Persons a
Functional Response
FEVj
Nonspecific
bronchial responsiveness a
Duration of response
Symptomatic Response
Cough
Chest pain
Duration of response
Impact of Responses
Interference with normal
activity
None
Within normal
range (±3%)
Within normal
range
None
Normal
Infrequent
cough
None
None
Normal
None
Small
Decrements of
3 to < 10%
Increases of
<100%
<4 hours
Mild
Cough with deep
breath
Discomfort just
noticeable on
exercise or
deep breath
<4 hours
Normal
None
Moderate
Decrements of
>10but<20%
Increases of <300%
>4 hours but
<24 hours
Moderate
Frequent
spontaneous cough
Marked discomfort
on exercise or deep
breath
>4 hours but
<24 hours
Mild
A few sensitive
individuals choose to
limit activity
Large
Decrements of
>20%
Increases of >300%
>24 hours
Severe
Persistent
uncontrollable
cough
Severe discomfort
on exercise or
deep breath
>24 hours
Moderate
Many sensitive
individuals choose
to limit activity
a An increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PD20 or PD1(
Source: This table is reproduced from the 1996 O3 AQCD (Table 9-1, page 9-24)
(U.S. Environmental Protection Agency, 1996a).
epidemiologically to be associated with ambient O3 exposures in children and adults vigorously
engaged in outdoor activities. The severity of the symptoms and magnitude of the pulmonary
decrements, as previously noted, depend on the inhaled O3 dose and individual O3 sensitivity of
the exposed persons. Controlled exposure chamber studies assessed in the 1996 O3 AQCD
provided very clear evidence that statistically significant reductions in lung function occurred in
healthy adults in response to 6.6-h O3 exposures as low as 0.08 ppm under moderate intermittent
exercise conditions. Of considerable importance, whereas none of the subjects showed marked
8-67
-------
Table 8-3. Gradation of Individual Responses to Short-Term Ozone Exposure in
Persons with Impaired Respiratory Systems
Functional
FEVj change
Nonspecific
bronchial
responsiveness a
Airway resistance
(SRaw)
Duration of
response
Symptomatic
Wheeze
Cough
None
Decrements of
Within normal
range
Within normal
range (±20%)
None
Normal
None
Infrequent
cough
Small
Decrements of
3 to < 10%
Increases of <100%
SRaw increased <100%
<4 hours
Mild
With otherwise normal
breathing
Cough with deep breath
Moderate
Decrements of
>10 but <20%
Increases of <300%
SRaw increased up
to 200% or up to
15cmH2O/s
>4 hours but
<24 hours
Moderate
With shortness of
breath
Frequent spontaneous
cough
Large
Decrements of
>20%
Increases of >300%
SRaw increased
>200% or more than
15cmH2O/s
>24 hours
Severe
Persistent with
shortness of breath
Persistent
uncontrollable
cough
Chest pain
Duration of
response
None
None
Discomfort just
noticeable on exercise
or deep breath
< 4 hours
Marked discomfort
on exercise or deep
breath
>4 hours, but
<24 hours
Severe discomfort
on exercise or deep
breath
>24 hours
Impact of
Responses
Normal
Mild
Moderate
Severe
Interference with
normal activity
None
Medical treatment No change
Few individuals choose
to limit activity
Normal medication as
needed
Many individuals
choose to limit
activity
Increased frequency
of medication use or
additional medication
Most individuals
choose to limit
activity
Physician or
emergency room
visit
a An increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PD20 or PD1(
Source: This table is reproduced from the 1996 O3 AQCD (Table 9-2, page 9-25) (U.S. Environmental
Protection Agency, 1996a).
8-68
-------
decrements in FEVj measures (e.g., >10% decreases in FEVj) at the end of the 6.6-h exposure to
filtered air, a notable percentage did so with O3 exposure. More specifically, the percentages
showing >10% FEVj decrements at 0.08, 0.10, and 0.12 ppm O3 were 26%, 31%, and 46%,
respectively, as illustrated in Figure 8-1 A. Newly available results from the Adams (2002, 2006)
studies confirm the finding of a notable percentage (17 to 23%) of healthy adult subjects
exhibiting >10% decrements in lung function indexed by FEVj measurements (as shown in
Figure 8-IB and Figure 8-2). The newly available Adams (2002, 2006) results also further
extend the earlier findings in terms of providing indications of >10% FEVj decrements among a
small percentage (3 to 7%) following 6.6-h exposure of experimentally tested healthy adults to
0.06 ppm O3 while engaged in moderate exercise (as also shown in Figures 8-1B and 8-2).
Given that comparable or enhanced pulmonary function decrements are observed in
asthmatic children (accompanied by respiratory symptoms), it is reasonable to project that at
least comparable percentages of such children would exhibit >10% FEVj decrements (and
respiratory symptoms) at 0.06 or 0.08 ppm if tested under similar exercise conditions. The
likelihood of notable lung function decrements (as indexed by >10% FEVj decrease) and/or
respiratory symptoms occurring among asthmatic children in response to relatively low ambient
O3 exposures (i.e., <0.08 O3 ppm, 1- to 8-h average) is further supported by the epidemiologic
observations of the Thurston et al. (1997), Mortimer et al. (2002), and Hoppe et al. (2003)
studies discussed above in Section 8.7.2. The Hoppe et al. (2003) study, in fact, suggests that
about 20% of asthmatic children may be at risk of experiencing >10% FEVj decrements, which
presumably could be sufficient to cause many of them to limit activities or increase medication
use (as indicated in Table 8-3). Such projected potential impacts on asthmatic children in
response to acute (1- to 8-h) ambient O3 exposures of <0.08 ppm may possibly occur in the range
of 0.06 to 0.07 ppm, based on the Hoppe et al. (2003) study results. The observed lung function
decrements, as well as any accompanying symptoms and/or increased medication use, would be
consistent with and lend plausibility to the results of currently available time-series studies that
found associations between ambient O3 concentrations and increased asthma-related ED visits or
hospitalizations, especially during the warm season.
8-69
-------
8.7.5.2 Estimation of Potential Numbers of Persons in At-Risk Susceptible Population
Groups in the United States
Although O3-related health risk estimates may appear to be small, they may well be
significant from an overall public health perspective due to large numbers of individuals in
potential risk groups. Several subpopulations may be identified as having increased
susceptibility or vulnerability to adverse health effects from O3, including: older adults,
children, individuals with preexisting pulmonary disease, and those with higher exposure levels,
such as outdoor workers.
One consideration in the assessment of potential public health impacts is the size of various
population groups that may be at increased risk for health effects associated with O3-related air
pollution exposure. Table 8-4 summarizes information on the prevalence of chronic respiratory
conditions in the U.S. population in 2002 and 2003 (Dey and Bloom, 2005; Lethbridge-£ejku
et al., 2004). Individuals with preexisting cardiopulmonary disease constitute a fairly large
proportion of the population, with tens of millions of people included in each disease category.
Of most concern here are those individuals with preexisting respiratory conditions, with
approximately 11% of U.S. adults and 13% of children having been diagnosed with asthma and
6% of adults having COPD (chronic bronchitis and/or emphysema). Table 8-5 provides further
information on the number of various specific respiratory conditions per 100 persons by age
among the U.S. population during the mid-1990s. Asthma prevalence tends to be higher
in children than adults.
In addition, subpopulations based on age group also comprise substantial segments of the
population that may be potentially at risk for O3-related health impacts. Based on U.S. census
data from 2003, about 26% of the U.S. population are under 18 years of age and 12% are
65 years of age or older. Hence, large proportions of the U.S. population are included in age
groups that are considered likely to have increased susceptibility and vulnerability for health
effects from ambient O3 exposure.
The health statistics data illustrate what is known as the "pyramid" of effects. At the top of
the pyramid, there are approximately 2.5 millions deaths from all causes per year in the U.S.
population, with about 100,000 deaths from chronic lower respiratory diseases (Kochanek et al.,
2004). For respiratory health diseases, there are nearly 4 million hospital discharges per year
(DeFrances et al., 2005), 14 million ED visits (McCaig and Burt, 2005), 112 million ambulatory
8-70
-------
oo
Table 8-4. Prevalence of Selected Respiratory Disorders by Age Group and by Geographic Region in the United States
(2002 [U.S. Adults] and 2003 [U.S. Children] National Health Interview Survey)
Chronic Condition/Disease
Respiratory Conditions
Asthma
COPD
Chronic Bronchitis
Emphysema
Chronic Condition/Disease
Respiratory Conditions
Asthma
Age (Years)
Adults (18+ Years) 18-44 45-64 65-74
Cases
(xlO6) % % % %
21.9 10.6 11.5 10.6 8.4
9.1 4.4 3.5 5.5 5.5
3.1 1.5 0.3 2 4.9
Age (Years)
Children
(<18 years) 0-4 5-11 12-17
Cases
(xlO6) % % % %
9.1 12.5 7.5 14 14.7
Region
75+ Northeast Midwest South West
% % % % %
7.6 11 10.9 9.8 11.8
5.3 3.8 4 5.4 3.8
4.7 1.5 1.8 1.7 1.1
Region
Northeast Midwest South West
% % % %
14 13.5 11.8 11.2
Source: Lethbridge-Cejku et al. (2004) for data on adults (18+ years); Dey and Bloom (2005) for data on children (<18 years).
-------
Table 8-5. Acute Respiratory Conditions per 100 Persons/Year by Age Group in the
United States (1996 National Health Interview Survey)
45+ Years
Type of Acute
Condition
Respiratory Conditions
Common Cold
Other Acute Upper
Respiratory Infections
Influenza
Acute Bronchitis
Pneumonia
Other Respiratory
Conditions
All
Ages
78.9
23.6
11.3
36
4.6
1.8
1.7
Under 5
Years
129.4
48.6
13.1
53.7
7.2a
3.9a
2.9a
5-17
Years
101.5
33.8
15
44.3
4.3
1.7a
2.4a
18-24
Years
86
23.8
16.1
40.5
3.9a
1.4a
0.4a
25-44
Years
76.9
18.7
11.6
38.1
5.1
1.3a
2.0a
Total
53.3
16.1
7
23.3
3.8
2.0a
l.la
45-64
Years
55.9
16.4
7.5
26.1
3.5
0.9a
1.5a
65+
Years
49
15.7
6.1
18.6
4.4a
3.8a
0.5a
a Figure does not meet standard of reliability or precision.
Source: Adams et al. (1999).
care visits (Woodwell and Cherry, 2004), and an estimated 700 million restricted activity days
per year due to respiratory conditions (Adams et al., 1999). Combining small risk estimates with
relatively large baseline levels of health outcomes can result in quite large public health impacts.
Thus, even a small percentage reduction in O3 health impacts on cardiopulmonary diseases
would reflect a large number of avoided cases.
Another key input for public health impact assessment is the range of concentration-
response functions for various health outcomes. Epidemiologic studies have reported
associations between short-term exposure to O3 with mortality, hospitalizations for pulmonary
diseases, ED visits for asthma, reduced lung function, and incidence of respiratory symptoms.
Effect estimates for morbidity responses to short-term changes in O3 tend to be larger and more
variable in magnitude than those for mortality.
In addition to attribution of risks for various health outcomes related to O3 and other
copollutants, important considerations in assessing the impact of O3 on public health include the
size of population groups at risk, as well as the concentration-response relationship and potential
8-72
-------
identification of threshold levels. Taken together, based on the above information, it can be
concluded that exposure to ambient O3 likely has a significant impact on public health in the
United States.
8.8 SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS
This section summarizes the main conclusions derived from this integrated synthesis of
information regarding health effects associated with ambient O3 exposures. The conclusions are
based on human clinical, animal toxicologic, and epidemiologic studies that have evaluated
health effects associated with short-term, repeated, and long-term exposures to O3 alone or in
combination with other ambient pollutants. The controlled human exposure (or "clinical")
studies provide the clearest and most compelling evidence for human health effects directly
attributable to acute exposures to O3 per se. The evidence from human and animal toxicologic
studies presented in Chapters 4, 5, and 6 are further useful in not only providing insights into
possible mechanisms of action underlying different types of O3-related health effects but, also,
in helping to provide biological plausibility for health effects observed in epidemiologic studies
assessed in Chapter 7. These empirical efforts are also aimed at identifying susceptible and
vulnerable populations that are at potentially greater risk for effects of O3 exposure. Overall, the
new evidence generally supports and builds further upon key health-related conclusions drawn in
the previous 1996 AQCD. The following conclusions integrate results from newly available
studies with the scientific evidence assessed in that document.
1. Health Effects of Short-term Exposures to Ozone
The 1996 O3 AQCD assessed a substantial body of evidence from toxicologic, human
clinical, and epidemiologic studies. That AQCD concluded that short-term ambient O3 exposure
resulted in various respiratory health effects, including lung function decrements and increased
respiratory symptoms in both healthy and asthmatic individuals exposed during moderate to
heavy exercise to O3 concentrations ranging down to the lowest levels (0.12 ppm for 1 h;
0.08 ppm for 6.6 to 8 h) tested in the available controlled human exposure studies. Such
experimentally demonstrated effects were consistent with and lent plausibility to epidemiologic
8-73
-------
observations highlighted in the 1996 AQCD of increases in daily hospital admissions and ED
visits for respiratory causes. Epidemiologic evidence also provided suggestive evidence for an
association between short-term O3 exposure and mortality. However, there was essentially no
evidence available in the 1996 O3 AQCD regarding potential cardiovascular effects of short-term
O3 exposure. The newly-available evidence assessed in this revised O3 AQCD notably enhances
our understanding of short-term O3 exposure effects, as summarized below, first in relation to
respiratory morbidity endpoints and then cardiovascular effects and, lastly, mortality.
A. Respiratory Morbidity
Lung Function: Controlled exposure studies clearly demonstrate acute reversible
decrements in lung function in healthy adults exposed to >0.08 ppm O3 when minute ventilation
and/or duration of exposure are increased sufficiently. On average, spirometric responses to O3
exposure appear to decline with increasing age starting at approximately 18 years of age. There
is considerable variability in responses between similarly exposed individuals, such that some
may experience distinctly larger effects even when small group mean responses are observed.
For example, healthy adults exposed to 0.08 ppm O3 for 6.6 h with moderate exercise exhibited a
group mean O3-induced decrement in FEVj of about 6%, but a decrement of >10% was seen in
23% of these individuals. Also, exposure to 0.06 ppm O3 caused >10% lung function decline in
a small percentage (7%) of the subjects. Summer camp field studies conducted in southern
Ontario, Canada, in the northeastern U.S., and in southern California have also reported lung
function responses in pre-adolescent children associated with ambient O3 levels.
With repeated acute O3 exposures (0.12 to 0.45 ppm for 1 h) over several days, controlled
exposure studies typically find that FEVj response to O3 is enhanced on the second of several
days of exposure, but spirometric responses become attenuated on subsequent days with these
repeated exposures. However, this tolerance is lost after about a week without exposure.
Animal toxicologic studies also provide extensive evidence that acute O3 exposures alter
breathing patterns so as to cause rapid shallow breathing (i.e., increased frequency and decreased
tidal volume), an effect which attenuates after several days of exposure. Such results from
controlled human exposure studies and animal toxicologic studies provide clear evidence of
causality for the associations observed between acute (<24 h) O3 exposure and relatively small,
but statistically significant declines in lung function observed in numerous recent epidemiologic
8-74
-------
studies. Declines in lung function are particularly noted in children, asthmatics, and adults who
work or exercise outdoors.
Respiratory Symptoms: Some young healthy adult subjects exposed in clinical studies to
O3 concentrations >0.08 ppm for 6 to 8 h during moderate exercise exhibit symptoms of cough
and pain on deep inspiration. An increase in the incidence of cough has been found in clinical
studies as low as 0.12 ppm in healthy adults during 1 to 3 h with very heavy exercise and other
respiratory symptoms, such as pain on deep inspiration and shortness of breath, have been
observed at 0.16 ppm to 0.18 ppm with heavy and very heavy exercise. These O3-induced
respiratory symptoms gradually decrease in adults with increasing age. With repeated O3
exposures over several days, respiratory symptoms become attenuated, but this tolerance is lost
after about a week without exposure. The epidemiologic evidence shows significant associations
between acute exposure to ambient O3 and increases in a wide variety of respiratory symptoms
(e.g., cough, wheeze, production of phlegm, and shortness of breath) in asthmatic children.
Epidemiologic studies also indicate that acute O3 exposure is likely associated with increased
asthma medication use in asthmatic children. On the other hand, an effect of acute O3 exposure
on respiratory symptoms in healthy children is not as clearly indicated by epidemiology studies,
consistent with diminished symptom responses seen in healthy children in human clinical
studies.
Airway Inflammation: Inflammatory responses have been observed subsequent to 6.6 h O3
exposures to the lowest tested level of 0.08 ppm in healthy human adults. Some studies suggest
that inflammatory responses may be detected in some individuals following O3 exposures even
in the absence of O3-induced pulmonary function decrements in those subjects. With repeated
O3 exposures over several days, an attenuation of most inflammatory markers occurs. However,
none of the several markers of lung injury and permeability evaluated show attenuation,
indicating continued lung tissue damage during repeated exposure. Animal toxicologic studies
provide extensive evidence that acute (1 to 3 h) O3 exposures as low as 0.1 to 0.5 ppm can cause
(1) lung inflammatory responses (typified by increased reactive oxygen species, inflammatory
cytokines, influx of PMNs, and activation of alveolar macrophages); (2) damage to epithelial
airway tissues, (3) increases in permeability of both lung endothelium and epithelium, and
(4) increases in susceptibility to infectious diseases due to modulation of lung host defenses.
Consistent with these experimental findings, there is also limited epidemiologic evidence
8-75
-------
showing an association between acute ambient O3 exposure and airway inflammation in children
acutely exposed to ambient O3 concentrations (1-h max O3 of approximately 0.1 ppm). The
extensive human clinical and animal toxicological evidence, together with the limited available
epidemiologic evidence, is clearly indicative of a causal role for O3 in inflammatory responses in
the airways.
Airway Responsiveness: Controlled human exposure studies have found that acute O3
exposure causes an increase in nonspecific airway responsiveness, as indicated by reductions in
concentrations of methacholine or histamine required to produce a given decrease in FEVj or
increase in SRaw. Acute (2- or 3-h) O3 exposure at 0.25 or 0.4 ppm of allergic asthmatic subjects,
who characteristically already have somewhat increased airway responsiveness at baseline, was
found to cause further increases in airway responsiveness in response to allergen challenges.
Also, repeated daily exposure to 0.125 ppm O3 for 4 days exacerbated lung function decrements
in response to bronchial allergen challenges among persons with preexisting allergic airway
disease, with or without asthma. Ozone-induced exacerbation of airway responsiveness persists
longer and attenuates more slowly than O3-induced pulmonary function decrements and
respiratory symptom responses. Heightened airway responsiveness (reactivity) has also been
observed in several laboratory animal species with acute exposures (1 to 3 h) to 0.5 to 1.0 ppm
O3. Ozone increases airway hyperreactivity to bronchoconstrictive agents (e.g., ovalbumin), and
there is a temporal relationship between inflammatory cell influx and O3-induced increases in
airway reactivity. Several studies of sensitized laboratory animals showing O3-induced increases
in airway hyperreactivity are consistent with O3 exacerbation of airway hyperresponsiveness
reported in atopic humans with asthma. Airway responsiveness has not been widely examined in
epidemiologic studies. However, the evidence from human clinical and animal toxicological
studies clearly indicate that acute exposure to O3 can induce airway hyperreactivity, thus likely
placing atopic asthmatics at greater risk for more prolonged bouts of breathing difficulties due to
airway constriction in response to various airborne allergens or other triggering stimuli.
Respiratory Hospital Admissions and Emergency Department Visits: Aggregate
population time-series studies observed that ambient O3 concentrations are positively and
robustly associated with respiratory-related hospitalizations and asthma ED visits during the
warm season. These observations are strongly supported by the human clinical, animal
toxicologic, and epidemiologic evidence for lung function decrements, increased respiratory
8-76
-------
symptoms, airway inflammation, and airway hyperreactivity. Taken together, the overall
evidence supports a causal relationship between acute ambient O3 exposures and increased
respiratory morbidity outcomes resulting in increased ED visits and hospitalizations during the
warm season.
B. Cardiovascular Morbidity
At the time of the 1996 O3 AQCD, the possibility of O3-induced cardiovascular effects was
a largely unrecognized issue. Newly-available evidence has emerged since then which provides
considerable plausibility for how O3 exposure could exert cardiovascular impacts. This includes
direct O3 effects such as O3-induced release from lung epithelial cells of platelet activating factor
(PAF) that may contribute to blood clot formation that would increase the risk of serious
cardiovascular outcomes (e..g, heart attack, stroke, mortality). Also, interactions of O3 with
surfactant components in epithelial lining fluid of the lung results in production of oxysterols
and reactive oxygen species that may exhibit PAF-like activity contributing to clotting and/or
exert cytotoxic effects on lung and heart cells. Other possible mechanisms may involve O3-
induced secretions of vasoconstrictive substances and/or effects on neuronal reflexes that may
result in increased arterial blood pressure and/or altered electrophysiologic control of heart rate
or rhythm. Consistent with the latter possibility, some field/panel studies that examined
associations between O3 and various cardiac physiologic endpoints have yielded limited
epidemiologic evidence suggestive of a potential association between acute O3 exposure and
altered HRV, ventricular arrhythmias, and incidence of MI. Also, highly suggestive evidence for
O3-induced cardiovascular effects is provided by a few population studies of cardiovascular
hospital admissions which reported positive O3 associations during the warm season between
ambient O3 concentrations and cardiovascular hospitalizations. On the other hand, the only
controlled human exposure study that evaluated effects of O3 exposure on cardiovascular health
outcomes found no significant O3-induced differences in ECG, heart rate, or blood pressure in
healthy or hypertensive subjects, but did observe an overall increase in myocardial work and
impairment in pulmonary gas exchange. Also, some animal toxicological studies have shown
O3-induced decreases in heart rate, mean arterial pressure, and core temperature. Overall, then,
this generally limited body of evidence is highly suggestive that O3 directly and/or indirectly
8-77
-------
contributes to cardiovascular-related morbidity, but much remains to be done to more fully
substantiate links between ambient O3 exposure and adverse cardiovascular outcomes.
C. Mortality
Numerous recent epidemiologic studies conducted in the United States and abroad have
investigated the association between acute exposure to O3 and mortality. Results from several
large U.S. multicity studies as well as several single-city studies indicate a positive association
between increases in ambient O3 levels and excess risk of all-cause (nonaccidental) daily
mortality. Determining cause-specific mortality is more difficult due to reduced statistical
power by which to examine cause-specific associations and the lack of clarifying information on
contributing causes of death. That is, attribution to one or the other of the more specific
cardiopulmonary causes may underplay contributions of chronic cardiovascular disease to
"respiratory" deaths (e.g., a heart attack victim succumbing to acute pneumonia) or vice versa.
Nevertheless, consistent with observed O3-related increases in respiratory- and cardiovascular-
related morbidity, several newer multicity studies, single-city studies, and several meta-analyses
of these studies have provided relatively strong epidemiologic evidence for associations between
short-term O3 exposure and all-cause mortality, even after adjustment for the influence of season
and PM. In addition, consistently positive associations have been reported for O3-related
cardiovascular mortality across approximately 30 studies, with two well-conducted multicity
studies in the United States and Europe yielding small, but statistically significant positive
associations. Also, as discussed in Section 8.6, newly available experimental data from both
animal and human studies provide evidence suggestive of plausible pathways by which risk of
respiratory or cardiovascular morbidity and mortality could be increased by ambient O3 either
acting alone or in combination with copollutants in ambient air mixes. This overall body of
evidence is highly suggestive that O3 directly or indirectly contributes to non-accidental and
cardiopulmonary-related mortality, but additional research is needed to more fully establish
underlying mechanisms by which such effects occur.
2. Health Effects of Long-term Exposures to Ozone
In the 1996 O3 AQCD, the available epidemiologic data provided only suggestive evidence
that respiratory health effects were associated with chronic O3 exposure. Animal toxicologic
8-78
-------
studies indicated that chronic O3 exposure caused structural changes in the respiratory tract, and
simulated seasonal exposure studies in animals suggested that such exposures might have
cumulative impacts. As summarized below, recent studies are generally consistent with the
conclusions drawn in the previous 1996 AQCD.
A. Respiratory Morbidity
Lung Function: Recent epidemiologic studies observed that reduced lung function growth
in children was associated with seasonal exposure to O3; however, cohort studies investigating
the effect of annual or multiyear O3 exposure observed little clear evidence for impacts of
longer-term, relatively low-level O3 exposure on lung function development in children. The
epidemiologic data, collectively, indicate that the current evidence is suggestive but inconclusive
for respiratory health effects from long-term O3 exposure.
Morphological Changes: Animal toxicologic studies continue to show chronic O3-induced
structural alterations in several regions of the respiratory tract including the centracinar region.
Morphologic evidence from some recent studies using exposure regimens that mimic seasonal
exposure patterns report increased lung injury compared to conventional chronic stable
exposures. Infant rhesus monkeys repeatedly exposed to 0.5 ppm 8h/day O3 for 11 episodes
exhibited: (1) remodeling of the distal airways; (2) abnormalities in tracheal basement
membrane; (3) eosinophil accumulation in conducting airways; and (4) decrements in airway
innervation. Long-term O3 exposure of rats to 0.5 or 1.0 ppm for 20 months resulted in upper
respiratory tract mucus metaplasia and hyperplasia in the nasal epithelium (0.25 or 0.5 ppm,
8h/day, 7days/wk for 13 weeks). The persistent nature of these cytological changes raise the
possibility of long-lasting alterations in human airways in response to chronic O3 exposure, but it
is highly uncertain as to what long-term patterns of exposure or O3 concentrations in humans
may be requisite to produce analogous morphological changes. Nor is it now possible to
characterize the possible magnitude or severity of any such effects occurring in humans in
response to ambient O3 exposures at levels observed in the United States.
Incidence of Lung Cancer: The weight of evidence from recent animal toxicological
studies and a very limited number of epidemiologic studies do not support ambient O3 as a
pulmonary carcinogen.
8-79
-------
B. Mortality
Results from the few available epidemiologic studies are inconsistent regarding the
association between long-term exposure to O3 and mortality. There is little evidence to suggest a
causal relationship between chronic O3 exposure and increased risk for mortality in humans.
3. Susceptibility or Vulnerability to Effects Associated with Exposure to Ozone
Various factors have been shown to influence individuals' responses to environmental air
pollutants. Factors that increase susceptibility to O3-related effects include innate factors, such
as genetic predisposition or developmental effects, or disease status. Other factors can lead to
enhanced vulnerability to O3-related effects, such as heightened exposures or activity patterns.
In the 1996 O3 AQCD, available evidence suggested that children, asthmatics, and outdoor
workers were populations that may be more susceptible or vulnerable to effects of O3 exposure.
In addition, controlled human exposure studies also demonstrated a large variation in sensitivity
and responsiveness to O3 in studies of healthy subjects, but the specific factors that contributed
to this intersubject variability were yet to be identified. Recent studies have built upon the
evidence available in the previous review. Factors related to susceptibility or vulnerability to O3
exposure-related effects are briefly summarized below:
People with Preexisting Pulmonary Diseases: Ozone-induced differential responses in
lung function and AFIR in people with allergic rhinitis suggest that asthmatics have potentially
greater responses than healthy people with exposure to O3. There is a tendency for slightly
increased spirometric responses in mild asthmatics and allergic rhinitics relative to healthy
young adults. Spirometric responses in asthmatics appear to be affected by baseline lung
function, i.e., responses increase with disease severity. In addition, repeated O3 exposure over
several days has been shown to increase responsiveness to bronchial allergen challenge in
subjects with preexisting allergic airway disease, with or without asthma. Asthmatics also show
a significantly greater neutrophil response (18 h postexposure) than similarly-exposed healthy
individuals. Epidemiologic studies have reported associations with a range of respiratory health
outcomes in asthmatics, from decreases in lung function to hospitalization or ED visits for
asthma, thus supporting this population group as being likely to experience increased risk for O3-
induced health effects. Although controlled human exposure studies have not found evidence of
larger spirometric changes in people with COPD relative to healthy subjects, this may be due to
8-80
-------
the fact that most people with COPD are older adults who would not be expected to have such
changes based on their age. However, new epidemiologic evidence indicates that people with
COPD may be more likely to experience other effects, including emergency room visits, hospital
admissions, or premature mortality.
Age-related: Controlled human exposure studies have shown that lung function responses
to O3 varies with age, with responsiveness generally diminishing after about 18 to 20 years of
age. Children and older adults thus have lesser respiratory symptoms with O3 exposure than
young healthy adults. Potentially increased O3 doses can be received by individuals
experiencing less severe respiratory symptoms. Evidence from newer epidemiologic studies
supports the 1996 O3 AQCD conclusions that children are more likely at increased risk for O3-
induced health effects. Notably, epidemiologic studies have indicated adverse respiratory health
outcomes associated with O3 exposure in children. In addition, recently published epidemiologic
studies also suggest that older adults (aged >65 years) appear to be at excess risk of O3-related
mortality or hospitalization.
Heightened vulnerability due to greater exposures: Epidemiologic studies have provided
some evidence to indicate that outdoor workers are more vulnerable to O3-related effects, which
is likely related to their increased exposure to ambient air pollution. Controlled human exposure
studies clearly established differential biological response to O3 based on physical activity
(exertion). Epidemiologic studies also suggest that exercising (moderate to high physical
exertion) children and adolescents appear to demonstrate increased responsiveness to ambient
concentrations of O3 and may be more likely to experience O3-induced health effects. Animal
studies show a similar impact of exercise on responsiveness to O3.
Genetic susceptibility: Animal toxicologic studies provide supportive evidence to the
observations of innate susceptibility. Various strains of mice and rats have demonstrated the
importance, in general, of genetic background in O3 susceptibility. Moreover, genetic and
molecular characterization studies in laboratory animals identified genetic loci responsible for
both sensitivity and resistance. Recent human clinical and epidemiologic studies also have
shown that genetic polymorphisms for antioxidant enzymes and inflammatory genes (GSTM1,
NQO1, and Tmf-a) may modulate the effect of O3 exposure on pulmonary function and airway
inflammation.
8-81
-------
4. Health Effects of Ozone-Containing Pollutant Mixtures
The potential interaction of pollutant mixtures with O3 is poorly understood and the animal
studies reviewed in the 1996 O3 AQCD reported additive, synergistic or even antagonistic effects
depending on the exposure regimen and the endpoint studied. A few new controlled human
exposure and animal toxicology studies reviewed in Chapters 4, 5, and 6 investigated health
effects associated with O3-containing pollutant mixtures of near ambient levels. As noted below,
recent studies, although generally consistent with conclusions drawn in the 1996 O3 AQCD, have
added some new information, particularly with respect to interactions between O3 and PM.
Controlled human exposure studies indicate that continuous exposure of healthy human
adults to SO2 or NO2 increases bolus dose O3 absorption, suggesting that co-exposure to other
gaseous pollutants in the ambient air may enhance O3 absorption. Other controlled human
exposure studies that evaluated response to allergens in asthmatics (allergic and dust-mite
sensitive) suggest that O3 enhances response to allergen challenge. Consistent with these
findings, animal toxicology studies also reported enhanced response to allergen on exposure
toO3.
A few other animal toxicology studies that exclusively investigated the co-exposure of PM
and O3 reported increased response (lung tissue injury, inflammatory and phagocytosis) to the
mixture of PM + O3 compared to either PM or O3 alone. Recent investigations on the
copollutant interactions using simulated urban photochemical oxidant mixes suggest the need for
similar studies in understanding the biological basis for air pollutant mixture effects observed in
epidemiologic studies.
8-82
-------
REFERENCES
Adams, W. C. (2002) Comparison of chamber and face-mask 6.6-hour exposures to ozone on pulmonary function
and symptoms responses. Inhalation Toxicol. 14: 745-764.
Adams, W. C. (2003) Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm ozone via square-wave
and triangular profiles on pulmonary responses. Inhalation Toxicol. 15: 265-281.
Adams, W. C. (2006) Comparison of chamber 6.6-h exposures to 0.04-0.08 PPM ozone via square-wave and
triangular profiles on pulmonary responses. Inhalation Toxicol. 18: 127-136.
Adams, P. F.; Hendershot, G. E.; Marano, M. A. (1999) Current estimates from the National Health Interview
Survey, 1996. Hyattsville, MD: U.S. Department of Health and Human Services, Public Health Service,
National Center for Health Statistics; DHHS publication no. (PHS) 99-1528. (Vital and health statistics: v. 10,
data from the National Health Survey, no. 200). Available:
http://www.cdc.gov/nchs/products/pubs/pubd/series/srlO/pre-200/pre-200.htm [12 March, 2001].
Alexis, N.; Urch, B.; Tarlo, S.; Corey, P.; Pengelly, D.; O'Byrne, P.; Silverman, F. (2000) Cyclooxygenase
metabolites play a different role in ozone-induced pulmonary function decline in asthmatics compared to
normals. Inhalation Toxicol. 12: 1205-1224.
American Thoracic Society. (1985) Guidelines as to what constitutes an adverse respiratory health effect, with
special reference to epidemiologic studies of air pollution. Am. Rev. Respir. Dis. 131: 666-668.
American Thoracic Society. (2000) What constitutes an adverse health effect of air pollution? Am. J. Respir. Crit.
CareMed. 161:665-673.
Arito, H.; Takahashi, M.; Iwasaki, T.; Uchiyama, I. (1997) Age-related changes in ventilatory and heart rate
responses to acute ozone exposure in the conscious rat. Ind. Health 35: 78-86.
Avol, E. L.; Trim, S. C.; Little, D. E.; Spier, C. E.; Smith, M. N.; Peng, R.-C.; Linn, W. S.; Hackney, J. D.; Gross,
K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone exposure and lung function in children
attending a southern California summer camp. Presented at: 83rd annual meeting and exhibition of the Air &
Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management
Association; paper no. 90-150.3.
Ball, B. A.; Folinsbee, L. J.; Peden, D. B.; Kehrl, H. R. (1996) Allergen bronchoprovocation of patients with mild
allergic asthma after ozone exposure. J. Allergy Clin. Immunol. 98: 563-572.
Basha, M. A.; Gross, K. B.; Gwizdala, C. J.; Haidar, A. H.; Popovich, J., Jr. (1994) Bronchoalveolar lavage
neutrophilia in asthmatic and healthy volunteers after controlled exposure to ozone and filtered purified air.
Chest 106: 1757-1765.
Bayram, H.; Rusznak, C.; Khair, O. A.; Sapsford, R. J.; Abdelaziz, M. M. (2002) Effect of ozone and nitrogen
dioxide on the permeability of bronchial epithelial cell cultures of non-asthmatic and asthmatic subjects.
Clin. Exp. Allergy 32: 1285-1292.
Bell, M. L.; McDermott, A.; Zeger, S. L.; Samet, J. M.; Dominici, F. (2004) Ozone and short-term mortality in
95 US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292: 2372-2378.
Bell, M. L.; Dominici, F.; Samet, J. M. (2005) A meta-analysis of time-series studies of ozone and mortality with
comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
Bell, M. L. (2006) Community-specific maximum likelihood estimates of O3-related excess risk in mortality for the
NMMAPS U.S. 95 communities study [personal communication with attachments to Jee Young Kim].
New Haven, CT: Yale University School of Forestry and Environmental Studies; January 6.
Bergamaschi, E.; De Palma, G.; Mozzoni, P.; Vanni, S.; Vettori, M. V.; Broeckaert, F.; Bernard, A.; Mutti, A.
(2001) Polymorphism of quinone-metabolizing enzymes and susceptibility to ozone-induced acute effects.
Am. J. Respir. Crit. Care Med. 163: 1426-1431.
Bhalla, D. K.; Hoffman, L. (1997) Time course of airway epithelial and inflammatory changes in rats exposed to
moderate levels of ozone. Inhalation Toxicol. 9: 829-842.
Bosson, J.; Stenfors, N.; Bucht, A.; Helleday, R.; Pourazar, J.; Holgate, S. T.; Kelly, F. J.; Sandstrom, T.; Wilson, S.;
Frew, A. J.; Blomberg, A. (2003) Ozone-induced bronchial epithelial cytokine expression differs between
healthy and asthmatic subjects. Clin. Exp. Allergy 33: 777-782.
Brauer, M.; Blair, J.; Vedal, S. (1996) Effect of ambient ozone exposure on lung function in farm workers. Am.
J. Respir. Crit. Care Med. 154: 981-987.
Broeckaert, F.; Clippe, A.; Wattiez, R.; Falmagne, P.; Bernard, A. (2003) Lung hyperpermeability, Clara-cell
secretory potein (CC16), and susceptibility to ozone of five inbred strains of mice. Inhalation Toxicol.
15: 1209-1230.
8-83
-------
Brook, R. D.; Brook, J. R.; Urch, B.; Vincent, R.; Rajagopalan, S.; Silverman, F. (2002) Inhalation of fine paniculate
air pollution and ozone causes acute arterial vasoconstriction in healthy adults. Circulation 105: 1534-1536.
Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997a) Association between ozone and
hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72: 24-31.
Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997b) The role of paniculate size and chemistry in the
association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
Environ. Health Perspect. 105: 614-620.
Burnett, R. T.; Smith-Doiron, M; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of paniculate and gaseous air
pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54: 130-139.
Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.; Leech, J. A.; Cakmak, S.;
Krewski, D. (2001) Association between ozone and hospitalization for acute respiratory diseases in children
less than 2 years of age. Am. J. Epidemiol. 153: 444-452.
Bush, M. L.; Zhang, W.; Ben-Jebria, A.; Ultman, J. S. (2001) Longitudinal distribution of ozone and chlorine in the
human respiratory tract: simulation of nasal and oral breathing with the single-path diffusion model. Toxicol.
Appl. Pharmacol. 173: 137-145.
Cassino, C.; Ito, K.; Bader, L; Ciotoli, C.; Thurston, G.; Reibman, J. (1999) Cigarette smoking and ozone-associated
emergency department use for asthma by adults in New York City. Am. J. Respir. Crit. Care Med.
159: 1773-1779.
Castillejos, M.; Gold, D. R.; Damokosh, A. I.; Serrano, P.; Allen, G.; McDonnell, W. F.; Dockery, D.; Velasco,
S. R.; Hernandez, M.; Hayes, C. (1995) Acute effects of ozone on the pulmonary function of exercising
schoolchildren from Mexico City. Am. J. Respir. Crit. Care Med. 152: 1501-1507.
Cho, H.-Y.; Zhang, L.-Y.; Kleeberger, S. R. (2001) Ozone-induced lung inflammation and hyperreactivity are
mediated via tumor necrosis factor-a receptors. Am. J. Physiol. 280: L537-L546.
Chock, D. P.; Winkler, S. L.; Chen, C. (2000) A study of the association between daily mortality and ambient air
pollutant concentrations in Pittsburgh, Pennsylvania. J. Air Waste Manage. Assoc. 50: 1481-1500.
Cohen, M. D.; Sisco, M.; Li, Y.; Zelikoff, J. T.; Schlesinger, R. B. (2001) Ozone-induced modulation of
cell-mediated immune responses in the lungs. Toxicol. Appl. Pharmacol. 171: 71-84.
Cohen, M. D.; Sisco, M.; Baker, K.; Li, Y.; Lawrence, D.; VanLoveren, H.; Zelikoff, J. T.; Schlesinger, R. B.
(2002) Effects of inhaled ozone on pulmonary immune cells critical to antibacterial responses in situ.
Inhalation Toxicol. 14: 599-619.
Coleridge, J. C. G.; Coleridge, H. M.; Schelegle, E. S.; Green, J. F. (1993) Acute inhalation of ozone stimulates
bronchial C-fibers and rapidly adapting receptors in dogs. J. Appl. Physiol. 74: 2345-2352.
Corradi, M.; Alinovi, R.; Goldoni, M.; Vettori, M.; Folesani, G.; Mozzoni, P.; Cavazzini, S.; Bergamaschi, E.;
Rossi, L.; Mutti, A. (2002) Biomarkers of oxidative stress after controlled human exposure to ozone.
Toxicol. Lett. 134: 219-225.
David, G. L.; Romieu, I.; Sienra-Monge, J. J.; Collins, W. J.; Ramirez-Aguilar, M.; Del Rio-Navarro, B. E.;
Reyes-Ruiz, N. I.; Morris, R. W.; Marzec, J. M.; London, S. J. (2003) Nicotinamide adenine dinucleotide
(phosphate) reduced:quinone oxidoreductase and glutathione s-transferase ml polymorphism and childhood
asthma. Am. J. Respir. Crit. Care Med. 168: 1199-1204.
DeFrances, C. J.; Hall, M. J.; Podgornik, M. N. (2005) 2003 National Hospital Discharge Survey. Hyattsville, MD:
National Center for Health Statistics; DHHS publication no. (PHS) 2004-1250. (Advance data from vital and
health statistics; no. 359). Available: http://www.cdc.gov/nchs/data/ad/ad359.pdf [3 August, 2005].
Delaunois, A.; Segura, P.; Montano, L. M.; Vargas, M. H.; Ansay, M.; Gustin, P. (1998) Comparison of
ozone-induced effects on lung mechanics and hemodynamics in the rabbit. Toxicol. Appl. Pharmacol.
150: 58-67.
Delfino, R. J.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Matteucci, R. M.; Anderson, P. R.; Koutrakis, P. (1997)
The effect of outdoor fungal spore concentrations on daily asthma severity. Environ. Health Perspect.
105: 622-635.
Depuydt, P.; Joos, G. F.; Pauwels, R. A. (1999) Ambient ozone concentrations induce airway hyperresponsiveness
in some rat strains. Eur. Respir. J. 14: 125-131.
Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the
lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
Devlin, R. B.; Folinsbee, L. J.; Biscardi, F.; Hatch, G.; Becker, S.; Madden, M. C.; Robbins, M.; Koren, H. S. (1997)
Inflammation and cell damage induced by repeated exposure of humans to ozone. Inhalation Toxicol.
9:211-235.
8-84
-------
Dey, A. N.; Bloom, B. (2005) Summary health statistics for U.S. children: National Health Interview Survey, 2003.
Hyattsville, MD: U.S. Department of Health & Human Services, National Center for Health Statistics. (Vital
and health statistics, series 10, no. 223). Available: http://www.cdc.gov/nchs/data/series/sr_10/srlO_223.pdf
[3 August, 2005].
Dietert, R. R.; Etzel, R. A.; Chen, D.; Halonen, M; Holladay, S. D.; Jarabek, A. M; Landreth, K.; Peden, D. B.;
Pinkerton, K.; Smialowicz, R. I; Zoetis, T. (2000) Workshop to identify critical window of exposure for
children's health: immune and respiratory systems work group summary. Environ. Health Perspect. Suppl.
108(3): 483-490.
Dockery, D. W.; Schwartz, I; Spengler, J. D. (1992) Air pollution and daily mortality: associations with particulates
and acid aerosols. Environ. Res. 59: 362-373.
Dominici, F.; McDermott, A.; Daniels, M.; Zeger, S. L.; Samet, J. M. (2003) Mortality among residents of 90 cities.
In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA: Health
Effects Institute; pp. 9-24. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
Dong, W.; Selgrade, M. K.; Gilmour,M. L; Lange, R. W.; Park, P.; Luster, M. L; Kari, F. W. (1998) Altered alveolar
macrophage function in calorie-restricted rats. Am. J. Respir. Cell Mol. Biol. 19: 462-469.
Dormans, J. A. M. A.; VanBree, L.; Boere, A. J. F.; Marra, M.; Rombout, P. J. A. (1999) Interspecies differences in
time course of pulmonary toxicity following repeated exposure to ozone. Inhalation Toxicol. 11: 309-329.
Evans, M. J.; Fanucchi, M. V.; Baker, G. L.; Van Winkle, L. S.; Pantle, L. M.; Nishio, S. J.; Schelegle, E. S.;
Gershwhin, L. J.; Miller, L. A.; Hyde, D. M.; Sannes, P. L.; Plopper, C. G. (2003) Atypical development of
the trachea! basement membrane zone of infant rhesus monkeys exposed to ozone and allergen. Am. J.
Physiol. 285:L931-L939.
Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
Perspect. 107: 637-641.
Fairley, D. (2003) Mortality and air pollution for Santa Clara County, California, 1989-1996. In: Revised analyses of
time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute;
pp. 97-106. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdftl8 October, 2004].
Fiore, A. M.; Jacob, D. J.; Mathur, R.; Martin, R. V. (2003) Application of empirical orthogonal functions
to evaluate ozone simulations with regional and global models. J. Geophys. Res. (Atmos.)
108(D14): 10.1029/2002JD003151.
Folinsbee, L. J.; Horstman, D. H.; Kehrl, H. R.; Harder, S.; Abdul-Salaam, S.; Ives, P. J. (1994) Respiratory
responses to repeated prolonged exposure to 0.12 ppm ozone. Am. J. Respir. Crit. Care Med. 149: 98-105.
Foster, W. M.; Freed, A. N. (1999) Regional clearance of solute from peripheral airway epithelia: recovery after
sublobar exposure to ozone. J. Appl. Physiol. 86: 641-646.
Foster, W. M.; Stetkiewicz, P. T. (1996) Regional clearance of solute from the respiratory epithelia: 18-20 h
postexposure to ozone. J. Appl. Physiol. 81: 1143-1149.
Foster, W. M.; Silver, J. A.; Groth, M. L. (1993) Exposure to ozone alters regional function and particle dosimetry in
the human lung. J. Appl. Physiol. 75: 1938-1945.
Foster, W. M.; Weinmann, G. G.; Menkes, E.; Macri, K. (1997) Acute exposure of humans to ozone impairs small
airway function. Ann. Occup. Hyg. 41(suppl. 1): 659-666.
Friedman, M. S.; Powell, K. E.; Hutwagner, L.; Graham, L. M.; Teague, W. G. (2001) Impact of changes in
transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality and
childhood asthma. JAMA J. Am. Med. Assoc. 285: 897-905.
Gamble, J. L. (1998) Effects of ambient air pollution on daily mortality: a time series analysis of Dallas, Texas,
1990-1994. Presented at: 91st annual meeting and exhibition of the Air & Waste Management Association;
June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 98-MP26.03.
Gauderman, W. J.; Avol, E.; Gilliland, F.; Vora, H.; Thomas, D.; Berhane, K.; McConnell, R.; Kuenzli, N.;
Lurmann, F.; Rappaport, E.; Margolis, H.; Bates, D.; Peters, J. (2004) The effect of air pollution on lung
development from 10 to 18 years of age. N. Engl. J. Med. 351: 1057-1067.
Gent, J. F.; Triche, E. W.; Holford, T. R.; Belanger, K.; Bracken, M. B.; Beckett, W. S.; Leaderer, B. P. (2003)
Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA
J. Am. Med. Assoc. 290: 1859-1867.
Goldsmith, C.-A. W.; Ning, Y.-Y.; Qin, G.; Imrich, A.; Lawrence, J.; Murthy, G. G., K.; Catalano, P. J.; Kobzik, L.
(2002) Combined air pollution particle and ozone exposure increases airway responsiveness in mice.
Inhalation Toxicol. 14: 325-347.
Gong, H., Jr.; Shamoo, D. A.; Anderson, K. R.; Linn, W. S. (1997a) Responses of older men with and without
chronic obstructive pulmonary disease to prolonged ozone exposure. Arch. Environ. Health 52: 18-25.
8-85
-------
Gong, H., Jr.; McManus, M. S.; Linn, W. S. (1997b) Attenuated response to repeated daily ozone exposures in
asthmatic subjects. Arch. Environ. Health 52: 34-41.
Gong, H., Jr.; Wong, R.; Sarma, R. J.; Linn, W. S.; Sullivan, E. D.; Shamoo, D. A.; Anderson, K. R.; Prasad, S. B.
(1998) Cardiovascular effects of ozone exposure in human volunteers. Am. J. Respir. Crit. Care Med.
158: 538-546.
Gwynn, R. C.; Thurston, G. D. (2001) The burden of air pollution: impacts among racial minorities. Environ. Health
Perspect. Suppl. 109(4): 501-506.
Gwynn, R. C.; Burnett, R. T.; Thurston, G. D. (2000) A time-series analysis of acidic paniculate matter and daily
mortality and morbidity in the Buffalo, New York, region. Environ. Health Perspect. 108: 125-133.
Hanania, N. A.; Tarlo, S. M.; Silverman, F.; Urch, B.; Senathirajah, N.; Zamel, N.; Corey, P. (1998) Effect of
exposure to low levels of ozone on the response to inhaled allergen in allergic asthmatic patients. Chest
114:752-756.
Harkema, J. R.; Wagner, J. G. (2005) Epithelial and inflammatory responses in the airways of laboratory rats
coexposed to ozone and biogenic substances: enhancement of toxicant-induced airway injury. Exp. Toxicol.
Pathol. 57(suppl. 1): 129-141.
Hatch, G. E.; Slade, R.; Harris, L. P.; McDonnell, W. F.; Devlin, R. B.; Koren, H. S.; Costa, D. L.; McKee, J. (1994)
Ozone dose and effect in humans and rats: a comparison using oxygen-18 labeling and bronchoalveolar
lavage. Am. J. Respir. Crit. Care Med. 150: 676-683.
Hazucha, M. J.; Sant'Ambrogio, G. (1993) Effects of ozone on the activity of slowly (SAR) and rapidly adapting
(RAR) receptors in cats. FASEB J. 7: 407A.
Hazucha, M. J.; Folinsbee, L. J.; Seal, E., Jr. (1992) Effects of steady-state and variable ozone concentration profiles
on pulmonary function. Am. Rev. Respir. Dis. 146: 1487-1493.
Hazucha, M. J.; Folinsbee, L. J.; Bromberg, P. A. (2003) Distribution and reproducibility of spirometric response to
ozone by gender and age. J. Appl. Physiol. 95: 1917-1925.
Higgins, I. T. T.; D'Arcy, J. B.; Gibbons, D. I.; Avol, E. L.; Gross, K. B. (1990) Effect of exposures to ambient
ozone on ventilatory lung function in children. Am. Rev. Respir. Dis. 141: 1136-1146.
Hiltermann, J. T. N.; Lapperre, T. S.; VanBree, L.; Steerenberg, P. A.; Brahim, J. J.; Sont, J. K.; Sterk, P. J.;
Hiemstra, P. S.; Stolk, J. (1999) Ozone-induced inflammation assessed in sputum and bronchial lavage fluid
from asthmatics: a new noninvasive tool in epidemiologic studies on air pollution and asthma. Free Radical
Biol. Med. 27: 1448-1454.
Holz, O.; Mticke, M.; Paasch, K.; Bohme, S.; Timm, P.; Richter, K.; Magnussen, H.; Jorres, R. A. (2002) Repeated
ozone exposures enhance bronchial allergen responses in subjects with rhinitis or asthma. Clin. Exp. Allergy.
32: 681-689.
Hoppe, P.; Praml, G.; Rabe, G.; Lindner, J.; Fruhmann, G.; Kessel, R. (1995) Environmental ozone field study on
pulmonary and subjective responses of assumed risk groups. Environ. Res. 71: 109-121.
Hoppe, P.; Peters, A.; Rabe, G.; Praml, G.; Lindner, J.; Jakobi, G.; Fruhmann, G.; Nowak, D. (2003) Environmental
ozone effects in different population subgroups. Int. J. Hyg. Environ. Health 206: 505-516.
Horstman, D. H.; Folinsbee, L. J.; Ives, P. J.; Abdul-Salaam, S.; McDonnell, W. F. (1990) Ozone concentration and
pulmonary response relationships for 6.6-hour exposures with five hours of moderate exercise to 0.08, 0.10,
and 0.12 ppm. Am. Rev. Respir. Dis. 142: 1158-1163.
Huang, Y.; Dominici, F.; Bell, M. L. (2005) Bayesian hierarchical distributed lag models for summer ozone
exposure and cardio-respiratory mortality. Environmetrics 16: 547-562.
Ito, K. (2003) Associations of paniculate matter components with daily mortality and morbidity in Detroit,
Michigan. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
Health Effects Institute; pp. 143-156. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
[12 May, 2004].
Ito, K.; Thurston, G. D. (1996) Daily PM10/mortality associations: an investigation of at-risk subpopulations.
J. Exposure Anal. Environ. Epidemiol. 6: 79-95.
Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and
meta-analysis. Epidemiology 16: 446-457.
Jaffe, D. H.; Singer, M. E.; Rimm, A. A. (2003) Air pollution and emergency department visits for asthma among
Ohio Medicaid recipients, 1991-1996. Environ. Res. 91: 21-28.
Jenkins, H. S.; Devalia, J. L.; Mister, R. L.; Bevan, A. M.; Rusznak, C.; Davies, R. J. (1999) The effect of exposure
to ozone and nitrogen dioxide on the airway response of atopic asthmatics to inhaled allergen: dose- and
time-dependent effects. Am. J. Respir. Crit. Care Med. 160: 33-39.
8-86
-------
Torres, R.; Nowak, D.; Magnussen, H.; Speckin, P.; Koschyk, S. (1996) The effect of ozone exposure on allergen
responsiveness in subjects with asthma or rhinitis. Am. J. Respir. Crit. Care Med. 153: 56-64.
Torres, R. A.; Holz, O.; Zachgo, W.; Timm, P.; Koschyk, S.; Mtiller, B.; Grimminger, F.; Seeger, W.; Kelly, F. J.;
Dunster, C.; Frischer, T.; Lubec, G.; Waschewski, M; Niendorf, A.; Magnussen, H. (2000) The effect of
repeated ozone exposures on inflammatory markers inbronchoalveolar lavage fluid and mucosal biopsies.
Am. J. Respir. Crit. Care Med. 161: 1855-1861.
Kehrl, H. R.; Vincent, L. M.; Kowalsky, R. J.; Horstman, D. H.; O'Neil, J. J.; McCartney, W. H.; Bromberg, P. A.
(1987) Ozone exposure increases respiratory epithelial permeability in humans. Am. Rev. Respir. Dis.
135: 1124-1128.
Kelly, F. J.; Dunster, C.; Mudway, I. (2003) Air pollution and the elderly: oxidant/antioxidant issues worth
consideration. Eur. Respir. J. Suppl. 40: 70S-75S.
Kim, S.-Y.; Lee, J.-T.; Hong, Y.-C.; Ahn, K.-J.; Kim, H. (2004) Determining the threshold effect of ozone on daily
mortality: an analysis of ozone and mortality in Seoul, Korea, 1995-1999. Environ. Res. 94: 113-119.
Kinney, P. L.; Ozkaynak, H. (1991) Associations of daily mortality and air pollution in Los Angeles County.
Environ. Res. 54: 99-120.
Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
Kinney, P. L.; Thurston, G. D.; Raizenne, M. (1996) The effects of ambient ozone on lung function in children:
a reanalysis of six summer camp studies. Environ. Health Perspect. 104: 170-174.
Kleeberger, S. R.; Levitt, R. C.; Zhang, L.-Y.; Longphre, M.; Harkema, J.; Jedlicka, A.; Eleff, S. M.; DiSilvestre, D.;
Holroyd, K. J. (1997) Linkage analysis of susceptibility to ozone-induced lung inflammation in inbred mice.
Nat. Genet. 17: 475-478.
Kleeberger, S. R.; Ohtsuka, Y.; Ahang, L.-Y.; Longphre, M. (2001) Airway responses to chronic ozone exposure are
partially mediated through mast cells. J. Appl. Physiol. 90: 713-723.
Kleinman, M. T.; Bufalino, C.; Rasmussen, R.; Hyde, D.; Bhalla, D. K.; Mautz, W. J. (2000) Toxicity of chemical
components of ambient fine paniculate matter (PM2.5) inhaled by aged rats. J. Appl. Toxicol. 20: 357-364.
Klemm, R. J.; Lipfert, F. W.; Wyzga, R. E.; Gust, C. (2004) Daily mortality and air pollution in Atlanta: two years
of data from ARIES. Inhalation Toxicol. 16(suppl. 1): 131-141.
Kochanek, K. D.; Murphy, S. L.; Anderson, R. N.; Scott, C. (2004) Deaths: final data for 2002. Hyattsville, MD:
U.S. Department of Health & Human Services, National Center for Health Statistics; DHHS publication no.
(PHS) 2005-1120. (National vital statistics reports: v. 53, no. 5). Available:
http://www.cdc.gov/nchs/data/nvsr/nvsr53/nvsr53_05.pdf [3 August, 2005].
Koenig, J. Q.; Covert, D. S.; Smith, M. S.; Van Belle, G.; Pierson, W. E. (1988) The pulmonary effects of ozone and
nitrogen dioxide alone and combined in healthy and asthmatic adolescent subjects. Toxicol. Ind. Health
4: 521-532.
Koenig, J. Q.; Covert, D. S.; Hanley, Q. S.; Van Belle, G.; Pierson, W. E. (1990) Prior exposure to ozone potentiates
subsequent response to sulfur dioxide in adolescent asthmatic subjects. Am. Rev. Respir. Dis. 141: 377-380.
Kopp, M. V.; Ulmer, C.; Ihorst, G.; Seydewitz, H. H.; Frischer, T.; Forster, J.; Kuehr, J. (1999) Upper airway
inflammation in children exposed to ambient ozone and potential signs of adaptation. Eur. Respir. J.
14: 854-861.
Korrick, S. A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K. D.; Rosner, B. A.;
Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
Environ. Health Perspect. 106: 93-99.
Krishna, M. T.; Blomberg, A.; Biscione, G. L.; Kelly, F.; Sandstrom, T.; Frew, A.; Holgate, S. (1997) Short-term
ozone exposure upregulates P-selectin in normal human airways. Am. J. Respir. Crit. Care Med.
155: 1798-1803.
Leikauf, G. D.; Simpson, L. G.; Santrock, J.; Zhao, Q.; Abbinante-Nissen, J.; Zhou, S.; Driscoll, K. E. (1995)
Airway epithelial cell responses to ozone injury. Environ. Health Perspect. 103(suppl. 2): 91-95.
Lethbridge-Cejku, M.; Schiller, J. S.; Bernadel, L. (2004) Summary health statistics for U.S. adults: National Health
Interview Survey, 2002. Hyattsville, MD: Centers for Disease Control and Prevention; DHHS publication no.
(PHS) 2004-1550. (Vital and health statistics, series 10, number 222). Available:
http://www.cdc.gov/nchs/data/series/sr_10/srlO_222.pdf [3, August, 2005].
Levy, J. I.; Chemerynski, S. M.; Sarnat, J. A. (2005) Ozone exposure and mortality, an empiric Bayes
metaregression analysis. Epidemiology 16: 458-468.
8-87
-------
Lin, M; Chen, Y.; Burnett, R. T.; Villeneuve, P. J.; Krewski, D. (2003) Effect of short-term exposure to gaseous
pollution on asthma hospitalisation in children: a bi-directional case-crossover analysis. J. Epidemiol.
Community Health 57: 50-55.
Lin, M.; Chen, Y.; Villeneuve, P. J.; Burnett, R. T.; Lemyre, L.; Hertzman, C.; McGrail, K. M.; Krewski, D. (2004)
Gaseous air pollutants and asthma hospitalization of children with low household income in Vancouver,
British Columbia, Canada. Am. J. Epidemiol. 159: 294-303.
Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996)
Short-term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal.
Environ. Epidemiol. 6: 449-472.
Linn, W. S.; Szlachcic, Y.; Gong, H., Jr.; Kinney, P. L.; Berhane, K. T. (2000) Air pollution and daily hospital
admissions in metropolitan Los Angeles. Environ. Health Perspect. 108: 427-434.
Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000) Daily mortality in the Philadelphia metropolitan area and
size-classified paniculate matter. J. Air Waste Manage. Assoc. 50: 1501-1513.
Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no.
95.
Luginaah, I. N.; Fung, K. Y.; Gorey, K. M.; Webster, G.; Wills, C. (2005) Association of ambient air pollution with
respiratory hospitalization in a government designated "area of concern": the case of Windsor, Ontario.
Environ. Health Perspect. 113: 290-296.
Madden, M. C.; Richards, J. H.; Dailey, L. A.; Hatch, G. E.; Ohio, A. J. (2000) Effect of ozone on diesel exhaust
particle toxicity in rat lung. Toxicol. Appl. Pharmacol. 168: 140-148.
Mannino, D. M.; Ford, E. S.; Redd, S. C. (2003) Obstructive and restrictive lung disease and markers of
inflammation: data from the Third National Health and Nutrition Examination. Am. J. Med. 114: 758-762.
McCaig, L. F.; Burt, C. W. (2005) National Hospital Ambulatory Medical Care Survey: 2003 Emergency
Department Summary. Hyattsville, MD: National Center for Health Statistics; DHHS publication no. (PHS)
2005-1250. (Advance data from vital and health statistics; no. 358). Available:
http://www.cdc.gov/nchs/data/ad/ad358.pdft3 August, 2005].
McConnell, R.; Berhane, K.; Gilliland, F.; London, S. J.; Islam, T.; Gauderman, W. J.; Avol, E.; Margolis, H. G.;
Peters, J. M. (2002) Asthma in exercising children exposed to ozone: a cohort study. Lancet 359: 386-391.
McDonnell, W. F. (1996) Individual variability in human lung function responses to ozone exposure. Environ.
Toxicol. Pharmacol. 2: 171-175.
McDonnell, W. F.; Stewart, P. W.; Andreoni, S.; Seal, E., Jr.; Kehrl, H. R.; Horstman, D. H.; Folinsbee, L. J.; Smith,
M. V. (1997) Prediction of ozone-induced FEVj changes: effects of concentration, duration, and ventilation.
Am. J. Respir. Crit. Care Med. 156: 715-722.
Michelson, P. H.; Dailey, L.; Devlin, R. B.; Peden, D. B. (1999) Ozone effects on the immediate-phase response to
allergen in the nasal airways of allergic asthmatic subjects. Otolaryngol. Head Neck Surg. 120: 225-232.
Moolgavkar, S. H. (2003) Air pollution and daily mortality in two U.S. counties: season-specific analyses and
exposure-response relationships. Inhalation Toxicol. 15: 877-907.
Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A.; Anderson, E. L. (1995) Air pollution and daily mortality in
Philadelphia. Epidemiology 6: 476-484.
Mortimer, K. M.; Tager, I. B.; Dockery, D. W.; Neas, L. M.; Redline, S. (2000) The effect of ozone on inner-city
children with asthma: identification of susceptible subgroups. Am. J. Respir. Crit. Care Med. 162: 1838-1845.
Mortimer, K. M.; Neas, L. M.; Dockery, D. W.; Redline, S.; Tager, I. B. (2002) The effect of air pollution on
inner-city children with asthma. Eur. Respir. J. 19: 699-705.
Mudway, I. S.; Kelly, F. J. (2000) Ozone and the lung: a sensitive issue. Mol. Aspects. Med. 21: 1-48.
Mudway, I. S.; Kelly, F. J. (2004) An investigation of inhaled ozone dose and the magnitude of airway inflammation
in healthy adults. Am. J. Respir. Crit. Care Med. 169: 1089-1095.
Naeher, L. P.; Holford, T. R.; Beckett, W. S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
Healthy women's PEF variations with ambient summer concentrations of PM10, PM2 5, SO42", H+, and O3.
Am. J. Respir. Crit. Care Med. 160: 117-125.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol.
141: 111-122.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Speizer, F. E. (1999) Fine particles and peak flow in children: acidity
versus mass. Epidemiology 10: 550-553.
-------
Ostro, B. (1995) Fine paniculate air pollution and mortality in two Southern California counties. Environ. Res.
70: 98-104.
Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coachella
Valley, California: a follow-up study. J. Exposure Anal. Environ. Epidemiol. 10: 412-419.
Passannante, A. N.; Hazucha, M. J.; Bromberg, P. A.; Seal, E.; Folinsbee, L.; Koch, G. (1998) Nociceptive
mechanisms modulate ozone-induced human lung function decrements. J. Appl. Physiol. 85: 1863-1870.
Peel, J. L.; Tolbert, P. E.; Klein, M.; Metzger, K. B.; Flanders, W. D.; Knox, T.; Mulholland, J. A.; Ryan, P. B.;
Frumkin, H. (2005) Ambient air pollution and respiratory emergency department visits. Epidemiology
16: 164-174.
Petroeschevsky, A.; Simpson, R. W.; Thalib, L.; Rutherford, S. (2001) Associations between outdoor air pollution
and hospital admissions in Brisbane, Australia. Arch. Environ. Health 56: 37-52.
Plunkett, L. M.; Turnbull, D.; Rodricks, J. V. (1992) Differences between adults and children affecting exposure
assessment. In: Guzelian, P. S.; Henry, D. J.; Olin, S. S., eds. Similarities and differences between children
and adults: implications for risk assessment. Washington, DC: ILSI Press, pp. 79-96.
Ponce de Leon, A.; Anderson, H. R.; Bland, J. M.; Strachan, D. P.; Bower, J. (1996) Effects of air pollution on daily
hospital admissions for respiratory disease in London between 1987-88 and 1991-92. In: StLeger, S., ed.
The APHEA project. Short term effects of air pollution on health: a European approach using epidemiological
time series data. J. Epidemiol. Commun. Health 50(suppl. 1): S63-S70.
Pope, C. A., Ill; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA J. Am. Med.
Assoc. 287: 1132-1141.
Pulfer, M. K.; Murphy, R. C. (2004) Formation of biologically active oxysterols during ozonolysis of cholesterol
present in lung surfactant. J. Biol. Chem. 279: 26,331-26,338.
Pulfer, M. K.; Taube, C.; Gelfand, E.; Murphy, R. C. (2005) Ozone exposure in vivo and formation of biologically
active oxysterols in the lung. J. Pharmacol. Exp. Ther. 312: 256-264.
Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. (1987) Acute respiratory function and transported air pollutants:
observational studies. Presented at: 80th annual meeting of the Air Pollution Control Association; June;
New York, NY. Pittsburgh, PA: Air Pollution Control Association; paper no. 87-32.6.
Raizenne, M. E.; Burnett, R. T.; Stern, B.; Franklin, C. A.; Spengler, J. D. (1989) Acute lung function responses to
ambient acid aerosol exposures in children. Environ. Health Perspect. 79: 179-185.
Rigas, M. L.; Ben-Jebria, A.; Ultman, J. S. (1997) Longitudinal distribution of ozone absorption in the lung: effects
of nitrogen dioxide, sulfur dioxide, and ozone exposures. Arch. Environ. Health 52: 173-178.
Romieu, I.; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. I.; Estela del
Rio-Navarro, B.; Hernandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and
antioxidant supplementation influence lung function in relation to ozone exposure in asthmatic children in
Mexico City. Thorax 59: 8-10.
Ross, M. A.; Persky, V. W.; Scheff, P. A.; Chung, J.; Curtis, L.; Ramakrishnan, V.; Wadden, R. A.; Hryhorczuk,
D. O. (2002) Effect of ozone and aeroallergens on the respiratory health of asthmatics. Arch. Environ. Health
57: 568-578.
Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, L; Dockery, D. W.; Schwartz, J.; Zanobetti, A.
(2000) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94, part II.
Sathishkumar, K.; Haque, M.; Perumal, T. E.; Francis, J.; Uppu, R. M. (2005) A major ozonation product of
cholesterol, 3p-hydroxy-5-oxo-5,6-secocholestan-6-al, induces apoptosis inH9c2 cardiomyoblasts. FEES
Lett. 579: 6444-6450.
Scannell, C.; Chen, L.; Aris, R. M.; Tager, I.; Christian, D.; Ferrando, R.; Welch, B.; Kelly, T.; Balmes, J. R. (1996)
Greater ozone-induced inflammatory responses in subjects with asthma. Am. J. Respir. Crit. Care Med.
154: 24-29.
Schelegle, E. S.; Walby, W. F.; Alfaro, M. F.; Wong, V. J.; Putney, L.; Stovall, M. Y.; Sterner-Kock, A.; Hyde,
D. M.; Plopper, C. G. (2003) Repeated episodes of ozone inhalation attenuates airway injury/repair and
release of substance P, but not adaptation. Toxicol. Appl. Pharmacol. 186: 127-142.
Schierhorn, K.; Zhang, M.; Matthias, C.; Kunkel, G. (1999) Influence of ozone and nitrogen dioxide on histamine
and interleukin formation in a human nasal mucosa culture system. Am. J. Respir. Cell Mol. Biol.
20: 1013-1019.
Schierhorn, K.; Hanf, G.; Fischer, A.; Umland, B.; Olze, H.; Kunkel, G. (2002) Ozone-induced release of
neuropeptides from human nasal mucosa cells. Int. Arch. Allergy Immunol. 129: 145-151.
8-89
-------
Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease. Epidemiology 7: 20-28.
Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature?
Am. J. Respir. Crit. Care Med. 171: 627-631.
Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.; Ponce de
Leon, A.; Ponka, A.; Rossi, G.; Saez, M; Schouten, J. P. (1996) Methodological issues in studies of air
pollution and daily counts of deaths or hospital admissions. In: St Leger, S., ed. The APHEA project.
Short term effects of air pollution on health: a European approach using epidemiological time series data.
J. Epidemiol. Commun. Health 50(suppl. 1): S3-S11.
Segura, P.; Montano, L. M.; Bazan-Perkins, B.; Gustin, P.; Vargas, M. H. (1997) Ozone at high-pollution urban
levels causes airway hyperresponsiveness to substance P but not to other agonists. Environ. Toxicol.
Pharmacol. 3: 91-95.
Sexton, K. G.; Jeffries, H. E.; Jang, M.; Kamens, R. M.; Doyle, M.; Voicu, L; Jaspers, I. (2004) Photochemical
products in urban mixtures enhance inflammatory responses in lung cells. Inhalation Toxicol.
16(suppl. 1): 107-114.
Sheppard, L. (2003) Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington,
1987-1994. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston,
MA: Health Effects Institute; pp. 227-230. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
[18 October, 2004].
Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air pollution on
nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
Sin, D. D.; Man, S. F. P. (2003) Why are patients with chronic obstructive pulmonary disease at increased risk of
cardiovascular diseases? Circulation 107: 1514-1519.
Sin, D. D.; Wu, L.-L.; Man, S. F. P. (2005) The relationship between reduced lung function and cardiovascular
mortality: a population-based study and a systematic review of the literature. Chest 127: 1952-1959.
Solway, J.; Leff, A. R. (1991) Sensory neuropeptides and airway function. J. Appl. Physiol. 71: 2077-2087.
Spannhake, E. W.; Reddy, S. P. M.; Jacoby, D. B.; Yu, X.-Y.; Saatian, B.; Tian, J. (2002) Synergismbetween
rhinovirus infection and oxidant pollutant exposure enhances airway epithelial cell cytokine production.
Environ. Health Perspect. 110: 665-670.
Spektor, D. M.; Lippmann, M.; Lioy, P. J.; Thurston, G. D.; Citak, K.; James, D. J.; Bock, N.; Speizer, F. E.;
Hayes, C. (1988) Effects of ambient ozone on respiratory function in active, normal children. Am. Rev.
Respir. Dis. 137:313-320.
Spektor, D. M.; Lippmann, M. (1991) Health effects of ambient ozone on healthy children at a summer camp. In:
Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
Association; pp. 83-89. (A&WMA transactions series no. TR-19).
Sprenger, M.; Maspoli, M. C.; Wernli, H. (2003) Tropopause folds and cross-tropopause exchange: a global
investigation based upon ECMWF analyses for the time period March 2000 to February 2001. J. Geophys.
Res. (Atmos.) 108(D12): 10.1029/2002JD002587.
Stieb, D. M.; Burnett, R. T.; Beveridge, R. C.; Brook, J. R. (1996) Association between ozone and asthma
emergency department visits in Saint John, New Brunswick, Canada. Environ. Health Perspect.
104: 1354-1360.
Stohl, A. (2001) A 1-yearLagrangian "climatology" of airstreams in the Northern Hemisphere troposphere and
lowermost stratosphere. J. Geophys. Res. (Atmos.) 106: 7263-7279.
Tenias, J. M.; Ballester, F.; Rivera, M. L. (1998) Association between hospital emergency visits for asthma and air
pollution in Valencia, Spain. Occup. Environ. Med. 55: 541-547.
Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. (1997) Summertime haze air pollution and children with
asthma. Am. J. Respir. Crit. Care Med. 155: 654-660.
Tolbert, P. E.; Mulholland, J. A.; Macintosh, D. L.; Xu, F.; Daniels, D.; Devine, O. J.; Carlin, B. P.; Klein, M.;
Dorley, J.; Butler, A. J.; Nordenberg, D. F.; Frumkin, H.; Ryan, P. B.; White, M. C. (2000) Air quality and
pediatric emergency room visits for asthma in Atlanta, Georgia. Am. J. Epidemiol. 151: 798-810.
Trenga, C. A.; Koenig, J. Q.; Williams, P. V. (2001) Dietary antioxidants and ozone-induced bronchial
hyperresponsiveness in adults with asthma. Arch. Environ. Health 56: 242-249.
Tuazon, E. C.; Winer, A. M.; Pitts, J. N., Jr. (1981) Trace pollutant concentrations in a multiday smog episode in the
California South Coast Air Basin by long path length Fourier transform infrared spectroscopy. Environ. Sci.
Technol. 15: 1232-1237.
8-90
-------
U.S. Environmental Protection Agency. (1990) National air quality and emissions trends report, 1988. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/450/4-90/002.
U.S. Environmental Protection Agency. (1996a) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
http://cfpub2.epa.gov/ncea/.
U.S. Environmental Protection Agency. (1996b) Review of national ambient air quality standards for ozone:
assessment of scientific and technical information. OAQPS staff paper. Research Triangle Park, NC:
Office of Air Quality Planning and Standards; report no. EPA/452/R-96/007. Available from: NTIS,
Springfield, VA; PB96-203435. Available: http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_pr_sp.html
(9 September 2003).
U.S. Environmental Protection Agency. (2003) National air quality and emissions trends report. 2003 special studies
edition. Research Triangle Park, NC: Office of Air Quality Standards; Emissions Monitoring and Analysis
Division; report no. EPA454/R-03-005. Available: http://www.epa.gov/air/airtrends/aqtrnd03/toc.html
(27 August, 2004).
U.S. Environmental Protection Agency. (2004) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
http://cfpub.epa.gov/ncea/ [9 November, 2004].
Ultman, J. S.; Ben-Jebria, A.; Arnold, S. F. (2004) Uptake distribution of ozone in human lungs: intersubject
variability in physiologic response. Boston, MA: Health Effects Institute.
Vagaggini, B.; Taccola, M.; Clanchetti, S.; Carnevali, S.; Bartoli, M. L.; Bacci, E.; Dente, F. L.; Di Franco, A.;
Giannini, D.; Paggiaro, P. L. (2002) Ozone exposure increases eosinophilic airway response induced by
previous allergen challenge. Am. J. Respir. Crit. Care Med. 166: 1073-1077.
Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
pollution. Environ. HealthPerspect. Ill: 45-51.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994a) Ozone increases amino- and
carboxy-terminal atrial natriuretic factor prohormone peptides in lung, heart, and circulation. J. Biochem.
Toxicol. 9: 107-112.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994b) Increase in atrial natriuretic factor in
the lungs, heart, and circulatory system owing to ozone. Chest 105: 1551-1554.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994c) Ozone increases atrial natriuretic
peptides in heart, lung and circulation of aged vs. adult animals. Gerontology (Basel) 40: 227-236.
Villeneuve, P. J.; Burnett, R. T.; Shi, Y.; Krewski, D.; Goldberg, M. S.; Hertzman, C.; Chen, Y.; Brook, J. (2003)
A time-series study of air pollution, socioeconomic status, and mortality in Vancouver, Canada. J. Exposure
Anal. Environ. Epidemiol. 13: 427-435.
Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of submicron
particle exposure. Environ. Health Perspect. 108: 1139-1145.
Wattiez, R.; Noel-Georis, L; Cruyt, C.; Broeckaert, F.; Bernard, A.; Falmagne, P. (2003) Susceptibility to oxidative
stress: proteomic analysis of bronchoalveolar lavage from ozone-sensitive and ozone-resistant strains of mice.
Proteomics 3: 658-665.
Wayne, R. P. (1991) Chemistry of Atmospheres: an introduction to the chemistry of the atmospheres of Earth, the
planets, and their satellites. 2nd ed. New York, NY: Oxford University Press, Inc.
Wentworth, P., Jr.; Nieva, J.; Takeuchi, C.; Galve, R.; Wentworth, A. D.; Dilley, R. B.; DeLaria, G. A.; Saven, A.;
Babior, B. M.; Janda, K. D.; Eschenmoser, A.; Lerner, R. A. (2003) Evidence for ozone formation in human
atherosclerotic arteries. Science (Washington, DC, U.S.) 302: 1053-1056.
Weschler, C. J. (2004) Chemical reactions among indoor pollutants: what we've learned in the new millennium.
Indoor Air 14(suppl. 7): 184-194.
Weymer, A. R.; Gong, H., Jr.; Lyness, A.; Linn, W. S. (1994) Pre-exposure to ozone does not enhance or produce
exercise-induced asthma. Am. J. Respir. Crit. Care Med. 149: 1413-1419.
Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. (199la) Activity
patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
ARB/R93/487. Available from: NTIS, Springfield, VA.; PB94-108719.
Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b) Study of children's activity
patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
Wilson, A. M.; Wake, C. P.; Kelly, T.; Salloway, J. C. (2005) Air pollution, weather, and respiratory emergency
room visits in two northern New England cities: an ecological time-series study. Environ. Res. 97: 312-321.
8-91
-------
Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; McMurry, P. H. (2001) Measurement of Atlanta aerosol size distributions:
observations of ultrafine particle events. Aerosol Sci. Technol. 34: 75-87.
Woodwell, D. A.; Cherry, D. K. (2004) National Ambulatory Medical Care Survey: 2002 summary. Hyattsville,
MD: National Center for Health Statistics; DHHS publication no. (PHS) 2004-1250. (Advance data from vital
and health statistics; no. 346). Available: http://www.cdc.gov/nchs/data/ad/ad346.pdf [3 August, 2005].
Yang, Q.; Chen, Y.; Shi, Y.; Burnett, R. T.; McGrail, K. M.; Krewski, D. (2003) Association between ozone and
respiratory admissions among children and the elderly in Vancouver, Canada. Inhalation Toxicol.
15: 1297-1308.
Yang, I. A.; Holz, O.; Torres, R. A.; Magnussen, H.; Barton, S. I; Rodriguez, S.; Cakebread, J. A.; Holloway, J. W.;
Holgate, S. T. (2005) Association of tumor necrosis factor-a polymorphisms and ozone-induced change in
lung function. Am. J. Respir. Crit. Care Med. 171: 171-176.
8-92
-------
APPENDIX 8A
Summary of New Animal Toxicology,
Human Clinical, and U.S./Canadian Epidemiologic Studies
of Health Effects Associated with Ambient or
Near-Ambient Ozone Exposures
8A-1
-------
Table 8A-1. Short-Term Ozone-Induced Health Effects Observed in Controlled Human Exposure Studies
oo
>
to
Health Ozone
Effects Concentration
Exposure
Du ration/Activity
Subject
Characteristics
Observable Effects
Reference
Pulmonary Function
Short-term 0.0-0.4
(1-2 h)
Prolonged 0.0-0.08
(6.6 h)
0.0-0.12
(a) 0.08
(b) 0.08 (mean)
varied from
0.03 to 0.15
Repeated 0.12
2 h rest or IE
(4 x 15 minat
VE = 25or35
L/min/m2 BSA)
6.6 h
IE (6 x 50min)
VE = 20 L/min/m2 BSA
6.6 h
IE (6 x 50min)
VE = 20 L/min/m2 BSA
6.6 h
IE (6 x 50 min)
VE = 20 L/min/m2 BSA
6.6 h
50 min exercise/10 min
Healthy NS
18 to 36 yr old
mean age 24 yr
Healthy NS
Males23.5±3.0yrs
Females
22.8±1.2yrs
Healthy NS,
22.4 ± 2.4 yrs old
Healthy NS,
18 to 25 yr old
Healthy NS
Statistical analysis of 8 experimental chamber
studies conducted between 1980 and 1993 by the
U.S. EPA in Chapel Hill, NC. Response
decreased with age, was minimally affected by
body size corrections, and was not more sensitive
to O3 concentration than VE.
FEVj and symptom responses after 6.6 h
exposure to 0.04 and 0.06 ppm not significantly
different from FA.
FEVj and total symptoms after 6.6 h exposure to
0.04 ppm not significantly different from FA.
(a) FEVj decreased 6.2% after 6.6 h in square-
wave exposures. Total symptoms significantly
increased at 5.6 and 6.6 h.
(b) FEVj decreased 5.6 to 6.2% after 4.6 to 6.6 h,
respectively, in varied exposure; total symptoms
significantly increased also after 4.6 to 6.6 h.
FEVj responses were maximal on first day of
exposure (- 13%), less on second day (- 9%),
McDonnell
etal. (1997)
Adams (2006)
Adams (2002)
Adams (2003)
Folinsbee
etal. (1994)
rest, 30 min lunch
VE = 38.8 L/min
absent thereafter. Symptoms only the first 2 d.
Methacholine airway responsiveness was at least
doubled on all exposure d, but was highest on the
second day of O3. Airway responsiveness was
still higher than air control after 5 d of O3
exposure. Trend to lessened response, but it was
not achieved after 5 d.
-------
Table 8A-1 (cont'd). Short-Term Ozone-Induced Health Effects Observed in Controlled Human Exposure Studies
oo
>
Health Ozone
Effects Concentration
Exposure
Du ration/Activity
Subject
Characteristics
Observable Effects
Reference
Pulmonary Function (cont'd)
Preexisting 0. 125 to 0.250
Disease
0.125
0.0
0.12
Copollutants
SO2 0.10 to 0.12
NO2 0.0 to 0.3
3h
IE (10 min rest, 15 min
exercise on bicycle)
VE = 30 L/min
3 h IE x 4 d
0.75 h
IE (15 min exercise,
15 min rest)
VE = 40-46 L/min
1 h (mouthpiece)
IE
VE ~ 30 L/min
45-min exposure to air
or O3, followed by
15-min exposure to O3
orSO2
1 h (mouthpiece)
IE
VE = 33 L/min
VE = 35 L/min
Mild bronchial
asthma
20-53 yr old
Allergic rhinitis
19-48 yr old
Asthmatics sensitive
toSO2 19to38yrs
old
Allergic asthmatics,
12 to 1 Syr old,
medications
withheld for at
least 4 h before
exposures
Healthy NS,
12 to 17 yr old
Asthmatic
13 to ISyrold
Mean early-phase FEVl response and number of
> 20% reductions in FEVl were significantly
greater after 0.25 ppm O3 or 4 x 0.125 ppm O3.
Most of the > 15% late-phase FEVl responses
occurred after 4 d of exposure to 0. 125 ppm O3,
as well as significant inflammatory effects, as
indicated by increased sputum eosinophils
(asthma and allergic rhinitis) and increased
sputum lymphocytes, mast cell tryptase,
histamine, and LDH (asthma only).
No significant differences due to O3 between
placebo and antioxidant supplement cohort in
either spirometric responses or bronchial
hyperresponsiveness to 0. 1 ppm SO2.
Prior exposure to O3 potentiated pulmonary
function responses to SO2; decrements in FEVj.
No significant differences in FEVl and RT
between asthmatics and healthy, or between
atmospheres and cohorts.
Holz et al.
(2002)
Trenga et al.
(2001)
Koenig et al.
(1990)
Koenig et al.
(1988)
-------
Table 8A-1 (cont'd). Short-Term Ozone-Induced Health Effects Observed in Controlled Human Exposure Studies
Health
Effects
Ozone
Concentration
Exposure
Du ration/Activity
Subject
Characteristics
Observable Effects
Reference
Airway Hyperresponsiveness
Preexisting
Disease
0.125
0.250
oo
>
3h
IE (10 min rest, 15 min
exercise on bicycle)
VE = 30 L/min
3hIE x 4 d
Mild bronchial Mean early-phase FEVl response and number
asthma of >20% reductions in FEV{ were significantly
greater after 0.25 ppm O3 or 4 x 0.125 ppm O3.
Allergic rhinitis Most of the > 15% late-phase FEVl responses
occurred after exposure to 4 x 0.125 ppm O3,
as well as significant inflammatory effects, as
indicated by increased sputum eosinophils
(asthma and allergic rhinitis) and increased
sputum lymphocytes, mast cell tryptase,
histamine, and LDH (asthma only).
Holz et al.
(2002)
0.12 45 min
IE (exercise, rest,
exercise)
VE = 3 x resting
0.12 Ihrest
Air antigen
0.12 Rest
0.10 to 0.40 Ih
Light IE (2 x 15 min on
treadmill)
VE = 27 L/min
Healthy 0.08 to 0.12 6.6 h
Subjects IE at -39 L/min
Physician diagnosed
asthma; SO2-
induced airway
hyperreactivity
Mild allergic
asthma;
18 to 49 yr of age
Atopic
asthma
Stable mild
asthmatics with
FEVj >70% and
methacholine
responsiveness
Healthy NS,
18 to 32 yr old
The authors concluded O3 exposure increases
bronchial responsiveness to SO2 in asthmatics
and that antioxidant supplementation has a
protective effects against this responsiveness,
especially in the "more-severe" responders.
No effect of O3 on airway response to grass or
ragweed allergen.
No effect of O3 on airway response to grass
allergen.
No significant differences in FEVl or FVC were
observed for 0. 10 and 0.25 ppm O3-FA exposures
or postexposure exercise challenge; 12 subjects
exposed to 0.40 ppm O3 showed significant
reduction in FEVj.
33, 47, and 55% decreases in cumulative dose of
methacholine required to produce a 100%
increase in SRaw after exposure to O3 at 0.08,
0.10, and 0.12 ppm, respectively.
Trenga et al.
(2001)
Hanania et al.
(1998)
Ball et al.
(1996)
Weymer et al.
(1994)
Horstman
etal. (1990)
-------
Table 8A-1 (cont'd). Short-Term Ozone-Induced Health Effects Observed in Controlled Human Exposure Studies
Observable Effects Reference
Health
Effects
Ozone
Concentration
Exposure
Du ration/Activity
Subject
Characteristics
Pulmonary Inflammation
Short-term
(1-2 h)
0.1
0.12
oo
>
2h
Mild IE
Healthy subjects
mean age ~30 yr
2h
IE (15 mm/30 min);
(VE) ~ 20 L/mm/m2
BSA
Healthy
nonsmokers; mean
age -28 yr
Markers of exposure in exhaled breath Corradi et al.
condensate, including increased 8-isoprostane, (2002)
TEARS and LTB-4, and a marker of ROS-DNA
interaction in peripheral blood leukocytes
(8-OHdG), were increased in a sub-set of
subjects bearing the wild genotype for
NAD(P)H:quinone oxidoreductase and the null
genotype for glutathione-S- transferase Ml.
Increase in the percentage of vessels expressing Krishna et al.
P-selectin in bronchial biopsies at 1.5 h (1997)
postexposure. No changes in FEVb FVC,
inflammatory cells or markers in B AL, or vessels
expressing VCAM-1, E-selectin or ICAM-1 in
biopsies.
Preexisting
Disease
0.125
0.25
3 h exposures to both O3
concentrations and to
FA; 3 h on four
consecutive days to
0.125; study arms
separated by >4 wks
IE (15 mm/30 min)
Allergic asthmatic
and allergic rhinitic
subjects; 19-53 yrof
age
Repeated exposure caused increases in neutrophil
and eosinophil numbers in both subject groups,
as well as increased percentage and number of
lymphocytes in the asthmatics.
Holz et al.
(2002)
Cardiovascular
Copollutants
PM25
0.0 to 0.12
2-2.5 h
rest
Healthy NS Neither systolic nor diastolic pressure has been
18 to 50 yr old affected by pollutants exposure despite a significant
brachial artery constriction and a reduction in arterial
diameter when compared to filtered air (p = 0.03).
Brook et al.
(2002)
-------
Table 8A-2. Effects of Acute O3 Exposure on Lung Function in the U.S. and Canada
oo
Reference, Study
Location and Period
Braueretal. (1996)
Fraser Valley, British
Columbia, Canada
Jun-Aug 1993
Mortimer et al.
(2000; 2002)
Eight urban areas in
the U.S.: Baltimore,
MD; Bronx, NY;
Chicago, IL;
Cleveland, OH;
Detroit, MI;
East Harlem, NY;
St. Louis, MO;
Washington, DC
Jun-Aug 1993
Ross et al. (2002)
East Moline, IL and
nearby communities
May-Octl994
Naeheretal. (1999)
Vinton, VA
Summers 1995, 1996
Study Population
Berry pickers aged
10-69 yr(n= 58)
repeatedly monitored
over a 59-d period.
Outdoor work shifts
averaged 1 1 h in
duration.
National Cooperative
Inner City Asthma
Study (NCICAS)
cohort. Asthmatic
children aged 4-9 yr
(n = 846) repeatedly
monitored over a
3-mo period.
Mild and severe
asthmatics aged 5-49
yr (n = 40) repeatedly
monitored for a 5 -mo
period.
Nonsmoking women
aged 19-43 yr
(n = 473) who
recently delivered
babies repeatedly
monitored for a
two-wk period.
Statistics for 8-h max O3
Air Quality Data (ppb) "
Mean O3 Levels
(ppb) 98th % 99th % Range
l-hmaxO3: 55 55 3-55
40.3
SD 15.2
8-havgO3 64.3 66 28.8-66
(10 a.m. -6 p.m.):
48
SD not provided.
Range of
medians across
cities:
Approximately
34 to 58
(<5%ofdays
exceeded 80).
8-h max 03: 68.8 75 8.9-78.3
41.5
SD 14.2
IQR20
24-havgO3: 74 79 13-87
34.87
SD 8.86
Range 8.74-56.63
Standardized Percent Change in Lung Function (95% CI)b
Morning Afternoon Cross-day
FEV,: FEV,: FEV,:
Lag 1: Lag 0: Lag 0:
-6.4% (-8.0, -4.7) -5.4% (-6.5, -4.3) 0.0% (-1.7, 1.7)
PEF: — —
Lag 1-5:
All areas:
-1.18% (-2.10, -0.26)
PEF: PEF: —
Lag 0-1: LagO:
-0.96% (-1.78, -0.14) -1.08%(-1.78, -0.37)
PEF: PEF: —
Lag 1 : Lag 0:
-0.31% (-0.68, 0.07) -0.36% (-0.73, 0.01)
Lag 1-3: Lag 1-5:
-0.52% (-1.11, 0.07) -1. 11%(-1. 88, -0.33)
-------
Table 8A-2 (cont'd). Effects of Acute O3 Exposure on Lung Function in the U.S. and Canada
oo
Reference, Study
Location and Period
Korricketal. (1998)
Mount Washington,
NH
Summers 1991, 1992
Neasetal. (1995)
Uniontown, PA
Summer 1990
Study Population
Hikers aged 15-64 yr
(n = 530) monitored
before and after their
hike. Hikes averaged
8 h in duration.
Symptomatic and
asymptomatic 4th and
5th grade children
(n = 83) who did not
use any asthma
medication during the
previous year. Each
child, on average, was
monitored on 43 d.
Statistics for 8-h max O3
Air Quality Data (ppb) "
(ppb) 98th % 99th % Range
Avg of hourly O3 87 89 24-91
during each hike
(approximately
8-h avg O3):
40
SD12
Range 2 1-74
A1112-havgO3 96.5 98 15-98
(8 a.m.-8 p.m.
and 8 p.m.-8
a.m.):
37.2
IQR29.9
Maximum 87.5
Daytime
12-h avg O3
(8 a.m.-8p.m.):
50.0
SD not provided.
Standardized Percent Change in Lung Function (95% CI)b
Morning Afternoon Cross-day
— — FEV,:
LagO:
All hikers (n = 530):
-1.53% (-2.82, -0.24)
Respiratory disease
status:
Wheeze/asthma
(n = 40):
-4.47% (-7.65, -1.29)
No wheeze/asthma
(n = 490):
-1.08% (-2.49, 0.33)
Hours hiked:
Hiked 2-8 h (n = 265):
-0.99% (-2.70, 0.72)
Hiked 8-12 h (n = 265):
-2.07% (-3.78, -0.36)
— PEF: —
Daytime 12-h avg O3:
LagO:
-0.62%(-1.23, -0.01)
Weighted by
proportion of time
spent outdoors during
prior 12 h:
LagO:
-0.78%(-1.86, -0.31)
-------
Table 8A-2 (cont'd). Effects of Acute O3 Exposure on Lung Function in the U.S. and Canada
oo
>
oo
Reference, Study
Location and Period
Neasetal. (1999)
Philadelphia, PA
M-Sep 1993
Delfinoetal. (1997)
Alpine, CA
May-Jul 1994
Linnetal. (1996)
Three towns in
California:
Rubidoux, Upland,
Torrance
Fall-spring 1992-
1993 and 1993- 1994
Study Population
Children aged 6-1 1 yr
(n= 156) at two
summer camps
repeatedly monitored
over 40 d.
Symptomatic
asthmatics, children
aged 10 to 15 yr
(n= 13) and adults
aged 24 to 47 yr
(n = 9), repeatedly
monitored for a
8-week period.
School children
(n = 269) repeatedly
monitored for one
week in fall, winter,
and spring during their
4th and 5th grade
school years.
Mean O3 Levels
(Ppb)
12-h avg O3
(9 a.m.-9 p.m.):
SW camp:
57.5
IQR 19.8
NE camp:
55.9
IQR 21. 9
12-h avg O3
(8 a.m.-8p.m.):
Ambient:
64
SD17
Range 34-103
Personal:
18
SD14
Range 0-80
5 5% of personal
O3 samples
below limit of
detection.
24-h avg O3:
23
SD12
Range 3-53
Statistics for 8-h max O3
Air Quality Data (ppb) " Standardized Percent Change in Lung Function (95% CI) b
98th % 99th % Range Morning
96.9 104.5 17.7-104.5 PEF:
Lagl:
-0.74% (-1.54, 0.07)
Lag 1-5:
-0.76% (-2.65, 1.13)
110 121 38-121 PEF:
No effects observed
using ambient or
personal O3
concentrations.
Effect estimates not
presented.
150 164 2.5-192.5 FEV,:
Lagl:
-0.27% (-0.79, 0.24)
Afternoon Cross-day
PEF: —
LagO:
-0.46%(-1.18,0.27)
Lag 1-5:
-0.26% (-1.40, 0.88)
PEF: —
No effects observed
using ambient or
personal O3
concentrations.
Effect estimates not
presented.
FEV,: FEV,:
Lag 0: Lag 0:
-0.1 9% (-0.76, 0.35) -0.61% (-1.09, -0.14)
-------
Table 8A-2 (cont'd). Effects of Acute O3 Exposure on Lung Function in the U.S. and Canada
oo
Reference, Study
Location and Period
Thurstonetal. (1997)
Connecticut River
Valley, CT
June 1991, 1992,
1993
Study Population
Children aged 7-13 yr
(total n= 166) with
moderate-to- severe
asthma repeatedly
monitored over a 5-d
period.
Statistics for 8-h max O3
Air Quality Data (ppb) "
(ppb) 98th % 99th % Range
l-hmaxO3: — — —
1991: 114.0
1992: 52.2
1993: 84.6
1991-1993:
83.6
Standardized Percent Change in Lung Function (95% CI)b
Morning Afternoon Cross-day
— — PEF:
LagO:
-1.2% (0.02, -2.4)
1 Using O3 data obtained for the study period in the location of the study, 8-h max O3 concentrations were derived and statistics were calculated. The 98th and 99th percentile values
for the full study period distribution are presented here (unless noted otherwise), along with the range (minimum-maximum) of concentrations. Since the time periods of the studies
vary in length, from several weeks to over 10 yr, the 98th and 99th percentile values were selected for presentation here as a high study period concentration that roughly
approximates a 4th maximum concentration, depending on the study period length.
•"Percent change in lung function per standard unit ppb O3: 40 ppb for 1-h max O3; 30 ppb for 8-h max O3 or 8-h avg O3; 25 ppb for 12-h avg O3; and 20 ppb for 24-h avg O3.
-------
Table 8A-3. Effects of Acute O3 Exposure on Asthma Emergency Department Visits in the U.S. and Canada
oo
>
Reference, Study
Location and Period
Stiebetal. (1996)
Saint John, New
Brunswick, Canada
May-Sep 1984-1992
Cassinoetal. (1999)
New York City
1992-1995
Wilson et al. (2005)
Manchester, NH
1996-2000
Friedman et al. (2001)
Atlanta, GA
Jun-Sep 1996
Mean O3 Levels
(Ppb)
1-hmax O3:
Warm season: 41.6
Range 0-160
95th % 75
24-havgO3:
Allyr: 17.5
IQR14
8-hmaxO3:
Spring (Mar-May):
43.4
SD9.7
Summer (Jun-Aug): 42.8
SD 14.6
1-hmax O3:
Baseline (Jun 21 -Jul 18
andAug5-Sepl): 81.3
SD not provided.
Intervention period
(Jul 19-Aug 4): 58.6
SD not provided.
Statistics for 8-h max O3
Air Quality Data (ppb) a Standardized Percent Excess Risk (95% CI) b
98th % 99th % Range All Year Warm Season
83 91 5-140.5 — All ages:
Lag 2: 9.3% (0.0, 18.7)
83.3 88.8 3-114.6 All ages: —
All subjects:
Lag 2: 8.3% (-0.8, 18.5)
Note: Used Poisson GAM with
default convergence criteria
Apr-Oct: 85 Apr-Oct: 93 Apr-Oct: — Apr-Sep:
5-121
All ages:
LagO: -3% (-14, 9)
Age 0-14 yr:
LagO: 6% (-25, 51)
Age 15-64yr:
LagO: -6% (-21, 12)
Jul 19-Aug Jul 19-Aug Jul 19-Aug — Age 1-16 yr:
4: 85.8 4: 85.8 4: 20-85.8 LagO: 19% (-1,43)
Lag 0-2: 29% (2, 64)
-------
Table 8A-3 (cont'd). Effects of Acute O3 Exposure on Asthma Emergency Department Visits in the U.S. and Canada
oo
Reference, Study
Location and Period
Jaffe et al. (2003)
Columbus, OH
Jun-Aug 1991-1996
Jaffe et al. (2003)
Cleveland, OH
Jun-Aug 1991-1996
Jaffe et al. (2003)
Cincinnati, Cleveland,
and Columbus, OH
Jun-Aug 1991-1996
Jaffe et al. (2003)
Cincinnati, OH
Jun-Aug 1991-1996
Statistics for 8-h max O3
Air Quality Data (ppb) a
(ppb) 98th % 99th % Range
8-h max O3: 98 106 25-117
Summer: 57
SD16
8-h max O3: 104 107 27-111
Summer: 50
SD17
8-h max O3: 104 108 24-124
Summer:
All three cities:
Not reported.
8-h max O3: 106 116 24-124
Summer: 60
SD20
Standardized Percent Excess Risk (95% CI) b
All Year Warm Season
— Age 5-34 yr
(medicaid recipients) :
Lag 3: 15. 8% (-3.0,
36.8)
— Age 5-34 yr
(medicaid recipients) :
Lag 2: 3.0% (-8.7, 15.8)
— Age 5-34 yr
(medicaid recipients):
All three cities:
Lag 2 or 3: 9.3% (0.0,
19.1)
— Age 5-34 yr
(medicaid recipients) :
Lag 2: 15.8% (0.0, 36.8)
-------
Table 8A-3 (cont'd). Effects of Acute O3 Exposure on Asthma Emergency Department Visits in the U.S. and Canada
oo
to
Reference, Study
Location and Period
Wilson et al. (2005)
Portland, ME
1998-2000
Mean O3 Levels
(Ppb)
8-hmaxO3:
Spring (Mar-May):
43.7
Statistics for 8-h max O3
Air Quality Data (ppb) a
98th % 99th % Range
Apr-Oct: Apr-Oct: 121 Apr-Oct:
108 15-142
Standardized Percent Excess Risk (95%
CI)b
Warm Season
All Year
— Apr-Sep:
All ages:
LagO: 9% (3,
16)
Tolbert et al. (2000)
Atlanta, GA
Jun-Aug 1993-1995
Peel et al. (2005)
Atlanta, GA
1993-2000
SD 10.2
Summer (Jun-Aug): 46.1
SD 15.4
8-hmax O3:
Summer: 59.3
SD19.1
8-hmax O3:
Mar-Nov: 55.6
SD23.8
108.9
Mar-Nov:
115
112.6
Mar-Nov:
124
16.2-135.8
Mar-Nov:
3-152
All available data (Mar-Nov):
All ages:
Lag 0-2: 2.6% (-0.5, 5.9)
Age 0-14 yr:
LagO: 13% (-5, 35)
Age 15-64yr:
LagO: -3% (-14, 9)
Age 65+ yr:
LagO: -9% (-28, 16)
Age 0-16yr:
Lagl: 6.1% (1.2, 11.3)
All ages:
Lag 0-2: 3.1% (0.2, 6.2)
a Using O3 data obtained for the study period in the location of the study, 8-h max O3 concentrations were derived and statistics were calculated. The 98th and 99th
percentile values for the full study period distribution are presented here (unless noted otherwise), along with the range (minimum-maximum) of concentrations.
Since the time periods of the studies vary in length, from several weeks to over 10 yr, the 98th and 99th percentile values were selected for presentation here as a
high study period concentration that roughly approximates a 4th maximum concentration, depending on the study period length.
bPercent change in lung function per standard unit ppb O3: 40 ppb for 1-h max O3; 30 ppb for 8-h max O3; and 20 ppb for 24-h avg O3.
-------
Table 8A-4. Effects of Acute O3 Exposure on Total Respiratory and Asthma Hospital Admissions
in the U.S. and Canada
oo
Reference, Study
Location and Period
Yang et al. (2003)
Vancouver, British
Columbia, Canada
1986-1998
Burnett etal. (1997a)
16 Canadian cities
1981-1991
Burnett etal. (1997b)
Toronto, Ontario, Canada
Summers 1992-1994
Sheppard et al.
(1999; reanalysis
Sheppard, 2003)
Seattle, WA
1987-1994
Statistics for 8-h max O3
Air Quality Data (ppb) ' Standardized Percent Excess Risk (95% CI) b
Mean O3 Levels (ppb) 98th % 99th % Range All Year Warm Season
24-havgO3: 42.7 47.3 1.1-71.9 Total respiratory: —
Allyr: 13.41 Age <3 yr:
SD6.61 Lag 4: 50.4% (33.2, 71.4)
IQR 9.74
Age 65+ yr:
Lag 4: 28.5% (19.4, 40.5)
l-hmaxO3: Apr-Dec: Apr-Dec: Apr-Dec: — Total respiratory:
47.1 51.3 6.2-68.4
Allyr: 31 Apr-Dec:
95th % 60 All ages:
Lagl: 5.6% (3.4, 7.9)
Mean range
across cities: 26-38
95th % 45-84
Apr-Dec: 32.9
95th % 64
l-hmaxO3: 62 64 0-79 — Total respiratory:
Summer: 41.2 12-havgO3:
IQR 22
All ages:
12-havgO3: Lag 1-3: 14.4% (8.7, 20.5)
Summer: IQR 11. 5
8-h max O,: 65 73 2-100 Asthma: —
Allyr: 30.4 Age 0-64 yr:
IQR20 Lag2: 10.7%(1.5, 20.1)
-------
Table 8A-4 (cont'd). Effects of Acute O3 Exposure on Total Respiratory and Asthma Hospital Admissions
in the U.S. and Canada
oo
>
Reference, Study
Location and Period
Lin et al. (2004)
Vancouver, British
Columbia, Canada
1987-1998
Burnett etal. (1999)
Toronto, Ontario, Canada
1980-1994
Lin et al. (2003)
Toronto, Ontario, Canada
1981-1993
Burnett etal. (2001)
Toronto, Ontario, Canada
1980-1994
Mean O3 Levels (ppb)
l-hmaxO3:
Allyr: 28.02
SD11.54
Range 1.93-105. 50
24-havgO3:
Allyr: 19.5
IQR19
l-hmaxO3:
Allyr: 30.39
SD 17.87
Range 0-141
l-hmaxO3:
Summer: 45.2
IQR25
Statistics for 8-h max O3
Air Quality Data (ppb) ' Standardized Percent Excess Risk (95% CI) b
98th % 99th % Range All Year Warm Season
— — — Asthma: —
Age 6-1 2 yr:
Males:
Low SES:
Lagl: -35.5%(-52.4, -15.4)
High SES:
Lagl: -17.8%(-39.6, 11.2)
Females:
Low SES:
Lagl: 32.6% (-7.9, 94.9)
High SES:
Lagl: -22.5% (-48.9, 14.1)
68.4 74.8 0.14-110.8 Asthma: —
All ages:
Lag 1-3: 6.5% (3.7, 9.4)
68.4 74.8 0.14-110.8 Asthma: —
Age 6-1 2 yr:
Males:
LagO: -7.8% (-22.6, 8.2)
Females:
LagO: -26.0% (-39.2, 8.2)
77.7 83.7 9-110.8 Total respiratory: Total respiratory:
Age <2 yr: Age <2 yr:
LagO: 1.9% (-2.7, 6.8) Lag 0: 6.7% (0.3, 13.6)
Lag 0-4: 14.1% (4.9, 24.1) Lag 0-4: 30.2% (16. 9,
45.2)
-------
Table 8A-4 (cont'd). Effects of Acute O3 Exposure on Total Respiratory and Asthma Hospital Admissions
in the U.S. and Canada
Statistics for 8-h max O3
Air Quality Data (ppb)"
Standardized Percent Excess Risk (95% CI)b
Reference, Study
Location and Period
Mean O3 Levels (ppb)
98th %
99th %
Range
All Year
Warm Season
oo
>
Luginaah et al. (2005) 1-h max O3:
Windsor, Ontario, Canada
1995-2000 Allyr: 39.3
SD21.4
Range 1-129
78
85
0-106 Total respiratory:
All ages:
Males:
Lagl: 5.4% (-10.5, 24.2)
Females:
Lagl: -7.2% (-24.1,13.5)
Age 0-14 yr:
Males:
Lagl: -7.6% (-33.4, 28.0)
Females:
Lagl: 6.7% (-22.7, 47.0)
Age 15-64 yr:
Males:
Lagl: -5.6% (-43.5, 58.0)
Females:
Lagl: -24.3%(-48.2,10.5)
Age 65+ yr:
Males:
Lagl: 13.0%(-10.5,42.8)
Females:
Lagl: -7.5% (-29.4, 21.3)
-------
oo
Table 8A-4 (cont'd). Effects of Acute O3 Exposure on Total Respiratory and Asthma Hospital Admissions
in the U.S. and Canada
Statistics for 8-h max O3
Air Quality Data (ppb)"
Standardized Percent Excess Risk (95% CI)b
Keierence, siuay
Location and Period
Schwartz (1996)
Spokane, WA
Apr-Oct 1988-1990
Mean O3 Levels (ppb) 98th % 99th % Range
Warm season: 41
IQR12
All Year Warm Season
— Total respiratory:
Age 65+ yr:
Gwynn and Thurston
(2001)
New York City
1988-1990
Schwartz etal. (1996)
Cleveland, OH
Apr-Oct 1988-1990
Gwynn et al. (2000)
Buffalo, NY
1988-1990
24-havgO3:
Warm season: 29
IQR9
24-h avg O3:
Allyr: 22.1
IQR14.1
Maximum 80.7
l-hmaxO3:
Warm season: 56
IQR28
24-h avg O3:
Allyr: 26.2
IQR 14.8
Maximum 87.6
106
6-125
91
92.5
99
104
5-120.3
4.5-123
Total respiratory:
All ages:
White:
Lagl: 1.1% (-0.4, 2.6)
Non-white:
Lagl: 4.0% (2.5,5.6)
1-hmax O3:
Lag 2: 40.3% (0.3, 96.1)
24-h avg O3:
Lag 2: 21.4% (-5.8, 56.2)
Note: Used Poisson GAM
with default convergence
criteria
Total respiratory:
Age 65+ yr:
Lag 1-2: 6.9% (1.5, 12.2)
Total respiratory:
All ages:
Lagl: 3.9% (1.8, 6.1)
Note: Used Poisson GAM with
default convergence criteria
-------
oo
Table 8A-4 (cont'd). Effects of Acute O3 Exposure on Total Respiratory and Asthma Hospital Admissions
in the U.S. and Canada
Statistics for 8-h max O3
Air Quality Data (ppb)"
Standardized Percent Excess Risk (95% CI)b
Keterence, Muay
Location and Period
Linn et al. (2000)
Los Angeles, CA
1992-1995
Mean O3 Levels (ppb) 98th %
24-havgO3: Allyr:
98.8
Winter: 14
SD 7 Summer:
175
Spring: 32
SD10
Summer: 33
SD8
Fall: 15
SD9
99th %
All yr:
106.9
Summer:
180
Range All Year
All yr: Total respiratory:
4.6-143.2
Age 30+ yr:
Summer: LagO: 1.61% (0.03, 3.22)
13.5-188
Warm Season
Total respiratory:
Summer:
Age 30+ yr:
LagO: 1.21%(-1
4.02)
.53,
"Using O3 data obtained for the study period in the location of the study, 8-h max O3 concentrations were derived and statistics were calculated. The 98th and 99th percentile
values for the full study period distribution are presented here (unless noted otherwise), along with the range (minimum-maximum) of concentrations. Since the time periods
of the studies vary in length, from several weeks to over 10 yr, the 98th and 99th percentile values were selected for presentation here as a high study period concentration
that roughly approximates a 4th maximum concentration, depending on the study period length.
•"Percent change in lung function per standard unit ppb O3: 40 ppb for 1-h max O3; 30 ppb for 8-h max O3; 25 ppb for 12-h avg O3; and 20 ppb for 24-h avg O3.
-------
Table 8A-5. Effects of Acute O3 Exposure on All-Cause Mortality in the U.S. and Canada
oo
oo
Reference, Study
Location and Period
Vedal et al. (2003)
Vancouver, British
Columbia, Canada
1994-1996
Villeneuve et al. (2003)
Vancouver, British
Columbia, Canada
1986-1999
Fairley (1999; reanalysis
Fairley, 2003)
Santa Clara County, CA
1989-1996
Moolgavkar (2003)
Cook County, IL
1987-1995
Ito and Thurston( 1996)
Cook County, IL
1985-1990
Lippmann et al. (2000;
reanalysis Ito, 2003)
Detroit, MI
1992-1994
Mean O3 Levels
(Ppb)
1-hmax O3:
Allyr: 27.3
SD 10.2
Range 3. 1-75.1
24-havgO3:
Allyr: 13.4
Range 0.6-38.6
8-h max O3:
Allyr: 29
SD15
24-havgO3:
Allyr: Median 18
Range 0.2-67
Summer: Median 28
Range 7-67
1-hmax O3:
Allyr: 38.1
SD 19.9
24-havgO3:
Allyr: 25
IQR 18-30
Statistics for 8-h max O3
Air Quality Data (ppb) a Standardized Percent Excess Risk (95% CI) b
98th % 99th % Range All Year Warm Season
40.5 47.3 1.6-58.7 — All ages:
LagO: 16.5% (5.3,28.4)°
42.7 47.3 1.1-71.9 Age65+yr: —
LagO: 2.1% (-1.3, 5. 3)
Lag 0-2: 0.4% (-3.2, 4.4)
67 74 2-105 All ages: —
LagO: 3.0% (-0.3, 6.4)
— — — All ages: All ages:
LagO: 0.28% (0.19, 0.36) LagO: 0.57% (0.42, 0.73)
Note: Used Poisson GAM Note: Used Poisson GAM
with default convergence with default convergence
criteria. criteria.
76 85.6 2.7-124 All ages: —
Lag 0-1: 3.9% (2.4, 5.8)
80 85 4.3-101.3 All ages: —
LagO: 1.84% (-1.73, 5. 53)
Lag 0-3: 0.81% (-3.78, 5.63)
-------
Table 8A-5 (cont'd). Effects of Acute O3 Exposure on All-Cause Mortality in the U.S. and Canada
oo
VO
Reference, Study
Location and Period
Chock et al. (2000)
Pittsburgh, PA
1989-1991
Gamble (1998)
Dallas, TX
1990-1994
Lippmann et al. (2000;
reanalysis Ito, 2003)
Detroit, MI
1985-1990
Dockeryetal. (1992)
St. Louis, MO
1985-1986
Dockeryetal. (1992)
Eastern Tennessee
1985-1986
Mean O3 Levels
(Ppb)
1-hmax O3:
Not reported.
24-havgO3:
Allyr: 22
Range 0-160
Summer: 30
Range 0-160
24-havgO3:
Allyr: 20.9
IQR 12.0-27.5
24-havgO3:
Allyr: 22.5
SD 18.5
24-havgO3:
Allyr: 23.0
SD11.4
Statistics for 8-h max O3
Air Quality Data (ppb) a Standardized Percent
98th % 99th % Range All Year
80 88.9 2.3-92.5 Age 0-74 yr:
LagO: -1.5% (-5.6, 2.8)
Age75+yr:
LagO: -1.8% (-6.0, 2.6)
81 86.3 2-98.7 All ages:
Lag 1-2: 3.7% (0.8, 6.6)
81.5 88.7 2-123.5 All ages:
LagO: 0.60% (-0.62, 1.83)
Lag 0-3: 0.95% (-0.43, 2.35)
— — — All ages:
Lagl: 0.6% (-2.4, 3.6)
— — — All ages:
Lagl: -1.3% (-7.9, 5.8)
Excess Risk (95% CI)b
Warm Season
Age 0-74 yr:
LagO: -1.7% (-6.2, 3.1)
Age 75+ yr:
LagO: -0.6% (-5.3, 4.4)
All ages:
Lag 1-2: 4.8%, p< 0.05
-------
Table 8A-5 (cont'd). Effects of Acute O3 Exposure on All-Cause Mortality in the U.S. and Canada
oo
to
o
Reference, Study
Location and Period
Lipfert et al. (2000)
7 counties in
Philadelphia, PA area
1991-1995
Schwartz (2005)
14 U.S. cities
1986-1993
Samet et al. (2000;
reanalysis Dominici
etal.,2003)
90 U.S. cities (80 U.S.
cities with O3 data)
1987-1994
Statistics for 8-h max O3
Air Quality Data (ppb) a
(ppb) 98th % 99th % Range
l-hmaxO3: May-Sep: May-Sep: May-Sep:
88.8 93.6 2.3-116.6
Allyr: 44.76
SD 25.68
l-hmaxO3: — — —
All yr: Median range 35.1
(Chicago, IL) to 60.0
(Provo, UT)
24-havgO3: — — —
All available data:
Mean range:
Approximately 12
(Des Moines, IA) to 36
(San Bernardino, CA)
Standardized Percent Excess Risk (95% CI) b
All Year Warm Season
All ages: —
Philadelphia:
Lag 0-1: 2.49%, p< 0.055
4 counties in PA:
Lag 0-1: 2.52%, p< 0.055
7 counties in PA and NJ:
Lag 0-1: 2.84%, p< 0.055
All ages: All ages:
Lag 0: 0.76% (0.13, 1.40) Lag 0: 1.04% (0.28, 1.77)
All available data: All ages:
Lagl: 1.02% (0.46, 1.57)
All ages:
LagO: 0.84% (0.45, 1.24)
Lagl: 0.38% (0.05, 0.71)
-------
Table 8A-5 (cont'd). Effects of Acute O3 Exposure on All-Cause Mortality in the U.S. and Canada
Reference, Study
Location and Period
Mean O3 Levels
Statistics for 8-h max O3
Air Quality Data (ppb)a
Standardized Percent Excess Risk (95% CI)l
98th %
99th %
Range
All Year
Warm Season
oo
Bell et al. (2004)
95 U.S. communities
1987-2000
Klemm et al. (2004)
Atlanta, GA
1998-2000
24-havgO3:
All available data from
95 communities: 26
55 communities with all
year data:
Median range:
14.38 (Newark, NJ)
to 37.30 (Bakersfield, CA)
40 communities with
warm season only data:
Median range:
20.41 (Portland, OR)
to 36.15 (Memphis, TN)
8-h max O3:
Allyr: 47.03
SD 24.71
Moolgavkar (2003) 24-h avg O3:
Los Angeles County, CA
1987-1995 Allyr: Median 24
Range 0.6-77
Ostro et al. (2000)
Coachella Valley, CA
1989-1998
Summer: Median 36
Range 5-77
1-hmax O3:
All yr:
Indio: 62
Range 0-180
Palm Springs: 67
Range 0-190
— All available data:
All ages:
Lag 0: 0.50% (0.24, 0.78)
Lag 0-6: 1.04% (0.54, 1.55)
Age < 65 yr:
Lag 0-6: 1.00% (0.20, 1.85)
Age 65-74 yr:
Lag 0-6: 1.40% (0.56, 2.25)
Age75+yr:
Lag 0-6: 1.04% (0.36, 1.75)
6.63-124.41 Age65+yr:
Lag 0-1: 4.2 (-2.4, 11.2)
All ages:
LagO: 0.08% (0.01, 0.15)
Note: Used Poisson GAM
with default convergence
criteria.
All ages:
-l%(-4, 3)
Note: Used Poisson GAM
with default convergence
criteria.
All ages:
LagO: 0.44% (0.16, 0.76)
Lag 0-6: 0.78% (0.26, 1.30)
All ages:
Lag 0: 0.20% (0.06, 0.34)
Note: Used Poisson GAM
with default convergence
criteria.
-------
Table 8A-5 (cont'd). Effects of Acute O3 Exposure on All-Cause Mortality in the U.S. and Canada
Statistics for 8-h max O3
Air Quality Data (ppb)a
Standardized Percent Excess Risk (95% CI)l
Reference, Study
Location and Period
Mean O3 Levels
98th %
99th %
Range
All Year
Warm Season
Moolgavkar et al. (1995)
Philadelphia, PA
1973-1988
Kinneyetal. (1995)
Los Angeles County, CA
1985-1990
24-havgO3:
Summer: 35.5
Range 1.3-159.0
1-hmax O3:
Allyr: 70
SD41
All ages:
Lagl: 3.1% (1.8, 4.4)
115.3 130 5.4-156.1 All ages:
Lagl: 0.6% (0.0, 1.4)
oo
to
to
Kinney and Ozkaynak
(1991)
Los Angeles County, CA
1970-1979
Ostro (1995)
San Bernardino County
and Riverside County,
CA
1980-1986
1-hmax O3:
All yr:
Total oxidants (O^: 75
SD45
1-hmax O3:
Warm season: 140
Range 20-370
— All ages:
Lagl: 0.79% (0.33, 1.26)
All ages:
LagO: 0.8% (0.0, 2.0)
a Using O3 data obtained for the study period in the location of the study, 8-h max O3 concentrations were derived and statistics were calculated. The 98th and
99th percentile values for the full study period distribution are presented here (unless noted otherwise), along with the range (minimum-maximum) of concentrations.
Since the time periods of the studies vary in length, from several weeks to over 10 yr, the 98th and 99th percentile values were selected for presentation here as a
high study period concentration that roughly approximates a 4th maximum concentration, depending on the study period length.
bPercent change in lung function per standard unit ppb O3: 40 ppb for 1-h max O3; 30 ppb for 8-h max O3; and 20 ppb for 24-h avg O3.
0 Due to the low mean and relatively small variability in 1 -h max O3 concentrations in the study by Vedal et al. (2003), the standardized incremental change of 40 ppb
may not be appropriate. The effect estimate in the warm season was 11.0% (95% CI: 3.6, 18.6) per mean level (27.3 ppb) increase in 1-hmax O3 at a 0-d lag.
-------
Table 8A-6. Toxicological Effects of Acute Ozone Exposure in Animals
oo
to
OJ
Effect
Inflammation and increased
epithelial/endothelial permeability
Cardiovascular
( 1 heart rate)
( T atrial natiuretic peptide)
Pulmonary Function
( T AHR)
(lVt,Tf)
Histological
Host Defense
Species
Rat
Mouse
Dog
Rat
Rat
Rat
Guinea Pig
Ratb
Rat
Mouse
Guinea Pig
Rat
O3 Concentration and Duration of Exposure a References
0.5 ppm x 3 h
0.26 ppm, 8 h/d, 5 d/wk x 1-90 d
0.4 ppm x 6 h
0.1 ppm x 5 h
0.5 ppm x 8 h
0.05 ppm x 4 h
0.3 ppm x 4 h
0.1 ppm x 5 h
0.2 ppm, 3-7 d
0.2 ppm, 3-7 d
0.2 ppm, 3-7 d
0.8 ppm x 3 h
Bhalla and Hoffman (1997)
Kleebergeretal. (2001)
Foster and Freed (1999)
Aritoetal. (1997)
Veselyetal. (1994a,b,c)
Depuydtetal. (1999)
Seguraetal. (1997)
Aritoetal. (1997)
Dormansetal. (1999)
Dormansetal. (1999)
Dormansetal. (1999)
Dong etal. (1998)
a Lowest reported ozone concentration from the current literature.
b Effects on ventilation only occurred in young rats exposed to 0.1 ppm.
0.5 ppm exposures.
No significant changes were seen between young and old animals at 0.3 and
-------
9. ENVIRONMENTAL EFFECTS: OZONE EFFECTS
ON VEGETATION AND ECOSYSTEMS
9.1 INTRODUCTION
A number of ozone (O3) effect studies were published between 1996 and 2005, and they
are reviewed in this document in the context of the previous O3 air quality criteria documents
(AQCDs) (U.S. Environmental Protection Agency, 1978, 1986, 1992, 1996b). The studies
include multiple plant and tree species, multiple venues, and multiple research approaches from
empirical data to process models. The research since 1996 continues to support and strengthen
the conclusions from the previous O3 AQCD, including that:
(1) the entrance of O3 into the leaf through the stomata is the critical step in O3 effects;
(2) current ambient concentrations in many areas of the country are sufficient to impair
growth of numerous common and economically valuable plant and tree species;
(3) effects can occur with only a few hourly concentrations above 80 ppb;
(4) a plant's response to O3 depends upon the cumulative nature of ambient exposure as
well as the temporal dynamics of concentrations;
(5) other environmental biotic and abiotic factors (e.g., insects, water, and nutrient
availability, elevated CO2, and temperature) are also influential to the overall impact
of O3 on plants and trees.
Research, to date, has continued to be focused at the species level with very few studies at the
ecosystem level. Consequently, a high degree of uncertainty remains in our ability to assess the
impact of O3 on ecosystem services.
Although chamber exposures still dominate the effects literature, there has been a general
shift away from chamber studies in favor of more field-based approaches. Field-based
approaches include surveys of visible injury as well as physiological and growth studies using
non-chambered free-air CO2 exposure (FACE) systems. The FACE system studies published
thus far have supported earlier observations of foliar injury and altered growth effects in open-
top chambers (OTC) systems. Increased emphasis has also been placed on quantifying aspects
9-1
-------
of O3 uptake to better link ambient exposure monitoring with plant or tree response. Much of
this progress has occurred in Europe in developing their O3 air quality management tool, the
"critical level." The European research has developed exposure-response functions for several
European crops and tree seedlings using OTC studies and for use in developing and testing
models that simulate uptake. Evaluation of the above newly available information has added to
our knowledge and provides new research directions, but it has not fundamentally altered the
conclusions of the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996).
It is well known that O3 is phytotoxic and that toxicity occurs only if O3 or its reaction
products reach the target tissues in the plant cell. Recent studies have provided an increased
understanding of how O3 interacts with the plant at the cellular level. This increased
understanding of cellular-level O3 effects has been translated into better models, more detailed
schemata of how O3 alters much of the basic metabolism of plants, and how to construct an
index that more aptly captures the species, climate, and site factors that alter uptake. These
results have and will continue to lead to better quantification of exposure and effect. However,
the translation of these mechanisms into how O3 is involved in altered cell metabolism and
subsequent reductions in whole-plant productivity and ecosystem-level responses remains to be
more fully resolved.
The ensuing sections of this chapter (Sections 9.2 to 9.8) are not intended to provide a
comprehensive review of the environmental effects of O3, but rather an assessment of key
information published since the 1996 O3 AQCD. More detailed discussion of the research since
1996 is provided in Chapter 9 Annex Sections AX9.1 to AX9.7 (in Volume 3 of this document).
The framework for Chapter 9 follows the environmental effects chapter of the 1996 O3 AQCD.
First, an overview of various methodologies that have been, and continue to be, central to
quantifying O3 effects on vegetation is provided in Section 9.2 below (see Section AX9.1 for
more detailed discussion). The adequacy of each methodology is discussed in the context of
developing statistically robust data appropriate for assessing and predicting the risk of O3 injury
to vegetation resources. In Section 9.3, research is reviewed from the molecular to the
biochemical and physiological levels in impacted plants, offering insight into the mode of action
of O3 (see also AX9.2). Then, the manner in which plants respond to O3, as influenced by
9-2
-------
numerous environmental biotic and abiotic factors, is next discussed (see also AX9.3).
Quantifying these various modifiers is critical to scaling the response of individual plants to the
community level and across varied landscapes and climates and is needed for regional-to-
national assessments of risk. The development of indices of O3 exposure or O3 uptake is
discussed in the context of their adequacy to realistically describe the ambient concentration-
response relationships (see also AX9.4). Exposure-response relationships for a large number of
crop species and cultivars, native vegetation, and tree species are also reviewed, tabulated, and
compared to form the basis for an assessment of the potential risk from current levels of O3 on
vegetation resources (see also AX9.5). Available research by which to assess the impact of O3
on ecosystems is also reviewed, along with data potentially available for estimating the loss of
various ecosystem services (see also AX9.6). Finally, available research on the economic
impact of O3 effects on vegetation resources is briefly discussed (see also AX9.7).
9.2 METHODOLOGIES USED IN VEGETATION RESEARCH
Methodological advancements since 1996 have not fundamentally altered our
understanding of O3 effects on plants or ecosystems. Most of the new information confirms
earlier conclusions and provides additional support for OTC use in assessing plant species and
developing exposure-response relationships. A more in-depth discussion of this topic can be
found in Annex Section AX9.1.
The majority of O3 effects studies are fumigation studies that were conducted in controlled
chambers, as noted in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996). That
document noted that OTCs represented the best technology for determining statistically robust
exposure-response models of O3 and crop yield and plant biomass at that time. While OTCs are
still the best method for conducting multiple, replicated controlled exposures of varying length
and frequency for developing exposure-response relationships, several new approaches have
been applied to O3 effects research, most notably free-air exposure systems. FACE systems
eliminate some concerns about closed or open-top chamber experiments including small plot
size, altered microclimate within OTCs, and the effect of charcoal filtering on overall air quality
-------
within OTCs. The FACE studies have, on the whole, confirmed what was already understood or
hypothesized about how plants and plant assemblages respond to O3. Some shortcomings of
using free-air systems in O3 research have also been identified, namely the relatively poor
control of exposure levels, the presence of "hotspots" and the inability to decrease O3
concentrations to below ambient levels when ambient concentrations are phytotoxic.
Nonetheless, the application of FACE systems and other open-air systems to O3-exposure
research have greatly helped our scaling efforts and are, perhaps, the best current approach for
studying the response of plant species mixtures to O3 (Nussbaum and Fuhrer, 2000).
One of the advantages of the application of free-air systems to O3 research is the ability to
compare response of plants in open-field systems with results from OTCs. In particular, studies
with quaking aspen (Populus tremuloides L.) performed in OTCand FACE systems and at sites
along an ambient O3 gradient have showed that O3 foliar symptom expression is generally
similar across these methodologies, supporting the previously observed level of variation among
aspen clones in OTC studies (Isebrands et al., 2000, 2001; Karnosky et al., 1999). Recent
exposure studies using the non-chambered FACE system with soybean cultivars in Illinois
reported reductions in soybean yield of 15 to 25% (compared to ambient) in two year-long
studies (Morgan et al., 2006). The results are similar to the reductions reported in multiple
soybean studies conducted in the 1980s using OTC systems. Multiple-year exposure studies
employing the FACE system reported foliar injury with reduced volume growth in aspen and
maple similar to results reported from earlier OTC studies (Karnosky et al., 1999; Isebrands
et al., 2001). Similar observations between the two exposure systems offer some corroborative
support for the use of either system. Each has advantages and disadvantages as experimental
tools and each can be used to effectively investigate O3 effects. Extrapolation of the results from
chamber studies depends on fully characterizing temperature, light, turbulence, and other
chamber characteristics during exposures (Nussbaum and Fuhrer, 2000), but study design is
equally important. Conducting studies with a large number of plant species across regions of the
country where those species are indigenous is important in delineating regional climatic
differences in order to reduce the uncertainty associated with extrapolating composited response
9-4
-------
functions across regions and to identify relative risk to vegetation in relation to given O3
exposure values (U.S. Environmental Protection Agency, 1996).
The lack of rural monitors continues to be a major problem in characterizing O3 exposures
in remote areas and complex landscapes so as to link effects to exposure in natural ecosystems.
Since the 1996 O3 AQCD, the use of passive samplers has expanded monitoring efforts to
include remote areas that were previously uncharacterized. This has greatly enhanced our ability
to link O3 symptomology with elevated O3 exposure in such remote areas. However, passive
samplers do not capture the temporal dynamics of exposure. Therefore, passive samplers cannot
substitute for active monitors when attempting to link exposure dynamics to plant response or
when developing exposure- or dose-response relationships of much value as inputs for the
standard setting process. To overcome this problem, Krupa et al. (2001, 2003) used models and
data from a collocated O3 monitor to estimate the underlying frequency distribution of hourly O3
concentrations from passive samplers. Future development of passive monitor technology and
data synthesis techniques holds promise, particularly as it is unlikely that extensive O3
monitoring networks will be established in rural areas in the near future.
Exclusion methods that employ protective chemicals such as ethylenediurea (EDU) are the
least disruptive of ambient culture conditions in the field, as noted in the 1996 O3 AQCD.
However, the level of protection afforded by EDU is site- and species-specific and is subject to
local meteorologic conditions. In addition, new evidence suggests that EDU does not always
provide greater protection at higher O3 exposures and that the degree of protection by EDU
largely depends on environmental conditions. Because of the variability observed in the level of
protection provided and the fact that mechanisms of protection afforded by EDU and other
exclusion methods are unknown, caution is needed in applying this approach to the study of O3
effects in the field.
Advancements in biomonitoring have been made since the 1996 O3 AQCD, primarily in
the area of identification and symptom verification of O3-sensitive species (Flagler, 1998; Krupa
et al., 1998; Innes et al., 2001; Smith et al., 2003). The U.S. Department of Agriculture (USDA)
Forest Service continues its program to monitor O3 effects in forested ecosystems throughout the
United States. Currently, 33 states participate in the program, which uses a grid system to
9-5
-------
identify the location of plants showing foliar injury. Although results cannot be used for
developing exposure-response relationships or for quantifying responses to O3, they can provide
an annual assessment and correlative information regarding the extent of O3 injury occurring
across many regions of the United States.
9.3 SPECIES RESPONSE AND MODE-OF-ACTION
Several steps in the process of O3 uptake and toxicity are now better understood than in
1996. These advancements are important in refining hypotheses on O3 uptake and mode of
action on plants and in developing a flux-based index for use in quantifying response and,
ultimately, for potential use in developing a secondary national ambient air quality standard
(SNAAQS) for O3. The new information available on the mode of action of O3 is, in part, a
result of improved molecular tools for following rapid changes that occur within the leaf (Ward
et al., 1991; Pell et al., 1997; Sandermann, 2000). This new information is discussed in greater
detail in Annex Section AX9.2.
Clearly, many changes occur within hours or possibly days following O3 exposure
(Sandermann, 1998). However, other O3 effects take longer to occur (e.g., "carry-over" effects
seen in the growing season following O3 exposure) and tend to be most obvious only under
exposure to low O3 concentrations for long periods (Hogsett et al., 1989; Andersen et al., 1997;
Langebartels et al., 1997). These low-exposure chronic effects have been linked to the
senescence process or to some physiological response very closely linked to senescence (e.g.,
translocation, reabsorption, storage, and allocation of nutrients and carbon).
Langebartels et al. (1997) discussed "memory" or carry-over effects within the plant to
explain sensitivity to frost in the winter following summertime O3 exposure. Others have argued
that this sensitivity is due to the nutrient status of the tree during the over-wintering phase of its
life and to chronic (ongoing, less severe levels with fewer peaks at very high levels) exposure to
ambient O3 inducing (1) mineral nutrient deficiency; (2) alterations of normal metabolism,
including translocation and allocation of carbohydrates and probably nitrogen; and/or
(3) disturbance of normal transpiration and diurnal cycling, leading to water stress (Schmieden
9-6
-------
and Wild, 1995). While generalized low nutrient concentrations within the foliage may not
occur, localized deficiencies might. This is difficult to observe or to prove without a great deal
of work on all portions of a tree and without a general hypothesis of what is occurring.
It is important to note that the dramatic strides made over the last few years in
understanding the genetic makeup of plants, gene control, and signal transduction and control
will accelerate in the future and translate into better models of the hypotheses listed above as
well as into more detailed schemes of how O3 alters basic plant metabolism. Thus, while our
understanding of how O3 interacts with the plant at the cellular level has dramatically improved
(Assmann, 2003; Assmann and Wang, 2001; Rao and Davis, 2001), the translation of those
mechanisms into how O3 is involved with altered cell metabolism and the subsequent reductions
in whole plant productivity and other physiological facts have not yet been fully achieved.
As the understanding of wounding responses in plants and more information on genome details
and varied plant mutants becomes available, the cellular and physiological responses of plants
to O3 exposures are slowly becoming clearer. However, more studies on a larger variety of
species are needed before this type of information can be incorporated into indices of response
and for consideration in developing SNAAQS.
9.4 MODIFICATION OF FUNCTIONAL AND GROWTH RESPONSES
It has been known for decades that several factors, both biotic and abiotic, alter plant
response to O3. However, only a few studies reported since the 1996 O3 AQCD have improved
our understanding of the role of these interactions in modifying plant O3 response. Quantifying
how these interactions alter plant O3 response is a critical first step to reducing the uncertainty in
extrapolating individual plant responses spatially or to higher levels of biological organization,
e.g., ecosystems. Although the recent studies have not improved our ability to quantify the
degree to which these factors modify plant O3 response, they have reinforced the conclusions of
the 1996 O3 AQCD with regard to factors known to alter plant response to O3. This new
information is discussed in greater detail in Annex Section AX9.3.
9-7
-------
In the area of biotic interactions, new evidence with regard to insect pests and diseases
(see Docherty et al., 1997, and Fliickiger et al., 2002 for recent reviews) has not altered the
conclusions and uncertainties noted in the 1996 O3 AQCD. Recent studies have supported the
earlier conclusion that O3 often increases the likelihood and success of insect attacks, but only
with respect to chewing insects (e.g., Percy et al., 2002; Kopper and Lindroth, 2003). With the
economically important group of sucking insects (e.g., aphids), no clear trends have been
revealed in the latest studies (see reviews by Docherty et al., 1997; Fliickiger et al., 2002).
It seems that some insect problems could increase as a result of greater O3 levels, but predicting
the likelihood and severity of any particular O3-plant-insect interaction is not possible at
this time.
The situation is somewhat clearer with respect to interactions involving facultative
necrotrophic plant pathogens, with O3 exposure generally contributing to increased disease
(Fliickiger et al., 2002). With obligate biotrophic fungal, bacterial, and nematode diseases,
however, twice as many reports indicate O3-induced inhibitions as enhancements. This pattern is
supported by the concept put forth by Dowding (1988) that pathogens that benefit from damage
to cells are enhanced by pollution stress of their hosts, whereas those pathogens and pests that
require healthy hosts are depressed by pollution stress. Despite frequent reports that infection by
obligate biotrophs reduces the severity of O3-induced foliar injury (e.g., Schraudner et al., 1996),
such infection does not result in true "protection", as the disease per se causes negative effects
on the host plant. With obligate biotrophs, the nature of any interaction with O3 is probably
dictated by the unique, highly specific biochemical relationships between the pathogen and the
host plant. At this time, therefore, although some diseases may become more widespread or
severe as a result of exposure to O3, it is still not possible to predict exactly which diseases are
likely to present the greatest risks to crops and forests.
Recent studies of interactions between O3 and root symbionts have supported conclusions
put forth in the 1996 O3 AQCD. Several studies have indicated that the functioning of tree root
symbioses with mycorrhizae may be adversely affected by O3 (e.g., Kytoviita et al., 2001),
but there is also evidence that the presence of mycorrhizae may overcome O3-enhanced root
diseases (Bonello et al., 1993). Also, there is evidence that O3 may encourage the spread of
9-8
-------
mycorrhizae to the roots of uninfected trees. The role of O3 in altering root symbionts, its
interactions with soil organisms, and the subsequent feedback effects on plant growth represent
one of the greatest areas of uncertainty in assessing the influence of O3 on ecosystems and
ecosystem services (Andersen, 2003).
The few recent studies of the impact of O3 on intraspecific plant competition confirmed
that grasses frequently show greater resilience than other types of plants. In grass-legume
pastures, the leguminous species tend to suffer greater growth inhibition (Johnson et al., 1996;
Nussbaum et al., 2000). The suppression of Ponderosa pine (Pinusponderosa Laws.) seedling
growth by blue wild-rye grass (Elymus glaucus Buckl.) was markedly increased by O3 (Andersen
et al., 2001). However, predicting the impact of O3 on specific competitive situations, such as
successional plant communities or crop-weed interactions, is not possible at this time.
Physical or abiotic factors play a large role in modifying plant response to O3, and new
information is available that supports the conclusions of the 1996 O3 AQCD. Although some
recent field studies have indicated that O3 impact significantly increases with increased ambient
temperature (Ball et al., 2000; Mills et al., 2000), other studies have indicated that temperature
has little effect (Balls et al., 1996; Fredericksen et al., 1996). Temperature affects the rates of all
physiological processes based on enzyme catalysis and diffusion; each process and overall
growth (the integral of all processes) has a distinct optimal temperature range. It is important to
note that a plant's response to changes in temperature will depend on whether it is growing near
its optimum temperature for growth or near its maximum temperature (Rowland-Bamford,
2000). But temperature is very likely an important variable affecting plant O3 response in the
presence of the elevated CO2 levels contributing to global climate change. In contrast, evidence
continues to accumulate that O3 exposure sensitizes plants to low temperature stress (Colls and
Unsworth, 1992) and, also, that O3 decreases below-ground carbohydrate reserves, which may
lead to responses in perennial species ranging from rapid demise to impaired growth in
subsequent seasons (i.e., carry-over effects) (Andersen et al., 1997).
Light, a component of the plant's physical environment, is an essential "resource" whose
energy content drives photosynthesis and CO2 assimilation. It has been suggested that increased
light intensity may increase the O3 sensitivity of light-tolerant species while decreasing that of
9-9
-------
shade-tolerant species, but this appears to be an oversimplification with many exceptions.
As pointed out by Chappelka and Samuelson (1998) and Topa et al. (2001), the interaction
between O3 sensitivity and light environment is complicated by the developmental stage as well
as the light environment of individual leaves in the canopy.
Although the relative humidity of the ambient air has generally been found to increase the
adverse effects of O3 by increasing stomatal conductance (thereby increasing O3 flux) abundant
evidence also indicates that the ready availability of soil moisture results in greater O3 sensitivity
(Mills, 2002). The partial "protection" against the adverse effects of O3 afforded by drought has
been observed in field experiments and modeled in computer simulations (Broadmeadow and
Jackson, 2000). There is also compelling evidence that O3 can predispose plants to drought
stress (Maier-Maercker, 1998). Hence, the response will depend to some extent upon the
sequence in which the stressors occur; but, even though the nature of the response is largely
species-specific, successful applications of model simulations have led to larger-scale
predictions of the consequences of O3-drought interactions. However, regardless of the
interaction, the net result on short-term growth is negative; although in tree species, other
responses (such as increased water use efficiency) could benefit long-term survival.
Mineral nutrients in the soil, other gaseous air pollutants, and agricultural chemicals
constitute environmental chemical factors. The evidence regarding interactions with specific
nutrients is still contradictory: some experimental evidence indicates that low general fertility
increases sensitivity to O3 (Whitfield et al., 1998; Landolt et al., 1997), while others have found
less sensitivity with decreased fertility (Cardoso-Vilhena and Barnes, 2001). Simulation
modeling of trees suggests that nutrient deficiency and O3 may act less than additively.
Interactions of O3 with other air pollutants have received relatively little attention since
1996 (see Barnes and Wellburn, (1998, and Fangmeier et al., 2002, for recent reviews). The
situation with SO2 remains inconsistent, but SO2 seems unlikely to pose any additional risk to
those related to other individual pollutants. With NO and NO2, the situation is complicated by
their nutritional value as a N source. More information is needed to predict the outcomes of
different O3-NO-NO2 scenarios. The latest research into O3-acid rain interactions has confirmed
that, at realistic acidities, significant interactions are unlikely (Laurence et al., 1997; Momen
9-10
-------
et al., 1997, 1999; Sayre and Fahey, 1999). A continuing lack of information precludes any
generalizations about interactive effects of O3 with NH3, HF, or heavy metals. More evidence
was reported for protective effects against O3 afforded by the application of fungicides (Wu and
Tiedemann, 2002).
Considerable emphasis during the last decade has been placed on research evaluating
potential O3 interactions with the components of global climate change: increased atmospheric
CO2, increased mean global temperatures, and increased surface level UV-B radiation.
However, it must be noted that most of these studies have tended to regard increased CO2 levels
and increased mean temperatures as unrelated phenomena. Experiments into the effects of
doubled CO2 levels at today's mean ambient temperatures are of questionable value in trying to
assess the impact of climate change on responses to O3. To date, the limited experimental
evidence and model simulations suggest that even though an enriched CO2 atmosphere
(-600 ppm) would more than offset the impact of O3 on responses as varied as wheat (Triticum
aestivum L.) yield or young Ponderosa pine growth, the concurrent increase in temperature
would reduce, but probably not eliminate, the net gain (Batts et al., 1997; Van Oijen and Ewart,
1999; Constable et al., 1996). There is also some recent evidence that O3 and ultraviolet
radiation of wavelengths 280 to 320 nm (UV-B) interact in their effects on plant injury and
photosynthesis (Schnitzler et al., 1999), but additional research is needed to fully understand
how O3 interacts with multiple climate change factors.
9.5 EFFECTS-BASED AIR QUALITY EXPOSURE INDICES
Exposure indices are metrics that relate measured plant damage (i.e., reduced growth) to
monitored ambient O3 concentrations over time to provide a consistent metric for reviewing and
comparing exposure-response effects obtained from various studies. No new information is
available since 1996 that alters the basic conclusions put forth in the 1996 O3 AQCD. The 1996
AQCD (U.S. Environmental Protection Agency, 1996) focused on the research used to develop
various exposure indices to quantify growth and yield effects in crops, perennials, and trees
(primarily seedlings) and not foliar injury. The proposed indices included various functional and
9-11
-------
statistical summaries of monitored hourly O3 concentrations over designated time periods. The
indices were developed through regression analyses of earlier exposure studies, which was
accomplished by ordering the measured responses of growth and/or yield of crops and tree
(seedling) species in response to O3. The indices' development focused on consideration and
inclusion of some, but not all, the factors that affect O3 uptake and expression of effects (e.g.,
Leeetal., 1988).
In the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996b), it was concluded,
based on the best available data, that those O3 exposure indices that cumulated differentially
weighted hourly concentrations were the best candidates for relating exposure to plant growth
response. Greater weight was given to higher O3 concentrations and daylight hours, and it was
noted that the timing of peak concentrations and maximum plant conductance was critical in
determining exposure impact on plants. Various weighting functions were used, including
threshold-weighted (e.g., SUM06) and continuous sigmoid-weighted (e.g., W126) functions.
Based on statistical goodness-of-fit tests, these cumulative, concentration-weighted indices could
not be differentiated one from another. Additional statistical forms for O3 indices have been
discussed in Lee et al. (1988) and in Chapter 3 of this AQCD. A detailed discussion of effects-
based O3 exposure indices research since 1996 can be found in Annex Section AX9.4.
The few studies that have been published since the 1996 O3 AQCD continue to support the
earlier conclusions, including the importance of peak concentrations, and the duration and
occurrence of O3 exposures in altering plant growth and yield. In addition, a large body of new
research, mostly out of Europe, addresses the need for an index related to the actual uptake of O3
by the plant and the flux of O3 from the atmosphere to the O3-affected plant tissues. Despite
additional research linking estimates of flux with plant response since 1996, information is still
insufficient to identify a flux-based model that incorporates the necessary complexity across
space and time to be non-site- or non-species-specific. Based on the current state of knowledge,
exposure indices that cumulate and differentially weight the higher hourly average O3
concentrations, but include the mid-level values, still represent the best approach for relating
vegetation effects to O3 exposure in the United States.
9-12
-------
The new studies substantiated earlier conclusions on the role of exposure components
including concentration, duration, and exposure patterns in determining plant growth response
to O3 (Yun and Laurence, 1999; Oksanen and Holopainen, 2001). Recent studies using different
exposure patterns have confirmed earlier studies on the role of higher concentrations and
exposure duration (Nussbaum et al., 1995). A role for higher concentrations is inferred based on
improved air quality in regions in the western United States (Lefohn and Shadwick, 2000).
For example, the O3 reductions in the San Bernardino Mountain area since the late 1970s are
associated with reductions in the higher hourly average O3 concentrations, the number of hours
of concentrations >0.95 ppm, and the cumulative concentration-weighted exposure index (Lee
et al., 2003). The mid-range concentrations appeared to be relatively unchanged over the period
of 1980 to 2000. General forest improvement has been reported following a decrease of O3
along a decreasing gradient of exposure (Miller and Rechel, 1999; Arbaugh et al., 2003; Tingey
et al., 2004). These studies suggest that the focus should be on the higher O3 concentrations,
while including the lower levels, when estimating the effects of O3 precursor emission reduction
strategies on vegetation. The area has also experienced increasing deposition of N over the same
time period and, by many indicators, the soil is considered to be N-saturated (Fenn et al., 1996).
The relative role of N, however, in the measured or simulated tree response in the area has not
been quantified.
New studies have demonstrated the potential disconnection of peak events and maximal
stomatal conductance at xeric to mesic sites in California (Grulke et al., 2002; Panek et al., 2002;
Panek, 2004). In addition, a few studies have indicated that O3 uptake during nighttime hours is
greater than previously thought (Grulke et al., 2004; Massman, 2004), and a review of the
literature suggests a large number of species exhibit some degree of conductance at night
(Musselman and Minnick, 2000). These studies suggest the need for a reconsideration of
cumulating exposure 24 h/day and not just during daylight hours in O3 exposure index
determinations. This lack of coincidence in temporal patterns of conductance and peak ambient
concentrations introduces uncertainty in assessing the impact of O3. The use of an exposure
index that does not consider regionally unique climate and site factors modifying stomatal
conductance may, as a result, under- or overestimate growth effects. The shortcomings of an
9-13
-------
ambient exposure-based index is especially apparent when assessing the potential impact of O3
across broad climatic regions of the United States or Europe. Various means to overcome this
potential problem were addressed in several new studies, and one solution would be to add other
components to the present statistical summaries of exposure indices (e.g., meteorological) to
develop flux-based indices. However, the increased biological and meteorological information
in these indices may make them more regional in their applicability.
A number of studies have taken a flux-based approach to improve upon the ambient air
concentration-based (i.e., exposure indices) approach as a means to address the issue of
assessing risk of O3 across different climatic regions. The European acceptance and use of the
flux-based critical values is, in part, a recognition of the landscape scaling problems associated
with ambient exposure-based indices. A great deal of progress has occurred in developing and
testing stomatal models that may be generally applicable across certain vegetation types
(Emberson et al., 2000; Pleijel et al., 2000; Danielsson et al., 2003; Griinhage and Jager, 2003;
Matyssek et al., 2004). While a flux-based approach is preferred, a cautionary argument has
been advanced in a few publications, based on the nonlinear relationship between O3 uptake and
foliar injury (growth was not assessed). The concern is that not all O3 stomatal uptake results in
a yield reduction, which depends to some degree on the amount of internal detoxification
occurring with each particular species. Those species having high amounts of detoxification
potential may, in fact, show little relationship between O3 stomatal uptake and plant response
(Musselman and Massman, 1999).
Given the current state of knowledge and the best available data, exposure indices that
cumulate and differentially weight the higher hourly average concentrations and also include the
mid-level values continue to offer the most defensible approach for use in developing response
functions and comparing studies, as well as for defining future indices for vegetation protection.
A large database exists that has been used for establishing exposure-response relationships;
however, at this time, such a database does not exist for relating O3 flux to growth response.
It is anticipated that, as the overlapping relationships of conductance, concentration, and
defense mechanisms are better defined, the flux-based indices may be able to predict vegetation
injury and/or damage across varied landscapes and climates with more accuracy than the
9-14
-------
exposure-response models. However, that is not the case at this time. The translation of these
indices from research and assessment tools to air quality standards has the additional need to be
simple, understandable, and adaptive to a manageable monitoring program.
9.6 OZONE EXPOSURE-PLANT RESPONSE RELATIONSHIPS
Data published since 1996 continue to support the conclusions of previous O3 AQCDs that
there is strong evidence that ambient concentrations of O3 cause foliar injury and growth and
yield damage to numerous common and economically valuable plant and tree species. For
annual vegetation, the data summarized in Table AX9-16 (see Annex Section AX9.5) show a
range of growth and yield responses both within species and among species. Nearly all of these
data were derived from OTC studies, with only two studies using open-air systems in the United
Kingdom (Ollerenshaw and Lyons, 1999; Ollerenshaw et al., 1999). It continues to be difficult
to compare studies that report O3 exposure using different indices, such as AOT40, SUM06,
W126, or 7-h or 12-h mean values. The AOT40, SUM06, and W126 indices are defined as
follows:
AOT40: the seasonal sum of the difference between an hourly concentration above
the threshold value of 40 ppb, minus the threshold value of 40 ppb;
SUM06: the seasonal sum of hourly concentrations at or above the threshold value
of 60 ppb; and
W126: a sigmoid functional weighting of all hourly concentrations for the season.
When such index comparisons can be made, the results of recent research confirm earlier
results assessed in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996).
A summary of earlier literature concluded that a 7-h, 3-month mean of 49 ppb O3 corresponding
to a SUM06 exposure of 26 ppm-h, would cause a 10% loss in 50% of 49 experimental cases
(Tingey et al., 1991). Recent data summarized in Table AX9-16 support this conclusion and,
more generally, indicate that ambient O3 exposures can reduce the growth and yield of annual
species. Some annual species such as soybean [Glycine max (L.) Merr.] are more sensitive, and
9-15
-------
greater losses in such species may be expected (Table 9-16). Thus, the recent scientific literature
supports the conclusions of the 1996 O3 AQCD that ambient O3 concentrations are reducing the
yield of major crops in the United States.
Much research in Europe has used the cutoff-concentration weighted, cumulative-exposure
statistic AOT40, and substantial effort has gone into developing "Level-1" critical levels for
vegetation using this index. Based on regression analysis of 15 OTC studies of spring wheat,
including one U.S. study and 14 studies from locations ranging from southern Sweden to
Switzerland, an AOT40 value of 5.7 ppm-h was found to correspond to a 10% yield loss, and a
value of 2.8 ppm-h corresponded to a 5% yield loss (Fuhrer et al., 1997). Because a 4 to 5%
decrease could be detected with a confidence level of 99%, 3 ppm-h was selected by the
European Union as the AOT40 critical level in 1996 (Karenlampi and Skarby, 1996).
In addition to reductions in crop yield, O3 may also reduce the quality or nutritive value of
annual species. Many studies have found O3 effects in various measures of plant organs that
affect quality, with most of those studies focusing on characteristics important for food or
fodder. These studies indicate that ambient O3 may have economically important effects on the
quality of crop and forage species. Previous O3 AQCDs have concluded that visible symptoms
on marketable portions of crops and ornamental plants can occur with seasonal 7-h mean O3
exposures of 40 to 100 ppb (U.S. Environmental Protection Agency, 1978; 1986; 1992; 1996).
The recent scientific literature does not refute this conclusion.
The use of OTCs may reverse the usual vertical gradient in O3 that occurs within a few
meters above the ground surface (see Annex Section AX9.1). This reversal suggests that OTC
studies may, to some degree, overestimate the effects of an O3 concentration as measured several
meters above the ground. However, such considerations do not invalidate the conclusion of the
1996 O3 AQCD that ambient O3 exposures are sufficient to reduce the yield of major crops in the
United States. Recent results from the non-chambered FACE systems, having different vertical
gradients than chambered systems, report similar results for foliar injury, growth response and
yield reductions as reported for the chambered OTC systems ( Karnosky et al., 1999; Isebrands
et al., 2001; Morgan et al., 2006).
9-16
-------
As found for single-season agricultural crops, yields of multiple-year forage crops are
reduced at O3 exposures that occur over large areas of the United States. This result is similar to
that reported in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996). When
species are grown in mixtures, O3 exposure can result in the increased growth of O3-tolerant
species, which exacerbates the growth decrease of O3-sensitive species. Because of this
competitive interaction, the total growth of the mixed-species community may not be affected
by O3 exposure. However, in some cases, mixtures of grasses and clover species have shown
significant decreases in total biomass growth in response to O3 exposure in studies in the United
States and in Sweden. In Europe, a provisional AOT40 critical level of 7 ppm-h over 6 months
has been proposed by the European Union as a value to protect sensitive herbaceous perennial
plant species from the adverse effects of O3.
For deciduous tree species, recent evidence from FACE and OTC studies supports results
observed in previous OTC studies. For example, a series of O3-FACE studies was undertaken in
Rhinelander, WI (Isebrands et al., 2000, 2001). These studies showed that O3 symptom
expression was generally similar in OTC and FACE studies and in studies at sites along an
ambient O3 gradient, supporting the previously observed variation among aspen clones obtained
using OTCs (Karnosky et al., 1999). As has been observed in previous O3 AQCDs, root growth
is often found to be the most sensitive biomass response to O3.
Results of studies since 1996 support the conclusion of the 1996 O3 AQCD (U.S.
Environmental Protection Agency, 1996) that deciduous trees are generally less sensitive to O3
than are most annual plants, with the exception of a few very sensitive genera such as sensitive
clones or genotypes of the Populus genera and sensitive species such as black cherry (Prunus
serotina Ehrh.). However, the data presented in Table AX9-18 (see Annex Section AX9.5)
suggest that ambient exposures that occur in the United States can sometimes reduce the growth
of seedlings of deciduous species. Results from some multiyear studies have shown a pattern of
increased effects in subsequent years (Hogsett et al., 1989; Anderson et al., 1997; Karlsson et al.,
1995). In some cases, however, growth decreases due to O3 have become less significant or even
disappeared over time (Karlsson et al., 2002). While some mature trees show greater O3
9-17
-------
sensitivity than do seedlings in physiological parameters such as net photosynthetic rate, these
effects may not translate into measurable reductions in biomass growth (Samuelson and Kelly,
2001). However, because even multiyear experiments do not expose trees to O3 for more than a
small fraction of their life span and because competition may, in some cases, exacerbate the
effects of O3 on individual species, determining O3 effects on mature trees remains a significant
challenge.
In Europe, a Level I critical level has been set for forest trees based on OTC studies of
European beech (Fagus sylvatica L.) seedlings and is defined as an AOT40 value of 10 ppm-h
for daylight hours for a 6-month growing season (Karenlampi and Skarby, 1996). However,
other studies show that other species, such as silver birch(Betulapendula Roth.), may be more
sensitive to O3 than beech (Paakkonen et al., 1996).
As found for other tree species, various evergreen tree species and genotypes have widely
varying O3 sensitivities. Based on OTC studies with seedlings, major evergreen species in the
United States are generally less sensitive than are most deciduous trees, and slower-growing
evergreen species are less sensitive than are faster-growing species. There is evidence that plant
interaction stress, such as competition stress, may increase the O3 sensitivity of trees. As in
studies of deciduous species, most experiments with evergreen species have only covered a very
small portion of the life span of a tree and have been conducted with seedlings, so estimating
effects for mature evergreens is difficult.
For all types of perennial vegetation, cumulative effects over more than one growing
season may be important, and studies for only a single season may underestimate effects.
Mature trees may be more or less sensitive to O3 than are seedlings, depending on the species,
but information on physiological traits can be used to predict such differences in specific cases.
In some cases, mature trees may be more sensitive to O3 than seedlings due to differences in gas
exchange rates or growth rates, to greater cumulative exposures, or to the interaction of O3 with
other stressors.
9-18
-------
9.7 EFFECTS OF OZONE EXPOSURE ON NATURAL ECOSYSTEMS
There is evidence that tropospheric O3 is an important stressor of ecosystems, with
documented impacts on the biotic condition, ecological processes, and chemical and/or physical
nature of natural ecosystems (See Table AX9-22; Annex Section AX9.6; Figure 9-1). In turn,
the effects of O3 on individual plants and processes are scaled up through the ecosystem,
affecting processes such as energy and material flow, inter- and intraspecies competition, and net
primary productivity (NPP). Thus, effects on individual keystone species and their associated
microflora and fauna, effects that have been shown experimentally, may cascade through the
ecosystem to the landscape level, although this has not yet been demonstrated. By affecting
water balance, cold hardiness, tolerance to wind and by predisposing plants to insect and disease
pests, O3 has the potential to impact the occurrence and severity of natural disturbances (e.g.,
fire, erosion). Despite the possible occurrence of such effects, there are essentially no instances
where ecosystem-level, highly integrated studies have conclusively shown that O3 is indeed
altering ecosystem structure and/or function.
Systematic injury surveys demonstrate that foliar injury occurs to O3-sensitive species in
many regions of the United States (Chappelka et al., 1997; Campbell et al., 2000; Coulston et al.,
2003; Smith et al., 2003) and Europe (Braun et al., 1999). However, the frequent lack of
correspondence between foliar symptoms and growth effects means that other methods must be
used to estimate the regional effects of O3 on tree growth rates (e.g., Rebbeck, 1996; Kouterick
et al., 2000). Investigations of the radial growth of mature trees in combination with data
from many controlled studies with seedlings and a few studies with mature trees suggest that
ambient O3 is reducing the growth of mature trees in some locations (Somers et al., 1998).
Studies using models based on tree physiology and forest stand dynamics suggest that modest
effects of O3 on growth may accumulate over time and may interact with other stressors
(Laurence et al., 2001, 2003). For mixed-species stands, such models predict that overall stand
growth rate is generally not likely to be affected. However, competitive interactions among
species may change as a result of growth reductions of O3-sensitive species (Weinstein et al.,
2001). These results suggest that O3 exposure over decades may alter the species composition of
forests in some regions.
9-19
-------
ffl
Hydrolic Alteration
Habitat Conversion
Habitat Fragmantation
Climate Change
Invasive Normative Species
Turbidity/Sedimentation
Pesticides
Disease/Pest Outbreaks
Nutrient Pulses
Metals
Dissolved Oxygen Depletion
Ozone (Tropospheric)
Nitrogen Oxides
Nitrates
Hydrolie Alteration
Habitat Conversion
Habitat Fragmentation
Climate Change
Over-Harvesting Vegetation
Large-Scale Invasive
Species introduction
Large-Scale Disease/Pest
Outbreaks
Hydroiic Alteration
Habitat Conversion
Climate Change
Over-Harvesting Vegetation
Diseasa/Pest Outbreaks
Altered Fire Regime
Hydroiic Alteration
Habitat Conversion
Climate Change
Turbidity/Sedimentation
Pesticides
Nutrient Pulses
Metals
Dissolved Oxygen Depletion
Ozone (Tropospheric)
Nitrogen Oxides
Nitrates
Suifates
Salinity
Acidic Deoosition
V)
Hydroiic Alteration
Habitat Conversion
Climate Change
Pesticides
Disease/Pest Outbreaks
Nutrient Pulses
Dissolved Oxygen Depletion
Nitrogen Oxides
Nitrates
Syifates
Figure 9-1. Common anthropogenic stressors and the essential ecological attributes
they affect.
Source: Modified from Young and Sanzone (2002).
Despite the increased understanding of possible ecosystem effects of O3, the database
demonstrating and quantifying the degree to which O3 is altering natural ecosystems is sparse.
Much of the speculative O3 impact on ecosystems must be inferred from a number of case
studies of forest-plot field-based data reporting on a number of different species. One means to
discuss our current knowledge is by listing the areas in which more information is needed, as
shown below:
9-20
-------
Ecosystem processes. Very little is known about the effects of O3 on water, carbon, and
nutrient cycling, particularly at the stand and community levels. Effects on below-ground
ecosystem processes in response to O3 exposure alone and in combination with other stressors
are critical to projections at the watershed and landscape levels. Little is yet known about the
effects of O3 on structural or functional components of soil food webs or how these impacts
could affect plant species diversity (Andersen, 2003).
Biodiversity and genetic diversity. The study of genetic aspects of O3 impacts on natural
ecosystems has been largely based on correlations, and it remains to be shown conclusively
whether O3 affects biodiversity or genetic diversity (Davison and Barnes, 1998; Pitelka, 1988;
Winner et al., 1991). Studies of competitive interactions under elevated O3 levels are needed
(Laurence and Andersen, 2003). Reexaminations via new sampling of population studies to
bring a time component into previous studies showing spatial variability in population responses
to O3 are also needed. These studies could be strengthened by modern molecular methodologies
to quantify impacts on diversity.
Natural ecosystem interactions with the atmosphere. Little is known about feedbacks
between O3 and climate change on the production of volatile organic compounds (VOCs), which,
in turn, could affect O3 production (Fuentes et al., 2001). At moderate to high O3 exposure
sites, aberrations in stomatal behavior could significantly affect individual tree water balance
in O3-sensitive trees, and if the sensitive tree species is dominant, the hydrologic balance at the
watershed and landscape levels could also be affected. This has not been addressed in any
model, because O3-exposure effects, if included at all in the modeling effort, have often assumed
a linear relationship between assimilation and stomatal conductance. Interaction studies with
other components of global change (i.e., warming, increasing atmospheric CO2, N deposition) or
with various biotic stressors are needed to better predict complex interactions likely to occur in
the future (Laurence and Andersen, 2003). Whether O3 will negate the positive effects of an
elevated CO2 environment on plant carbon and water balances is not yet known; nor is it known
if these effects will scale up through the ecosystem. How O3 affects the progress of pest
epidemics and insect outbreaks as concentrations increase is also unclear (Ball et al., 1998).
Information is needed with regard to the impact of O3 on plant pest and insect reproductive
9-21
-------
processes and reproductive development under realistic field or forest conditions, as well as
examination of reproductive effects under interacting pollutants (Black et al., 2000).
Scaling. The vast majority of O3 tree studies have been conducted with young, immature
trees and with trees that have not yet formed a closed canopy. Questions remain as to the
comparability of O3 effects on juvenile versus mature trees, and on trees grown in the open
versus those in a closed forest canopy, in a competitive environment (Chappelka and Samuelson,
1998; Kolb and Matyssek, 2001; Samuelson and Kelly, 2001). Merging the effects of O3 across
spatial scales is also difficult. Scaling responses of a single or a few plants to effects on
communities and ecosystems are complicated matters that will require a combination of
manipulative experiments with model ecosystems; community and ecosystem studies along
natural O3 gradients; and extensive modeling efforts to project landscape-level, regional, national
and international impacts of O3. Linking these various studies via impacts on common research
quantification across various scales using measures of such factors as leaf area index or spectral
reflective data, which could eventually be remotely sensed (Kraft et al., 1996; Panek et al.,
2003), would provide powerful new tools for ecologists.
Identifying endpoints. In general, methodologies to determine the important values of
services and benefits derived from natural ecosystems are lacking. Identifying and quantifying
factors that could be used in comprehensive risk assessment for O3 effects on natural ecosystems
would increase societal awareness of the importance of protecting ecosystems (Heck et al.,
1998).
9.8 ECONOMICS
Substantial progress has been made over the past two decades in our understanding of the
effects of O3 and other oxidants on vegetation, particularly for agriculturally important plant
species (see Annex Section AX9.7 for a more detailed discussion). The physical and economic
effects on agriculture are well documented and provide useful information for consideration in
establishing air quality standards for crops (e.g., Spash, 1997).
9-22
-------
Since the completion of the National Crop Loss Assessment Newwork (NCLAN) program
in the late 1980s, the number of economic assessments of air pollution studies focusing on
terrestrial ecosystems in general, and agriculture in particular, has declined. For example,
33 economic studies of O3 and other air pollutant effects on U.S. crops were published in
peer-reviewed journal outlets from 1980 to 1990 (Spash, 1997). However, in preparing this
section of the current O3 AQCD, only four peer-reviewed economic assessments addressing
vegetation in the United States were found for the decade of 1991 to 2000. In addition, one
peer-reviewed article (Kuik et al., 2000) was found dealing with agriculture in the Netherlands.
Recent interest in global climate change and the potential effects of global warming on O3 and
other photochemical oxidants has renewed interest in the effects of air pollution on both
managed and unmanaged terrestrial ecosystems (Adams et al., 1998). In addition, concern is
growing regarding the effects of air pollutants on natural ecosystems and on the services they
provide (Daily, 1997). Unfortunately, this interest has not yet translated into additional
peer-reviewed publications addressing O3 or other air pollutants' effects on ecosystems.
A study by Murphy et al. (1999) of the economic effects of tropospheric O3 on U.S.
agriculture is of note here, because it confirms the general magnitude of economic effects
reported by the two key studies performed a decade earlier (Adams, 1985, 1986). Specifically,
Murphy et al. (1999) evaluated benefits to eight major crops associated with several scenarios
concerning the reduction or elimination of O3 precursor emissions from motor vehicles in the
United States. Their analysis reported a $2.8 billion to $5.8 billion (1990 dollars) benefit from
complete elimination of O3 exposures from all sources, i.e., ambient O3 reduced to a background
level assumed to be 0.025 to 0.027 ppm. While the analytical framework is similar to Adams
et al. (1986) in the use of NCLAN-based yield response functions and a mathematical
programming-based economic optimization model, the study is novel in its focus on the role of
motor vehicle emissions of VOCs and nitrogen oxides (NOX) in total anthropogenic O3 levels.
The study is also notable in its careful attention to federal farm program effects, particularly the
deficiency payment component.
9-23
-------
A number of recent studies of air pollutant effects on tree species have appeared in the
literature. Some studies have reported changes in total biomass and focused on European
species (Kurczynska et al., 1997). Other studies have assessed changes in composition of forest
species (biodiversity) or forest health due to exposure to air pollutants (Bringmark and
Bringmark, 1995; McLaughlin and Percy, 1999; Vacek et al., 1999). As noted previously,
changes in forest biomass and composition are more difficult to value than marketable products.
However, measures of forest composition or health have implications for an area of increasing
policy concern, that being the effect of air pollutants and other environmental stressors on
unmanaged (natural) ecosystems and the services they provide (Goulder and Kennedy, 1997;
Pimentel et al., 1997). Considerable discussion has occurred among ecologists and economists
as to the appropriate means for valuing these services (Carpenter and Dixon, 1985; Anderson,
1990; Common and Perrings, 1992). A number of conceptual articles have been published on
this issue in both economic and ecological journals (Bergstrom, 1990; Suter, II, 1990; Castle,
1993; Pearce, 1993).
Effects on forests and natural ecosystems remain problematic because of limitations in
biological response data and economic methods. The problem is even more acute for valuing
natural ecosystem service flows. The current limitations surrounding forests and natural
ecosystems present a rich research agenda. Areas of greatest potential value in terms of regional
policymaking need to be prioritized. Such priority setting can be assisted by sensitivity analyses
with existing economic models. By measuring the changes in economic effects arising from
changes in key parameters, it is possible to identify those research data gaps most likely to affect
economic values.
9-24
-------
REFERENCES
Adams, R. M. (1986) Agriculture, forestry, and related benefits of air pollution control: a review and some
observations. Am. J. Agric. Econ. 68: 464-472.
Adams, R. M.; Hamilton, S. A.; McCarl, B. A. (1985) An assessment of the economic effects of ozone on U.S.
agriculture. J. AirPollut. Control Assoc. 35: 938-943.
Adams, R. M.; Hamilton, S. A.; McCarl, B. A. (1986) The benefits of pollution control: the case of ozone and U.S.
agriculture. Am. J. Agric. Econ. 68: 886-893.
Adams, R. M.; Kurd, B. H.; Lenhart, S.; Leary, N. (1998) Effects of global climate change on agriculture: an
interpretative review. Clim. Res. 11: 19-30.
Anderson, E. (1990) The ethical limitations of the market. Econ. Philos. 6: 179-205.
Andersen, C. P. (2003) Source-sink balance and carbon allocation below ground in plants exposed to ozone.
New Phytol. 157: 213-228.
Andersen, C. P.; Wilson, R.; Plocher, M.; Hogsett, W. E. (1997) Carry-over effects of ozone on root growth and
carbohydrate concentrations of ponderosa pine seedlings. Tree Physiol. 17: 805-811.
Andersen, C. P.; Hogsett, W. E.; Plocher, M.; Rodecap, K.; Lee, E. H. (2001) Blue wild-rye grass competition
increases the effect of ozone on ponderosa pine seedlings. Tree Physiol. 21: 319-327.
Arbaugh, M.; Bytnerowicz, A.; Grulke, N.; Fenn, M.; Poth, M.; Temple, P.; Miller, P. (2003) Photochemical smog
effects in mixed conifer forests along a natural gradient of ozone and nitrogen deposition in the San Bernardino
Mountains. Environ. Int. 29: 401-406.
Assmann, S. M. (2003) Open stomata 1 opens the door to ABA signaling inArabidopsis guard cells. Trends Plant
Sci. 8: 151-153.
Assmann, S. M.; Wang, X. Q. (2001) From milliseconds to millions of years: guard cells and environmental
responses. Curr. Opin. Plant Biol. 4: 421-428.
Ball, G. R.; Benton, J.; Palmer-Brown, D.; Fuhrer, J.; Skarby, L.; Gimeno, B. S.; Mills, G. (1998) Identifying factors
which modify the effects of ambient ozone on white clover (Trifolium repens) in Europe. Environ. Pollut.
103: 7-16.
Ball, G. R.; Palmer-Brown, D.; Fuhrer, J.; Skarby, L.; Gimeno, B. S.; Mills, G. (2000) Identification of non-linear
influences on the seasonal ozone dose-response of sensitive and resistant clover clones using artificial neural
networks. Ecol. Model. 129: 153-168.
Balls, G. R.; Palmer-Brown, D.; Sanders, G. E. (1996) Investigating microclimatic influences on ozone injury in
clover (Trifolium subterraneum) using artificial neural networks. New Phytol. 132: 271-280.
Barnes, J. D.; Wellburn, A. R. (1998) Air pollutant combinations. In: De Kok, L. J.; Stulen, L, eds. Responses of
plant metabolism to air pollution and global change. Leiden, The Netherlands: Backhuys Publishers;
pp. 147-164.
Batts, G. R.; Morison, J. I. L.; Ellis, R. H.; Hadley, P.; Wheeler, T. R. (1997) Effects of CO2 and temperature on
growth and yield of crops of winter wheat over four seasons. Eur. J. Agron. 7: 43-52.
Bergstrom, J. C. (1990) Concepts and measures of the economic value of environmental quality: a review.
J. Environ. Manage. 31: 215-228.
Black, V. J.; Black, C. R.; Roberts, J. A.; Stewart, C. A. (2000) Impact of ozone on the reproductive development of
plants. New Phytol. 147: 421-447.
Bonello, P.; Heller, W.; Sandermann, H., Jr. (1993) Ozone effects on root-disease susceptibility and defence
responses in mycorrhizal and non-mycorrhizal seedlings of Scots pine (Pinus sylvestris L). New Phytol.
124: 653-663.
Braun, S.; Rihm, B.; Schindler, C.; Fltickiger, W. (1999) Growth of mature beech in relation to ozone and nitrogen
deposition: an epidemiological approach. Water Air Soil Pollut. 116: 357-364.
Bringmark, E.; Bringmark, L. (1995) Disappearance of spatial variability and structure in forest floors - a distinct
effect of air pollution? Water Air Soil Pollut. 85: 761-766.
Broadmeadow, M. S. J.; Jackson, S. B. (2000) Growth responses of Quercuspetraea, Fraxinus excelsior and
Pinus sylvestris to elevated carbon dioxide, ozone and water supply. New Phytol. 146: 437-451.
Campbell, S.; Temple, P.; Pronos, J.; Rochefort, R.; Andersen, C. (2000) Monitoring for ozone injury in west coast
(Oregon, Washington, California) forests in 1998. Portland, OR: U.S. Department of Agriculture,
Forest Service, Pacific Northwest Research Station; general technical report no. PNW-GTR-495. Available:
http://www.fs.fed.us/pnw/gtrs.htm [11 April, 2003].
Cardoso-Vilhena, J.; Barnes, J. (2001) Does nitrogen supply affect the response of wheat (Triticum aestivum cv
Hanno) to the combination of elevated CO2 and O3? J. Exp. Bot. 52: 1901-1911.
9-25
-------
Carpenter, R. A.; Dixon, J. A. (1985) Ecology meets economics: a guide to sustainable development. Environment
27: 6-32.
Castle, E. N. (1993) A pluralistic, pragmatic and evolutionary approach to natural resource management. For. Ecol.
Manage. 56: 279-295.
Chappelka, A. H.; Samuelson, L. J. (1998) Ambient ozone effects on forest trees of the eastern United States:
a review. New Phytol. 139: 91-108.
Chappelka, A.; Renfro, J.; Somers, G.; Nash, B. (1997) Evaluation of ozone injury on foliage of black cherry
(Primus semtina) and tall milkweed (Asclepias exaltata) in Great Smoky Mountains National Park. Environ.
Pollut. 95: 13-18.
Colls, J. J.; Unsworth, M. H. (1992) Air pollution interactions with natural stressors. In: Barker, J. R.; Tingey, D. T.,
eds. Air pollution effects on biodiversity. New York, NY: Van Nostrand Reinhold; pp. 93-108.
Common, M.; Perrings, C. (1992) Towards an ecological economics of sustainability. Ecol. Econ. 6: 7-34.
Constable, J. V. H.; Taylor, G. E., Jr.; Laurence, J. A.; Weber, J. A. (1996) Climatic change effects on the
physiology and growth of Pinusponderosa: expectations from simulation modeling. Can. J. For. Res.
26: 1315-1325.
Coulston, J. W.; Smith, G. C.; Smith, W. D. (2003) Regional assessment of ozone sensitive tree species using
bioindicator plants. Environ. Monit. Assess. 83: 113-127.
Daily, G. C. (1997) Introduction: what are ecosystem services? In: Daily, G. C., ed. Nature's services: societal
dependence on natural ecosystems. Washington, DC: Island Press; pp. 1-10.
Danielsson, H.; Karlsson, G. P.; Karlsson, P. E.; Pleijel, J. (2003) Ozone uptake modelling and flux-response
relationships—an assessment of ozone-induced yield loss in spring wheat. Atmos. Environ. 37: 475-485.
Davison, A. W.; Barnes, J. D. (1998) Effects of ozone on wild plants. New Phytol. 139: 135-151.
Docherty, M.; Salt, D. T.; Holopainen, J. K. (1997) The impacts of climate change and pollution on forest pests.
In: Watt, A. D.; Stork, N. E.; Hunter, M. D., eds. Forests and insects. Chapman and Hall; pp. 229-247.
Dowding, P. (1988) Air pollutant effects on plant pathogens. In: Schulte-Hostede, S.; Darrall, N. M.; Blank, L. W.;
Wellburn, A. R., eds. Air pollution and plant metabolism. London, United Kingdom: Elsevier Applied Science
Publishers; pp. 329-355.
Emberson, L.; Ashmore, M. R.; Cambridge, H. M.; Simpson, D.; Tuovinen, J. P. (2000) Modelling stomatal ozone
flux across Europe. Environ. Pollut. 109: 403-413.
Fangmeier, A.; Bender, J.; Weigel, H. J.; Jager, H. J. (2002) Effects of pollutant mixtures. In: Bell, J. N. B.;
Treshow, M., eds. Air Pollution and Plant Life. 2nd ed. Chichester, United Kingdom: John Wiley & Sons Ltd.;
pp. 251-272.
Flagler, R. B. (1998) Recognition of air pollution injury to vegetation: a pictorial atlas. 2nd ed. Pittsburgh, PA: Air &
Waste Management Association.
Fluckiger, W.; Braun, S.; Hiltbrunner, E. (2002) Effects of air pollutants onbiotic stress. In: Bell, J. N. B.; Treshow,
M., eds. Air pollution and plant life. 2nd ed. Chichester, United Kingdom: John Wiley & Sons Ltd.; pp.
379-406.
Fredericksen, T. S.; Skelly, J. M.; Snyder, K. R.; Steiner, K. C. (1996) Predicting ozone uptake from meteorological
and environmental variables. J. Air Waste Manage. Assoc. 46: 464-469.
Fuentes, J. D.; Hayden, B. P.; Garstang, M.; Lerdau, M.; Fitzjarrald, D.; Baldocchi, D. D.; Monson, R.; Lamb, B.;
Geron, C. (2001) New directions: VOCs and biosphere-atmosphere feedbacks. Atmos. Environ. 35: 189-191.
Fuhrer, J. H. (1994) Effects of ozone on managed pasture. 1. Effects of open-top chambers on microclimate, ozone
flux and plant growth. Environ. Pollut. 86: 297-305.
Fuhrer, J.; Skarby, L.; Ashmore, M. R. (1997) Critical levels for ozone effects on vegetation in Europe. Environ.
Pollution 97: 91-106.
Goulder, L. H.; Kennedy, D. (1997) Valuing ecosystem services: philosophical bases and empirical methods. In:
Daily, G. C. Nature's Services: Societal Dependence on Natural Ecosystems. Washington, DC: Island Press;
pp. 23-47.
Grulke, N. E.; Preisler, H. K.; Rose, C.; Kirsch, J.; Balduman, L. (2002) O3 uptake and drought stress effects on
carbon acquisition of ponderosa pine in natural stands. New Phytol. 154: 621-631.
Grulke, N. E.; Alonso, R.; Nguyen, T.; Cascio, C.; Dobrowolski, W. (2004) Stomata open at night in pole-sized and
mature ponderosa pine: implications for O3 exposure metrics. Tree Physiol. 24: 1001-1010.
Grunhage, L.; Jager, H. J. (2003) From critical levels to critical loads for ozone: a discussion of a new experimental
and modelling approach for establishing flux-response relationships for agricultural crops and native plant
species. Environ. Pollut. 125: 99-110.
9-26
-------
Heck, W. W.; Furiness, C. S.; Cowling, E. B.; Sims, C. K. (1998) Effects of ozone on crop, forest, and natural
ecosystems: assessment of research needs. EM (October): 11-22.
Hogsett, W. E.; Tingey, D. T.; Hendricks, C.; Rossi, D. (1989) Sensitivity of western conifers to SO2 and seasonal
interaction of acid fog and ozone. In: Olson, R. K.; Lefohn, A. S., eds. Effects of air pollution on western forests
[an A&WMA symposium; June; Anaheim, CA]. Air Pollution Control Association; pp. 469-491
(APCA transactions series: no. 16).; pp. Anaheim,CA-Anaheim491.
Innes, J. L.; Skelly, J. M; Schaub, M. (2001) Ozone and broadleaved species A guide to the identification of ozone-
induced foliar injury. Bern, Switzerland: Paul Haupt Publishers.
Isebrands, J. G.; Dickson, R. E.; Rebbeck, J.; Karnosky, D. F. (2000) Interacting effects of multiple stresses on
growth and physiological processes in northern forest trees. In: Mickler, R. A.; Birsdey, R. A.; Horn, J., eds.
Responses of northern U.S. forests to environmental change. New York, NY: Springer-Verlag; pp. 149-180.
(Ecological studies: v. 139).
Isebrands, J. G.; McDonald, E. P.; Kruger, E.; Hendrey, G.; Percy, K.; Pregitzer, K.; Sober, J.; Karnosky, D. F.
(2001) Growth responses ofPopulus tremuloides clones to interacting carbon dioxide and tropospheric ozone.
Environ. Pollut. 115: 359-371.
Johnson, B. G.; Hale, B. A.; Ormrod, D. P. (1996) Carbon dioxide and ozone effects on growth of a legume-grass
mixture. J. Environ. Qual. 25: 908-916.
Karenlampi, L.; Ska'rby, L. (1996) Critical levels for ozone in Europe: testing and finalizing the concepts UN-ECE
workshop report. In: Proceedings of UN-ECE convention on long-range transboundary air pollution workshop;
April; Kuopio, Finland; Kupio, Finland: University of Kuopio, Department of Ecology and Environmental
Science.
Karnosky, D. F.; Mankovska, B.; Percy, K.; Dickson, R. E.; Podila, G. K.; Sober, J.; Noormets, A.; Hendrey, G.;
Coleman, M. D.; Kubiske, M.; Pregitzer, K. S.; Isebrands, J. G. (1999) Effects of tropospheric ozone on
trembling aspen and interaction with CO2: results from an O3-gradient and a FACE experiment. Water Air Soil
Pollut. 116:311-322.
Kolb, T. E.; Matyssek, R. (2001) Limitations and perspectives about scaling ozone impacts in trees. Environ. Pollut.
115:373-393.
Kopper, B. J.; Lindroth, R. L. (2003) Effects of elevated carbon dioxide and ozone on the phytochemistry of aspen
and performance of an herbivore. Oecologia 134: 95-103.
Kouterick, K. B.; Skelly, J. M.; Fredericksen, T. S.; Steiner, K. C.; Kolb, T. E.; Ferdinand, J. A. (2000) Foliar injury,
leaf gas exchange and biomass responses of black cherry (Primus serotina Ehrh) half-sibling families to ozone
exposure. Environ. Pollut. 107: 117-126.
Kraft, M.; Weigel, H. J.; Mejer, G. J.; Brandes, F. (1996) Reflectance measurements of leaves for detecting visible
and non-visible ozone damage to crops. J. Plant Physiol. 148: 148-154.
Krupa, S. V.; Tonneijck, A. E. G.; Manning, W. J. (1998) Ozone. In: Flagler, R. B., ed. Recognition of air pollution
injury to vegetation - a pictorial atlas. Air& Waste Management Association; pp. 2-11.
Krupa, S. V.; Nosal, M.; Peterson, D. L. (2001) Use of passive ozone O3 samplers in vegetation effects assessment.
Environ. Pollut. 112: 303-309.
Krupa, S. V.; Nosal, M.; Ferdinand, J. A.; Stevenson, R. E.; Skelly, J. M. (2003) A multi-variate statistical model
integrating passive sampler and meteorology data to predict the frequency distributions of hourly ambient ozone
(O3) concentrations. Environ. Pollut. 124: 173-178.
Kuik, O. J.; Helming, J. F. M.; Dorland, C.; Spaninks, F. A. (2000) The economic benefits to agriculture of a
reduction of low-level ozone pollution in The Netherlands. Eur. Rev. Agric Econ. 27: 75-90.
Kurczynska, E. U.; Dmuchowski, W.; Wloch, W.; Bytnerowicz. A. (1997) The influence of air pollutants on needles
and stems of scots pine (Pinus Sylvestris L.) trees. Environ. Pollut. 98: 325-334.
Kytoviita, M. M.; Le Thiec, D.; Dizengremel, P. (2001) Elevated CO2 and ozone reduce nitrogen acquisition by
Pinus halepensis from its mycorrhizal symbiont. Physiol. Plant. Ill: 305-312.
Landolt, W.; Gtinthardt-Goerg, M. S.; Pfenninger, L; Einig, W.; Hampp, R.; Maurer, S.; Matyssek, R. (1997) Effect
of fertilization on ozone-induced changes in the metabolism of birch (Betula pendula) leaves. New Phytol.
137: 389-397.
Langebartels, C.; Ernst, D.; Heller, W.; Lutz, C.; Payer, H. D.; Sandermann, H., Jr. (1997) Ozone responses of trees:
results from controlled chamber exposures at the GSF phytotron. In: Sandermann, H.; Wellburn, A. R.; Heath,
R. L., eds. Forest decline and ozone. New York, NY: Springer-Verlag; pp. 163-200 (Ecological studies: v. 127).
Laurence, J. A.; Andersen, C. P. (2003) Ozone and natural systems: understanding exposure, response, and risk.
Environ. Int. 29: 155-160.
9-27
-------
Laurence, J. A.; Amundson, R. G.; Kohut, R. I; Weinstein, D. A. (1997) Growth and water use of red spruce
(Picea mbens Sarg) exposed to ozone and simulated acidic precipitation for four growing seasons. For. Sci.
(Bethesda, Md.) 43: 355-361.
Laurence, J. A.; Retzlaff, W. A.; Kern, J. S.; Lee, E. H.; Hogsett, W. E.; Weinstein, D. A. (2001) Predicting the
regional impact of ozone and precipitation on the growth of loblolly pine and yellow poplar using linked
TREGRO andZELG models. For. Ecol. Manage. 146: 247-263.
Laurence, J. A.; Retzlaff, W. A.; Kern, J. S.; Lee, E. H.; Hogsett, W. E.; Weinstein, D. A. (2003) Corrigendum to
"Predicting the regional impact of ozone and precipitation on the growth of loblolly pine and yellow-poplar
using linked TREGRO and ZELIG models" [Forest Ecol Manage 146 (2001) 247-263]. For. Ecol. Manage.
174: 607.
Lee, E. H.; Tingey, D. T.; Hogsett, W. E. (1988) Evaluation of ozone exposure indices in exposure-response
modeling. Environ. Pollut. 53: 43-62.
Lee, E. H.; Tingey, D. T.; Hogsett, W. E.; Laurence, J. A. (2003) History of tropospheric ozone for the
San Bernardino Mountains of southern California, 1963-1999. Atmos. Environ. 37: 2705-2717.
Lefohn, A. S.; Shadwick, D. S. (2000) Differences in trending estimates in the United States using several ozone
metrics. In: Proceedings of the 93rd annual meeting of the Air & Waste Management Association; June;
Salt Lake City, UT; Pittsburgh, PA: Air & Waste Management Association; paper no. AS ld-645.
Maier-Maercker, U. (1998) Predisposition of trees to drought stress by ozone. Tree Physiol. 19: 71-78.
Manning, W. J.; Krupa, S. V. (1992) Experimental methodology for studying the effects of ozone on crops and trees.
In: Lefohn, A. S., ed. Surface level ozone exposures and their effects on vegetation. Chelsea, MI: Lewis
Publishers, Inc.; pp. 93-156.
Massman, W. J. (2004) Toward an ozone standard to protect vegetation based on effective dose: a review of
deposition resistance and a possible metric. Atmos. Environ. 38: 2323-2337.
Matyssek, R.; Wieser, G.; Nunn, A. J.; Kozovits, A. R.; Reiter, I. M.; Heerdt, C.; Winkler, J. B.; Baumgarten, M.;
Haberle, K. H.; Grams, T. E. E.; Werner, H.; Fabian, P.; Havranek, W. M. (2004) Comparison between AOT40
and ozone uptake in forest trees of different species, age and site conditions. Atmos. Environ. 38: 2271-2281.
McLaughlin, S.; Percy, K. (1999) Forest health in North America: some perspectives on actual and potential roles of
climate and air pollution. Water Air Soil Pollut. 116: 151-197.
Miller, P. R.; Rechel, J. (1999) Temporal changes in crown condition indices, needle litterfall, and collateral needle
injuries of Ponderosa and Jeffrey pines. In: Miller, P. R.; McBride, J. R., eds. Oxidant air pollution impacts in
the Montane forests of southern California: a case study of the San Bernardino Mountains. New York, NY:
Springer; pp. 164-178.
Mills, G. (2002) Modification of plant response by environmental conditions. In: Bell, J. N. B.; Treshow, M., eds.
Air pollution and plant life. 2nd ed. Chichester, United Kingdom: John Wiley & Sons Ltd.; pp. 343-358.
Mills, G.; Ball, G.; Hayes, F.; Fuhrer, J.; Skarby, L.; Gimeno, B.; De Temmerman, L.; Heagle, A.; Members of the
ICP Vegetation programme. (2000) Development of a multi-factor model for predicting the effects of ambient
ozone on the biomass of white clover. Environ. Pollut. 109: 533-542.
Momen, B.; Anderson, P. D.; Helms, J. A.; Houpis, J. L. J. (1997) Acid rain and ozone effects on gas exchange of
Pinusponderosa: a comparison between trees and seedlings. Int. J. Plant Sci. 158: 617-621.
Momen, B.; Anderson, P. D.; Helms, J. A. (1999) Temperature dependency of acid-rain effect on photosynthesis of
Pinus ponderosa. For. Ecol. Manage. 113: 223-230.
Morgan, P. B.; Mies, T. A.; Bollero, G. A.; Nelson, R. L.; Long, S. P. (2006) Season-long elevation of ozone
concentration to projected 2050 levels under fully open-air conditions substantially decreases the growth and
production of soybean. New Phytol.: in press.
Murphy, J. J.; Deluki, M. A.; McCubbin, D. R.; Kim, H. J. (1999) The cost of crop damage caused by ozone air
pollution from mo tor vehicles. J. Environ. Manage. 55: 273-289.
Musselman, R. C.; Massman, W. J. (1999) Ozone flux to vegetation and its relationship to plant response and
ambient air quality standards. Atmos. Environ. 33: 65-73.
Musselman, R. C.; Minnick, T. J. (2000) Nocturnal stomatal conductance and ambient air quality standards for
ozone. Atmos. Environ. 34: 719-733.
Nussbaum, S.; Fuhrer, J. (2000) Difference in ozone uptake in grassland species between open-top chambers and
ambient air. Environ. Pollut. 109: 463-471.
Nussbaum, S.; Geissmann, M.; Fuhrer, J. (1995) Ozone exposure-response relationships for mixtures of perennial
ryegrass and white clover depend on ozone exposure patterns. Atmos. Environ. 29: 989-995.
Nussbaum, S.; Bungener, P.; Geissmann, M.; Fuhrer, J. (2000) Plant-plant interactions and soil moisture might be
important in determining ozone impacts on grasslands. New Phytol. 147: 327-335.
9-28
-------
Oksanen, E.; Holopainen, T. (2001) Responses of two birch (Betula pendula Roth) clones to different ozone profiles
with similar AOT40 exposure. Atmos. Environ. 35: 5245-5254.
Ollerenshaw, J. H.; Lyons, T. (1999) Impacts of ozone on the growth and yield of field-grown winter wheat.
Environ. Pollut. 106: 67-72.
Ollerenshaw, J. H.; Lyons, T.; Barnes, J. D. (1999) Impacts of ozone on the growth and yield of field-grown winter
oilseed rape. Environ. Pollut. 104: 53-59.
Paakkonen, E.; Holopainen, T.; Karenlampi, L. (1996) Relationships between open-field ozone exposures and
growth and senescence of birch (Betula pendula and Betula pubescens). In: Karenlampi, L.; Skarby, L. eds.
Critical levels for ozone in Europe: testing and finalizing the concepts: UN-ECE workshop report. UN-ECE
Convention on Long-Range Transboundary Air Pollution; April; Kuopio, Finland; Kuopio, Finland: University
of Kuopio, Department of Ecology and Environmental Science; pp. 298-302.
Panek, J. A. (2004) Ozone uptake, water loss and carbon exchange dynamics in annually drought-stressed
Pinusponderosa forests: measured trends and parameters for uptake modeling. Tree Physiol. 24: 277-290.
Panek, J.; Kurpius, M. R.; Goldstein, A. H. (2002) An evaluation of ozone exposure metrics for a seasonally
drought-stressed ponderosa pine ecosystem. Environ. Pollut. 117: 93-100.
Panek, J. A.; Baldocchi, D. D.; Goldstein, A. H. (2003) The need for spatially and functionally integrated models of
ozone deposition to Sierra Nevada forests. In: Bytnerowicz, A.; Arbaugh, M. J.; Alonso, R., eds. Ozone air
pollution in the Sierra Nevada: distribution and effects on forests. New York, NY: Elsevier Science Ltd;
pp. 325-357. (Developments in environmental science: v. 2).
Pearce, D. W. (1993) Economic values and the natural world. London, United Kingdom: Earthscan
Publications, Ltd.
Pell, E. J.; Schlagnhaufer, C. D.; Arteca, R. N. (1997) Ozone-induced oxidative stress: mechanisms of action and
reaction. Physiol. Plant. 100: 264-273.
Percy, K. E.; Awmack, C. S.; Lindroth, R. L.; Kubiske, M. E.; Kopper, B. J.; Isebrands, J. G.; Pregitzer, K. S.;
Hendry, G. R.; Dickson, R. E.; Zak, D. R.; Oksanen, E.; Sober, J.; Harrington, R.; Karnosky, D. F. (2002)
Altered performance of forest pests under atmospheres enriched with CO2 and O3. Nature (London)
420: 403-407.
Pimentel, D.; Wilson, C.; McCullum, C.; Huang, R.; Dwen, P.; Flack, J.; Tran, Q.; Saltman, T.; Cliff, B. (1997)
Economic and environmental benefits of biodiversity. BioScience 47: 747-757.
Pitelka, L. F. (1988) Evolutionary responses of plants to anthropogenic pollutants. Trends Ecol. Evol. 3: 233-236.
Pleijel, H.; Danielsson, H.; Karlsson, G. P.; Gelang, J.; Karlsson, P. E.; Sellden, G. (2000) An ozone flux-response
relationship for wheat. Environ. Pollut. 109: 453-462.
Rao, M. V.; Davis, K. R. (2001) The physiology of ozone induced cell death. Planta 213: 682-690.
Rebbeck, J. (1996) Chronic ozone effects on three northeastern hardwood species: growth and biomass. Can. J. For.
Res. 26: 1788-1798.
Rowland-Bamford, A. J. (2000) Plant responses to changing carbon dioxide and temperature. In: Singh, S. N., ed.
Trace gas emissions and plants. Dordrecht, The Netherlands: Kluwer Academic Publishers; pp. 63-74.
Samuelson, L.; Kelly, J. M. (2001) Scaling ozone effects from seedlings to forest trees. Tansley review no 21.
NewPhytol. 149:21-41.
Sandermann, H., Jr. (1998) Ozone: an air pollutant acting as a plant-signaling molecule. Naturwissenschaften
85: 369-375.
Sandermann, H., Jr. (2000) Ozone^iotic disease interactions: molecular biomarkers as a new experimental tool.
Environ. Pollut. 108: 327-332.
Sayre, R. G.; Fahey, T. J. (1999) Effects of rainfall acidity and ozone on foliar leaching in red spruce Picea mbens.
Can. J. For. Res. 29: 486-496.
Schmieden, U.; Wild, A. (1995) The contribution of ozone to forest decline. Physiol. Plant. 94: 371-378.
Schnitzler, J. P.; Langebartels, C.; Heller, W.; Liu, J.; Lippert, M.; Dohring, T.; Bahnweg, G.; Sandermann, H.
(1999) Ameliorating effect of UV-B radiation on the response of Norway spruce and Scots pine to ambient
ozone concentrations. Global Change Biol. 5: 83-94.
Schraudner, M.; Langebartels, C.; Sandermann, H., Jr. (1996) Plant defence systems and ozone. Biochem. Soc.
Trans. 24: 456-461.
Smith, G.; Coulston, J.; Jepsen, E.; Prichard, T. (2003) A national ozone biomonitoring program - results from field
surveys of ozone sensitive plants in northeastern forests (1994-2000). Environ. Monit. Assess. 87: 271-291.
Somers, G. L.; Chappelka, A. H.; Rosseau, P.; Renfro, J. R. (1998) Empirical evidence of growth decline related to
visible ozone injury. For. Ecol. Manage. 104: 129-137.
9-29
-------
Spash, C. L. (1997) Assessing the economic benefits to agriculture from air pollution control. J. Econ. Surv.
11:47-70.
Suter, G. W., II. (1990) Endpoints for regional ecological risk assessments. Environ. Manage. (N. Y.) 14: 9-23.
Tingey, D. T.; Hogsett, W. E.; Lee, E. H.; Herstrom, A. A.; Azevedo, S. H. (1991) An evaluation of various
alternative ambient ozone standards based on crop yield loss data. In: Berglund, R. L.; Lawson, D. R.; McKee,
D. J. eds. Tropospheric ozone and the environment: papers from an international conference; March 1990; Los
Angeles, CA; Pittsburgh, PA: Air & Waste Management Association (A&WMA transactions series no. TR-19);
pp. 272-288.
Tingey, D. T.; Hogsett, W. E.; Lee, E. H.; Laurence, J. A. (2004) Stricter ozone ambient air quality standard has
beneficial effect on ponderosa pine in California. Environ. Manage. 34: 397-405.
Topa, M. A.; Vanderklein, D. W.; Corbin, A. (2001) Effects of elevated ozone and low light on diurnal and seasonal
carbon gain in sugar maple. Plant Cell Environ. 24: 663-677.
U.S. Environmental Protection Agency (1978) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report no. EPA-600/8-78-004. Available from: NTIS, Springfield, VA; PB80-124753.
U.S. Environmental Protection Agency (1986) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA;
PB87-142949.
U.S. Environmental Protection Agency (1992) Summary of selected new information on effects of ozone on health
and vegetation: supplement to 1986 air quality criteria for ozone and other photochemical oxidants. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; report no. EPA/600/8-88/105F. Available from: NTIS, Springfield, VA; PB92-235670.
U.S. Environmental Protection Agency (1996) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
http://cfpub2.epa.gov/ncea/.
Vacek, S.; Bastl, M.; Leps, J. (1999) Vegetation changes in forests of the Krkonose Mountains over a period of air
pollution stress (1980-1995). Plant Ecol. 143: 1-11.
VanOijen, M.; Ewart, F. (1999) The effects of climatic variation in Europe on the yield response of spring wheat cv
Minaret to elevated CO2 and O3: an analysis of open-top chamber experiments by means of two crop growth
simulation models. Eur. J. Agron. 10: 249-264.
Ward, E. R.; Uknes, S. J.; Williams, S. C.; Dincher, S. S.; Wiederhold, D. L.; Alexander, D. C.; Ahi-Goy, P.;
Metraux, J. P.; Ryals, J. A. (1991) Coordinate gene activity in response to agents that induce systemic acquired
resistance. Plant Cell 3: 1084-1094.
Weinstein, D. A.; Gollands, B.; Retzlaff, W. A. (2001) The effects of ozone on a lower slope forest of the Great
Smoky Mountain National Park: simulations linking an individual tree model to a stand model. For. Sci.
47: 29-42.
Whitfield, C. P.; Davison, A. W.; Ashenden, T. W. (1998) The effects of nutrient limitation on the response of
Plantago major to ozone. New Phytol. 140: 219-230.
Winner, W. E.; Coleman, J. S.; Gillespie, C.; Mooney, H. A.; Pell, E. J. (1991) Consequences of evolving resistance
to air pollutants. In: Taylor, G. E.; Pitelka, L. F.; Clegg, M. T., eds. Ecological genetics and air pollution.
Springer-Verlag; pp. 177-202.
Wu, Y. X.; Tiedemann, A. V. (2002) Impact of fungicides on active oxygen species and antioxidant enzymes in
spring barley (Hordeum vulgare L) exposed to ozone. Environ. Pollut. 116: 37-47.
Yun, S. C.; Laurence, J. A. (1999) The response of sensitive and tolerant clones ofPopulus tremuloides to dynamic
ozone exposure under controlled environmental conditions. New Phytol. 143: 305-313.
9-30
-------
10. THE ROLE OF TROPOSPHERIC OZONE
IN UVB-RELATED HUMAN HEALTH OUTCOMES
AND IN CLIMATE CHANGE
10.1 INTRODUCTION
In addition to directly affecting human health, vegetation and ecosystems, tropospheric
ozone (O3) influences the ground-level flux of solar ultraviolet (UV) radiation and alters the
Earth's radiative balance by functioning as a greenhouse gas, therefore contributing to climate
change. This chapter addresses these effects.
10.2 THE ROLE OF TROPOSPHERIC OZONE IN DETERMINING
GROUND-LEVEL UV-B FLUX
Atmospheric O3 plays a crucial role in reducing the exposure of living organisms to solar
UV radiation. The stratospheric O3 layer is responsible for nearly all of this shielding effect, as
90% or more of the total atmospheric burden of O3 is located there. Specific quantification of
the role that tropospheric O3 plays in screening the earth's surface from harmful UV radiation
has not been accomplished, to date.
Reasonable estimation of the importance of ground-level O3 in US urban and suburban
areas, i.e., the fraction of tropospheric O3 subject to regulation by the US NAAQS, in altering
human UVB exposure and the incidence of UVB-induced human diseases requires an explicit
accounting for key factors that influence (a) ground-level flux in US urban and suburban areas,
(b) the probability of UV-B exposure experienced by sensitive populations, and (c) links
between exposure and the incidence rates of UVB-induced diseases. This section summarizes,
in this order, available information on such factors.
10.2.1 Factors Governing Ultraviolet Radiation Flux at the Earth's Surface
Given its role in protecting life from harmful UV radiation, photochemical processes that
alter the concentration of stratospheric O3 are of particular concern to the global community.
Scientific understanding of the significant losses of stratospheric O3 due to chemistry with the
10-1
-------
degradation products of the long-lived anthropogenic chlorinated- and fluorinated-hydrocarbons
evolved during the 1970s and early 1980s. This realization led to the international treaty for the
protection of stratospheric O3, i.e., the 1987 Montreal Protocol on Substances that Deplete the
Ozone Layer. The scientific community continues to track stratospheric O3 levels and to
document any changes in ground-level UV-B flux due to anthropogenic depletion. An outcome
of these efforts is a limited body of literature that describes the effects of tropospheric pollutants,
particulate matter (PM), clouds and O3 on ground-level UV radiation flux.
The Montreal Protocol requires routine review of the latest scientific information available
on the status of the O3 layer and of UV radiation levels at the Earth's surface. The World
Meteorology Organization (WMO) and U.N. Environmental Program (UNEP) are responsible
for assessing the state of the science regarding the O3 layer and for reporting on trends in surface
UV radiation levels. The latest WMO/UNEP assessment was published in 2002 (WMO, 2002)
and includes an evaluation of the role of the troposphere in determining UV flux at the earth's
surface.
The mixture of gases, clouds, and particles that comprise the troposphere scatter and/or
absorb incident solar radiation (see Figure 10-1). In general, these effects are greater in the
troposphere than in the stratosphere, due to the higher atmospheric pressures, particle and cloud
densities present there. Solar flux intensity has a temporal dependence, while radiative
scattering and absorption have strong wavelength, pathlength, and/or particle concentration
dependencies. These combine to create nonlinear effects on UV flux at the earth's surface.
Thus, careful quantification of atmospheric absorbers and scatterers, along with a well-resolved
description of the physics of these interactions, is necessary for predicting the impact of ground-
level O3 on UV-B flux. The following sections (10.2.1.1-10.2.1.4) summarize the relevant
physics that govern solar flux and the nature of radiative interactions in the atmosphere. The
sources for this information are a selection of atmospheric physics and chemistry texts, the
WMO (2002) report, and other peer-reviewed literature on the role of the troposphere in defining
UV surface flux.
10.2.1.1 UV Radiation: Wavelengths and Energies
The energy possessed by a photon is inversely proportional to its wavelength. For
example, gamma rays, having wavelengths -0.1 nm, are especially damaging high-energy
10-2
-------
V
Backscattered
Radiation
Incident Solar UV Radiation
Stratospheric O3
Figure 10-1. Complexity of factors that determine human exposure to UV radiation.
In addition to the geophysical/atmospheric factors (e.g., stratospheric and
tropospheric O3, clouds, aerosols, and Rayleigh scattering) that affect the
solar flux of UV radiation at surface level, there are human physical,
behavioral and demographic factors that influence human exposure to
UV radiation.
photons (>106 kJ einsteiiT1) emitted during radioactive decay and by stellar activity.
Radiowaves, having wavelengths ~108 nm, are very low in energy (~10"3-10"8 kJ einstein"1)
and function as carriers for broadcast communications (Finlayson-Pitts and Pitts, 1986).
The wavelengths ranging between 50 and 400 nm in length are denoted "ultraviolet,"
and are subdivided into classes, where UV-C corresponds to wavelengths <280 nm, UV-B refers
to the range 280-320 nm, and UV-A to 320-400 nm. UV-C is almost entirely blocked by the
Earth's upper atmosphere, where it participates in photoionization and photodissociation
processes. Solar UV-B radiation is absorbed or scattered in part within the atmosphere. UV-A
radiation can be scattered but is not absorbed to any meaningful degree by atmospheric gases.
Both UV-B and UV-A photons contain sufficient energy to break (photolyze) chemical
bonds in biomolecules and are associated with human health and ecosystem damage. However,
10-2
-------
because UV-B is more energetic, it is capable of producing substantially more biological damage
than UV-A.
10.2.1.2 Temporal Variations in Solar Flux at the Earth's Surface
The magnitude of the solar radiation flux entering the atmosphere depends upon long-term
solar activity, sunspot cycle (11 years), solar rotation (27 days), and the position of the Earth in
its orbit around the sun. At any given point on the Earth's surface, solar flux is dependent upon
solar zenith angle, a quantity that varies with time of day, season, and latitude.
Solar Rotation and the Sunspot Cycle
A variety of changes in solar irradiance can be found in historical data, from 1700 to the
present. Solanki and Fligge (2000) concluded that solar irradiation changes, on time-scales of
days to centuries, can be attributed to variations in solar magnetic features. The maximum level
of radiation (solar-max) differs from the minimum (solar-min) by as much as 10% for
wavelengths near 160 nm. This peak-to-trough difference declines to -1% for 300 nm (Salby,
1996).
On the decadal time scale, solar rotation and sunspot activity have the largest effects on
overall solar irradiance. Since the last Maunder minimum in 1700, and consistent with
increasing sunspot activity, solar irradiance has increased by -3.0% for wavelengths in the
UV-C range and by -1.3% for wavelengths in the UV-B and UV-A ranges. Including visible
wavelengths, Solanki and Fligge (2000) estimated that the overall increase in solar irradiance
was -0.3%. Rozema et al. (2001) pointed out that any increase in wavelengths <300 nm (UV-C)
would initiate additional O3 formation in the stratosphere. This suggests that any increase in
UV-B and/or UV-A solar flux could, therefore, be offset by a more absorptive stratosphere.
The Position of the Earth with Respect to the Sun
The combined effects of the Earth's obliquity (the angle of the Earth's axis of rotation with
respect to the plane of its orbit around the sun) and its precession (the rotation of the Earth's axis
with respect to a perpendicular line through the plane of its solar orbit) yield variations of up to
30% in total summertime solar flux at the top of the earth's atmosphere, depending on latitude
(Hartmann, 1994).
10-4
-------
Zenith Angle: Latitude, Season, and Time of Day
The sun's relative elevation from a point on the earth's surface is measured with respect to
the vertical and is known as its solar zenith angle (SZA). This angle varies hourly, seasonally,
and with latitude. Diurnal and seasonal changes in SZA result in the largest changes in the
magnitude of solar radiation flux at the surface, with larger zenith angles corresponding to lower
solar flux. The largest natural fluxes occur in the tropical regions, where solar noon occurs at an
SZA at or near 0°. Seasonal variation in solar flux ranges from small changes at the equator to
very large changes at high latitudes. Diurnal variations in solar flux, from sunrise to sunset,
show added wavelength dependence as a function of SZA. This wavelength dependence is a
function of the concentrations of radiation scattering and absorbing gases and particles in the
atmospheric pathlength traversed by the photon. These processes will be discussed further
below.
10.2.1.3 Atmospheric Radiative Interactions with Solar Ultraviolet Radiation
When solar radiation enters the earth's atmosphere, it can be either scattered or absorbed
by the gases and particles it encounters. The main mechanisms of atmospheric scattering are
Rayleigh, in the case of scattering by gas molecules, and Mie, in the case of scattering by aerosol
particles. The intensity of the Rayleigh effect, i.e., the extent to which the gas molecule is
capable of perturbing the trajectory of the incident photon, depends on the size of the gas
molecule in relation to the photon, and is maximized when the photon is comparable in size or
smaller than the gas molecule. Light is scattered symmetrically with respect to the gas molecule,
with similar forward and backward intensities. This effect explains the color of the clear sky,
as blue photons are comparable in size to the dominant gases in the atmosphere (N2 and O2).
With Mie scattering, the result of light encountering larger particles, photons are scattered with a
strong forward tendency. This tendency explains why dense water clouds, i.e., cumulus clouds,
appear brilliantly white.
The lower atmospheric pressures in the stratosphere mean fewer gas molecules are present
that can absorb or scatter radiation. Stratospheric clouds and aerosols are also thinner and more
dispersed than those in the troposphere. In the language of the radiative transfer literature, these
conditions make the stratosphere a "single scattering" regime for UV radiation. The
troposphere, due to its high gas and particle concentrations is referred to as a "multiple
10-5
-------
scattering" regime. In practical terms, UV radiation traverses the stratosphere with a
substantially lower probability of encountering a gas molecule or cloud or an aerosol particle
than it would in the troposphere.
The multiple scattering of solar photons in the troposphere accounts for the
"disproportionate" role that tropospheric O3 is said to play in absorbing UV radiation versus
stratospheric O3 on a molecular per molecule basis (Briihl and Crutzen, 1989; Balis et al., 2002;
Zerefos et al., 2002). The probability that an individual photon will encounter an O3 molecule is
increased when its effective pathlength through the atmosphere is increased with multiple
scattering. This leads to an increase in the probability of absorption by O3. This effect is well
understood in principle, but its importance has not been quantified in relation to stratospheric O3
or with respect to other tropospheric absorbers.
The importance of atmospheric pressure and particle concentrations is evident in the
altitude dependence of solar flux. Solar flux has been observed to increase with altitude above
sea level, consistent with an inverse relationship between cloud levels, air pollution
concentrations, and intensity of Rayleigh scattering with altitude. A number of measurements of
UV radiation have been taken at various altitudes and are reviewed by Xenopoulos and Schindler
(2001). Increases in flux as a function of altitude are given as percent irradiance enhancement
per 1000 m relative to sea level. The effect can range from 9 to 24% /1000 m, depending upon
the altitude at which the measurement was taken (Xenopoulos and Schindler, 2001).
Similarly, this effect is seen as a function of SZA. For example, at very small zenith
angles, near solar noon, photons have a shorter atmospheric pathlength to the surface and a
smaller probability of encountering gas molecules, clouds, or particles. Conversely, photons that
enter the atmosphere at larger zenith angles, i.e., increasingly tangential to the surface, are more
likely to be scattered or absorbed.
Radiative Interactions in the Stratosphere: Absorption by Ozone
As noted earlier, the stratosphere contains 90% or more of the total column O3. Ozone
interacts with UV radiation by scattering the photon, or absorbing it. Photoabsorption by O3
occurs with very high efficiency (Note: for a discussion of the metrics used and relative
efficiencies of photoabsorption by gas-phase molecules, the reader is referred to text books such
as Finlayson-Pitts and Pitts [1986]). After electronically-excited O3 (O3*) is formed, it can
10-6
-------
dissociate into ground-state oxygen, O2, and an electronically excited oxygen radical, O(1D)
(see Reaction 10-1). At low pressures and with irradiation at 300 nm (UV-B range), O(1D) is
produced with a quantum efficiency near unity.
2 (10-1)
->O3*+M-»O3 + M* (10-2)
However, in the presence of other atmospheric gases, intermolecular collisions may also
disperse the excess electronic energy of the O3* molecule by transferring it to other molecules as
vibrational, rotational, and translational energies, warming the atmosphere, and reducing the
production of O(JD) (see Reaction 10-2). The efficiency of the energy transfer process is a
function of atmospheric pressure and the quantum mechanical properties of the colliding
molecules. A deeper discussion of molecular energy transfer processes can be found in the text
by Levine and Bernstein (1987).
The Dobson Unit (DU) is conventionally used for discussing stratospheric O3
concentrations. One DU = 2.687 x 1016 molecules of O3/cm2. Alternatively, a DU corresponds
to the column height in hundredths of a millimeter of O3 at standard temperature and pressure
(273 K and 1 atmosphere), integrated along the total height of the atmospheric column
(Wayne, 2000).
As previously noted, the total O3 column effectively prevents any UV-C from reaching the
surface and reduces the penetration of UV-B to the surface, but it does little to attenuate the
intensity of UV-A except at the shorter wavelengths close to the cutoff for UV-B. Figure 10-2
compares the solar flux above the atmosphere with ground-level flux. Cutchis (1974) calculated
that with overhead sun, a 10% decrease in the O3 column would lead to 20, 250, and 500%
increases in flux at 305, 290, and 287 nm, respectively. These estimates have been supported by
ground observations in Toronto, ON (49° N; Kerr and McElroy, 1993). Rapid changes of this
magnitude appear to happen naturally. As seen in data collected by the Total Ozone Mapping
Satellite (TOMS) (Figure 10-3), the total O3 column undergoes wide natural variation on short
timescales (Cockell, 2001).
10-7
-------
a>
o
I
1.0000
I 0.1000
cs
E
x 0.0100
V)
0.0010
0.0001
290
Solar Flux
I...-1 \/'A :•/'::' V ' :/V
Transmittance
300 310
Wavelength (nm)
320
330
Figure 10-2. Comparison of solar flux above the atmosphere with flux at the Earth's
surface. The dotted line represents extraterrestrial solar flux measured by
the satellite UARS SOLSTICE instrument (dotted line). The dashed line
represents calculated atmospheric transmittance and the solid line is the
calculated absolute flux of UV radiation for a solar zenith angle of 50°, total
column O3 of 275 DU, and a surface reflectivity of 8%. The fine structure on
the surface flux trace results from Fraunhofer lines (absorption specific
wavelengths within the solar atmosphere).
Source: Krotkov et al. (1998).
The WMO (2002) assessment reported that global average total column O3 had declined by
3% from pre-1980 levels, due to the presence of anthropogenic O3-depleting substances in the
atmosphere. Ozone depletion has a strong latitude and seasonal dependence. The seasonality of
total O3 changes differs between the Northern and Southern Hemispheres. In the northern
midlatitudes, total column O3 declined by -4% during the winter/spring seasons and by
approximately half that amount in the summer/fall of the 1997-2001 time period, relative to pre-
1980 total column O3 levels. In southern midlatitudes, total column O3 declined -6% during all
seasons, possibly tied to the enhanced photochemical losses associated with the meteorological
dynamics peculiar to the antarctic region.
10-8
-------
Ozone Column (1990-1992)
c
o
o
a
600
500
400
300
200
100
\
\\ \\ \\
Date
Figure 10-3. Ozone column abundances from the years 1990 to 1992 for 0, 40, and 80° N
as well as 80° S. The data for 80° S are incomplete, but the graph shows the
effects of the Antarctic O3 hole on total column abundances at this latitude.
The data for the Northern Hemisphere illustrate the natural variations in the
O3 column over time. The data are taken from the TOMS (Total Ozone
Monitoring Satellite) data set (1979 to 1993).
Source: Cockell (2001).
Radiative Interactions in the Stratosphere: Scattering by Clouds and Particles
During periods of extreme cold in the stratosphere above the Earth's poles, nacreous or
polar stratospheric clouds (PSCs) form due to the condensation of sulfuric and nitric acids or
water vapor. Like tropospheric water clouds, PSCs have the capacity to scatter incident solar
radiation. Due to their infrequent appearance, location, and limited optical density, they are not,
however, an important factor in determining ground-level UV flux for human health impacts
assessment.
Stratospheric aerosols such as those produced by explosive volcanic eruption, i.e.,
Mt. Pinatubo, have the potential for playing a greater role in altering ground-level UV flux.
In the case of Mt. Pinatubo, a large quantity of sulfur dioxide was injected into the stratosphere,
which reacted to form a layer of sulfate aerosols. These aerosols increased the Earth's albedo for
10-9
-------
a period of years, reducing solar flux at the Earth's surface. Volcanic eruptions that lead to the
injection of aerosols containing light-absorbing-material, such as crustal material (dust), can also
reduce surface flux (Herman et al., 1997). However, these events are also too infrequent to be of
importance in this discussion.
Radiative Interactions in the Troposphere: Gases, Particles, Clouds, and the Earth's Surface
The components of the troposphere that are of most importance in relation to the prediction
of ground-level UV radiation flux are clouds, PM, and UV-absorbing trace gases. The
UV-shielding efficiencies of these components are highly dependent on density (concentration),
particle size, and the altitude at which they are present in the troposphere. All of these properties
are sensitive to meteorology, which introduces an element of temporal dependency. The
following section discusses some of the issues specific to each component.
Gases
UV-absorbing gases, including O3, NOx, VOCs, and SO2, are vented into the upper
troposphere, injected downwards from the stratosphere (O3) or form in situ, i.e., O3, VOC
oxidation products (see Chapter 3). Ozone concentrations decrease with increasing altitude from
the surface up to, roughly, the mid-troposphere, then increase up into the stratosphere.
Figure 10-4 shows a series of O3 vertical profiles for 4 sites within the continental U.S., i.e.,
plots of O3 concentrations as a function of atmospheric pressure (correlating to altitude). The
mean values of O3 in the free troposphere reported in the literature range from -50 to -80 ppbv,
with higher values occurring at the tropopause. For example, a series of ozonesonde soundings
over France from 1976 to 1995 showed an O3 increase from 48.9 ppbv in the 2.5 to 3.5 km layer
to 56.5 ppbv in the 6.5 to 7.5 km layer (Ancellet and Beekmann, 1997).
Within the planetary boundary layer, photochemistry produces a diurnal rise and fall in O3
and PM concentrations, especially in polluted urban settings. Temperature inversions that often
occur in these settings prevent the upward mixing and dilution of ground-level O3, also trapping
primary and secondary PM within the boundary layer. These effects are described in Chapters 2
and 3 of this document. Such conditions substantially increase the UV absorptive capacity of the
atmosphere, at the surface.
10-10
-------
Mean Monthly Ozone (mPa)
Trinidad Head, CA Boulder, CO Huntsville, AL
Wallops Island, VA
/ y y Number of profiles
1000
10
10 15 20 0 5 10 15 20 0 5 10 15 20
Figure 10-4. Monthly averaged vertical O3 profiles (partial pressure in mPa) as a function
of atmospheric pressure (in mBar) for Trinidad Head, CA (solid line);
Boulder, CO (dot-dashed line); Huntsville, AL (dotted line); and Wallops
Island, VA (dashed line). The number of launches at each site for each
month are indicated on the charts.
Source: Newchurch et al. (2003).
10-11
-------
Ultraviolet absorption by gases becomes significant under aerosol- and cloud-free
conditions. Figure 10-5 shows a calculation by Krotkov et al. (1998) of the sensitivity, as a
function of wavelength, of ground-level UV flux to a 1-DU decrease in total column O3 under
cloud- and aerosol-free conditions. (Note: 1 DU = 2.687 x 1016 molecules of O3 per cm2 or if
calculated for the entire troposphere, 1 DU = 10.9 Tg(O3) with 1 ppb of tropospheric O3 =
0.65 DU).
A study in Chicago in 1991 and 1992, during which ambient O3, broadband UV irradiance,
and total sunlight were monitored (Frederick et al., 1993), found a significant negative
correlation between the UV irradiance and ambient O3 when the atmosphere was relatively free
of clouds and haze. Although Frederick et al. (1993) estimated that a 10 ppbv reduction in O3
was associated with a 1.3% increase in erythemally-weighted UV-B, they cautioned that this
value had a comparatively large uncertainty (-1.2%, or nearly 100% of the estimated increase).
Matthijsen et al. (2000) noted that scattering effects differ for UV versus total solar irradiance:
UV radiation, even under clear sky conditions is mostly diffuse, whereas total solar is largely
direct. This is consistent with Rayleigh scattering of UV wavelengths by N2 and O2.
Clouds
Clouds have the largest observed influence on solar irradiance-scattering as much as 100%
of incoming direct beam radiation versus the 10-20% scattering that occurs on clear days due to
Rayleigh and PM. The long term average for the U.S., based on measurements taken at various
locations, show monthly average reductions in UV levels ranging from 10% to 50%, due to
cloud cover (Madronich et al., 1995). An overall reduction in normalized global irradiance has
been attributed, in part, to the increase in the extent (8%) and optical density of cirrus clouds
observed since the middle of the last century (Trepte and Winkler, 2004).
The depth, particle size distribution, and coverage determine, in large part, the amount and
wavelengths of radiation that clouds will scatter. In general, cloud effects are weaker for the UV
wavelengths than for total solar radiation. Alados-Arboledas et al. (2003) observed that the
cloud effect under overcast conditions for total solar radiation is 33% greater than for
erythemally-weighted UV radiation.
Geometry is also an especially important factor: scattered or broken clouds can either
slightly reduce or enhance irradiance due to scattering between clouds (WMO, 2002).
10-12
-------
290
300 310 320
Wavelength (nm)
330
Figure 10-5. The sensitivity of ground-level UV flux to a 1 DU change in total column O3,
under clear sky conditions, as a function of solar zenith angle (SZA).
Source: Krotkovetal. (1998).
Palancar and Toselli (2004) estimated that broken clouds can increase noontime radiation by up
to 34% when compared to calculated clear sky values. Mims and Fredrick (1994) reported
observations of UV-B radiation enhancements of up to 20% over maximum noontime values due
to scattering from cumulus clouds. Overcast conditions have been estimated, for SZAs
averaging 50°, to reduce UV-B surface radiation to 29% of clear sky irradiances (Schafer et al.,
1996). Schafer et al. (1996) also observed, as a percentage of clear sky values, 61% for 80-90%
coverage, 75% for 60-70% coverage, 79% for 40-50% coverage, with frequent incidences of
enhancement above clear sky levels due to reflections during partial cloud cover.
The altitude at which clouds are located also substantially alters their effectiveness in
scattering radiation back into space. High clouds have been observed to be more effective, by a
factor of 5, at scattering 300 nm wavelength radiation (SZA = 0°) than low clouds (Wen and
Frederick, 1995). At SZA = 60°, 305 nm wavelength radiation scatters from high clouds
10-13
-------
(8-10 km) at twice the efficiency of low clouds (1-2 km). The authors attribute this effect to
the reduction in the atmospheric pathlength that UV photons must travel, and therefore a
reduction in the odds of scattering due to other particle or gas interactions, before encountering
the cloud.
Quantifying the effect of clouds on surface UV flux, therefore, requires detailed
information on cloud depth, particle size distribution, geometry, altitude, and the position of the
sun relative to the cloud as a function of time.
Particulate Matter
On a zonally averaged basis, PM does not contribute significantly to lower tropospheric
absorption of UV radiation. However, in urban areas or other areas subject to high smog levels,
such as areas adjacent to significant biomass combustion, PM may be the most important
determinant of ground-level erythemal UV flux, excepting clouds (U.S. Environmental
Protection Agency, 2004; WMO, 2002). Model-to-measurement comparisons of ground-level
flux for Cordoba City (Argentina), the Aegean Sea (Greece) and Toronto (Canada), have shown
20%, 20% and 5-10% reductions, respectively (McKenzie et al., 2003). Liu et al. (1991)
estimated similar potential reductions (20%) in UV-B radiation in the presence of large local PM
concentrations.
Barnard et al. (2003) measured UV-B penetration to the surface at 3 sites: Rubidoux
(California); the University of California, Riverside; and Mt. Gibbes in the North Carolina Blue
Ridge Mountains. They found that UV-B is primarily sensitive, under cloud-free conditions,
to O3 values measured by satellite, which do not account for ground level O3, and concentrations
of PM and specifically PM containing black carbon. They found that ground level O3
concentrations did not meet the 0.05 significance level in their model for urban environments.
Barnard et al. (2003) further suggested that increases over the past 20 to 30 years in combustion-
associated PM and black carbon may account for the inability to detect a surface trend in UV-B
radiation caused by a known decrease in stratospheric O3 over the Northern Hemisphere.
Surface Albedo
The Earth's surface, by absorbing or reflecting radiation, directly influences UV flux in the
lower troposphere. The extent to which the surface reflects incident radiation back into the
10-14
-------
atmosphere is known as its "albedo." Ultraviolet flux is directly proportional to surface albedo
(Wendisch and Mayer, 2003).
Surface albedo is a very strongly wavelength-dependent property. For example, fresh and
wet snow reflect 60 to 90% of incident violet light, while soil and grass surfaces reflect <5%
(Xenopoulos and Schindler, 2001). In their in situ measurement and modeling study of the
vertical distribution of solar irradiance, Wendisch and Mayer (2003) found that surface albedos
must be specifically measured in order to accurately simulate solar flux, due to the large
variations that may occur within a given surface type.
Snow cover, even many kilometers from measurement sites is known to increase detected
UV irradiances. Complicated interactions result when radiation is scattered by snow (or other
bright surfaces) and backscattered or absorbed by atmospheric particles and clouds in the same
vicinity (WMO, 2002).
10.2.1.4 Data Requirements for a Surface UV-B Climatology
A means of establishing the range of variability in UV-B at ground level for a populated
area of interest would be to the development of a map of flux levels under typical seasonal
conditions based on historical records. In the atmospheric sciences community, a map of this
type is referred to as a "climatology."
The WMO (2002) stated that, in principle, if the spatial distribution of all UV absorbers
and scatterers were fully known, the wavelength and angular distribution of the UV irradiance at
the Earth's surface could be determined with model calculations. However, the very limited
information available on the distribution of the primary components (i.e., clouds, particles, O3,
other UV-absorbing trace gases and surface albedo) makes detailed predictions impossible.
Nearly all of the routine data on tropospheric O3 concentrations in the U.S. is from ground-level
O3 monitors, such as those used to determine the attainment of the O3 air quality standards. Such
measurements, alone, are not sufficient information for making reliable estimates of ambient O3
concentrations above the boundary layer—information needed to establish the vertical
distribution of O3 up to the tropopause. In an earlier assessment of the environmental effects of
stratospheric O3 depletion (WMO, 1999), the UNEP concluded that, in view of the high spatial
and temporal variability of surface UV radiation and the difficulty in maintaining calibration
within networks of UV monitoring instruments, satellite-based observations are necessary to
10-15
-------
develop a satisfactory UV climatology. However, satellite-derived estimates of surface UV are
limited by the availability of instruments in orbit, with available datasets comprising
interpolations based upon a single satellite overpass per day for a given region.
Furthermore, the current limitation in all satellite retrievals is that backscattered radiation,
from which flux rates are derived, are limited to atmospheric layers above the surface.
Assumptions must be made concerning aerosol and ozone absorption below the atmospheric
limit of detection by satellites (McKenzie, et al., 2001). Above densely populated or industrial
areas, aerosol extinctions are large and reduce UV appreciably (Gonzalez et al., 2000).
Complete assessment of the uncertainties in predictions of UV surface flux would require
comparisons between sparse available ground-level observations and satellite data over longer
periods of time and for different geographical locations. No such assessment has been reported
in the scientific literature.
10.2.2 Factors Governing Human Exposure to Ultraviolet Radiation
An assessment of public health benefits due to the attenuation of UV-B radiation by
surface-level O3 requires appropriate consideration of: (1) the multiple factors that alter the flux
of UV-B radiation at ground-level, as described above; (2) the factors that influence the extent of
human exposure to UV-B radiation, particularly behavioral decisions; and (3) the effects of
UV-B radiation exposure on human health. Consideration must also be given to the public
health benefits from exposure to UV-B radiation. The present section outlines the most recent
information on the determinants of exposure to UV-B radiation in human populations.
Quantitative evaluation of human exposure to UV-B radiation is scientifically necessary to
perform health risk assessment and to define subpopulations at risk for UV-B-related health
effects.
10.2.2.1 Outdoor Activities
Exposure to solar UV radiation is related to one predominating factor: time spent outdoors
during daylight hours. A large U.S. study was conducted using the EPA National Human
Activity Pattern Survey (NHAPS) to assess UV radiation dose in Americans (Godar, 2001;
Godar et al., 2001, 2003). The EPA NHAPS recorded the activity profiles of 9,386 Americans
(age 0 to 60+ years) over a 24-month period to assess their exposure to various environmental
10-16
-------
pollutants, including UV radiation. Available UV radiation was assessed using the EPA
UV-monitoring program. Solar radiation in the UV-A and UV-B waveband regions were
measured daily at a monitoring site in each quadrant of the U.S. There is considerable error
associated with quantifying UV radiation dose from exposure surveys and four UV-monitoring
sites across the country; however, the qualitative information regarding factors that increase
human exposure to UV radiation is still of relevance. The EPA-UV monitoring network has
since expanded to 21 sites, located in 14 U.S. national parks and 7 urban areas across the
United States (http://www.epa.gov/uvnet/). A UV-B monitoring network by the U.S.
Department of Agriculture is also available for the quantitative assessment of UV radiation
exposure (http://uvb.nrel.colostate.edu/UVB/). This monitoring network has 30+ monitoring
sites across the United States and three additional sites in Canada and New Zealand.
Godar et al. (2001) observed a strong seasonal preference for outdoor activities, with
people spending the most time outdoors during the summer followed by spring, fall, and, lastly,
winter. Because the solar erythemal (i.e., skin reddening) UV radiation dose is also highest
during the summer, the estimated UV radiation dose of Americans was more than 10-fold greater
in the summer compared to the winter season (Godar et al., 2001).
Vacationing at the beach in the summer was associated with higher UV radiation exposures
(Godar et al., 2001; Thieden et al., 2001). Even after accounting for sunscreen use at the beach,
the erythemal UV radiation doses were more than 40% higher during a 3-week beach vacation
compared to a 3-week stay at home (Godar et al., 2001). Danish children and adolescents were
found to receive >50% of their annual UV radiation dose while vacationing at European beaches
(Thieden et al., 2004a). Sunbathing also was associated with increased annual UV radiation
dose in the Canadian National Survey on Sun Exposure and Protective Behaviours (Shoveller
et al., 1998). Among the 3,449 adults (age 25+ years) who completed the telephone household
survey, 21% stated that they spent time actively sunbathing. In a Danish study with 164
participants, all children (age 1 to 12 years) and teenagers (age 13 to 19 years) as well as 94% of
adults (age 20 to 76 years) had days with risk behavior (Thieden et al., 2004b). Teenagers, who
had the highest number of risk-behaviors days, were found to have the highest annual UV
radiation doses. Among teenagers, 76% (95% CI: 41, 98) of their UV radiation dose during the
measurement period was received on risk-behavior days, as determined using personal electronic
UV dosimeters and exposure diaries (Thieden et al., 2004b).
10-17
-------
An Australian study examining time profiles of daily UV radiation exposure among
8th grade students observed that up to 47% of the daily UV radiation dose fell within the time
periods when students were outdoors during school hours, sitting under shaded structures during
lunch breaks and participating in routine outdoors or sports activities (Moise et al., 1999).
Other studies also have found that participation in outdoor sports (e.g., basketball, soccer,
golfing, swimming, cycling) significantly increased UV radiation exposure (Moehrle, 2001;
Moehrle et al., 2000; Thieden et al., 2004a,b).
10.2.2.2 Occupation
Of the various factors that affect human exposure to UV radiation, occupation is also
important. Approximately 5% of the American workforce work outdoors, as determined by the
EPA NHAPS (Godar et al., 2001). On average, American indoor workers spend -10% of their
day outdoors. During their time outdoors, they are exposed to -30% of the total ground-level
UV flux, as measured by the EPA UV-monitoring program (Godar et al., 2001). Compared to
indoor or in-home workers, outdoor workers are exposed to much higher levels of UV radiation
(Kimlin et al., 1998a; Thieden et al., 2004a), frequently at levels that are above current exposure
limits set by the International Commission on Non-Ionizing Radiation Protection (ICNIRP,
2004). For example, Thieden et al. (2004a) observed that the annual UV radiation dose,
estimated using personal electronic UV dosimeters and exposure diaries, was -70% higher for
gardeners than indoor workers. The gardeners received the majority (55%) of their UV radiation
dose on working days (Thieden et al., 2004a). Another study found that outdoor workers
received three to four times the annual UV radiation exposure of indoor workers (Diffey, 1990).
At-risk working populations include farmers (Airey et al., 1997; Schenker et al., 2002),
fishermen (Rosenthal et al., 1988), landscapers (Rosenthal et al., 1988), building and
construction workers (Gies and Wright, 2003), physical education teachers (Vishvakarman et al.,
2001), mail delivery personnel (Vishvakarman et al., 2001), and various other workers who
spend the majority of their day outdoors during peak UV radiation hours.
10.2.2.3 Age
Age may be a factor that influences human exposure to UV radiation. In a large U.S. study
using the EPA NHAPS, the average UV radiation dose among American children (age
10-18
-------
<12 years) was estimated to be slightly higher (-20%) than that of adolescents (age 13 to
19 years) (Godar, 2001). A large Canadian survey found that 89% of children (age <12 years)
had 30 minutes or more of daily UV exposure compared to 51% for both adults (age 25+ years)
and youth (age 15 to 24 years) (Lovato et al., 1998a, 1998b; Shoveller et al., 1998). In an
English study (Diffey et al., 1996), UV radiation exposure was estimated in 180 children (age 9
to 10 years) and adolescents (age 14 to 15 years) using personal film badges and exposure
records. Once again, children were found to have received higher UV radiation exposure
compared to adolescents (Diffey et al., 1996). However, as discussed earlier, a Danish study
found that the annual UV radiation dose in teenagers (age 13 to 19 years) was 14-24% higher
compared to children (age 1 to 12 years) and adults (Thieden et al., 2004b). This increase in UV
radiation dose in the Danish teenagers was attributed to their increased risk-behavior days.
Therefore, age may affect human exposure to UV radiation by influencing other factors of
exposure, such as outdoor activity and risk behavior.
Two studies examined lifetime UV radiation exposure among persons in the U.S. (Godar
et al., 2003) and Denmark (Thieden et al., 2004b). Both studies observed that while there are
slight differences in UV radiation dose by age, generally people receive fairly consistent UV
doses at different age intervals throughout their lives.
10.2.2.4 Gender
Studies have indicated that females generally spend less time outdoors and, consequently,
have lower UV radiation exposure compared to males (Gies et al., 1998; Godar et al., 2001;
Shoveller et al., 1998). The U.S. study by Godar et al. (2001) observed that while both males
and females had relatively consistent erythemal UV radiation doses throughout their lives, males
consistently received higher overall UV doses compared to females at all age groups. Among all
Americans, the lowest exposure to UV radiation was received in females during their child-
raising years (age 22 to 40 years) (Godar et al., 2001). The highest exposure was observed in
males aged 41 to 59 years in the U.S. study (Godar et al., 2001). A similar Canadian survey
found that younger adult males had the greatest exposures to UV radiation (Shoveller et al.,
1998).
10-19
-------
10.2.2.5 Geography
In the U.S. study by Godar et al. (2001), erythemal UV radiation doses were examined in
persons living in northern and southern regions. Northerners and southerners were found to
spend an equal amount of time outdoors; however, the higher solar flux at lower latitudes
significantly increased the annual UV radiation dose for southerners (Godar et al., 2001). The
annual UV radiation doses in southerners were 25 and 40% higher in females and males,
respectively, compared to northerners (Godar et al., 2001). Other studies also have shown that
altitude and latitude influence personal exposure to UV radiation (Kimlin et al., 1998b; Rigel
etal., 1999).
10.2.2.6 Protective Behavior
Protective behaviors such as using sunscreen (e.g., Nole and Johnson, 2004), wearing
protective clothing (e.g., Rosenthal et al., 1988; Sarkar, 2004; Wong et al., 1996), and spending
time in shaded areas (Moise et al., 1999; Parisi et al., 1999) have been shown to reduce exposure
to UV radiation. In one study, the use of sunscreen was associated with extended intentional UV
radiation exposure (Autier et al., 1999); however, a follow-up study indicated that sunscreen use
increased duration of exposures to doses of UV radiation that were below the threshold level for
erythema (Autier et al., 2000).
In a national study of U.S. youths aged 11 to 18 years, the most prevalent protective
behavior was sunscreen use (39.2%) followed by use of a baseball hat (4.5%) (Davis et al.,
2002). There were significant differences in the use of sunscreen by age group and gender, with
the younger age group (age 11 to 13 years) and girls having greater likelihood (47.4 and 48.4%,
respectively) of using sunscreen (Davis et al., 2002). The Canadian National Survey on Sun
Exposure and Protective Behaviours observed that less than half of the adults (age 25+ years,
n = 3,449) surveyed took adequate protective actions (Shoveller et al., 1998). Once again,
children (age <12 years, n = 1,051) were most protected from exposure to UV radiation, with
76% using sunscreen and 36% avoiding the sun, as reported by their parents (Lovato et al.,
1998a). However, the protection level was still not adequate, as indicated by the high 45% rate
of erythema in children. Among Canadian youth (age 15 to 24 years, n = 574), protective
actions from UV radiation exposure included wearing a hat (38%) and seeking shade and
avoiding the sun between the peak hours of 11:00 a.m. to 4:00 p.m. (26%) (Lovato et al., 1998b).
10-20
-------
The lowest prevalence of protective behavior among the youth was likely responsible for the
highest proportion of erythema (68%) experienced in this age group. A Danish study observed
that both children and teenagers applied sunscreen on more days than adults, but teenagers had
the most days with erythema, due to their increased risk behavior (Thieden et al., 2004b).
A survey in Switzerland of 1,285 individuals, including children and parents, indicated that
sunscreen use was the protective action most commonly used, but only at the beach and not in
routine daily exposure (Berret et al., 2002). In general, protective clothing and avoiding the sun
were not highly used among these individuals to protect against UV-related health effects.
10.2.2.7 Summary of Factors that Affect Human Exposures to Ultraviolet Radiation
The factors that potentially influence UV radiation doses were discussed in the previous
sections and include outdoor activities, occupation, age, gender, geography, and protective
behavior. Results from the various studies indicate that the following subpopulations may be at
risk for higher exposures to UV radiation:
• Individuals who engage in high-risk behavior, viz., sunbathing;
• Individuals who participate in outdoor sports and activities;
• Individuals who work outdoors with inadequate shade, e.g., farmers, fishermen,
gardeners, landscapers, building and construction workers; and
• Individuals living in geographic areas with higher solar flux (i.e., lower latitudes
[e.g., Honolulu, HI] and higher altitudes [e.g., Denver, CO]).
10.2.3 Factors Governing Human Health Effects due to Ultraviolet Radiation
Ultraviolet radiation occupies a specific region of the electromagnetic spectrum of
wavelengths and can be further subdivided into three parts, UV-A (320 to 400 nm), UV-B
(280 to 320 nm), and UV-C (200 to 280 nm). Most of the health risks associated with UV
radiation exposure are wavelength dependent. Wavelengths <180 nm are of little practical
biological significance as they are almost completely absorbed by the stratosphere (ICNIRP,
2004).
"Action spectra" of a given biological response to UV radiation across its spectral range
are used to estimate exposure by weighting individual wavelength intensities according to the
10-21
-------
associated response. The overall effectiveness of the incident flux at inducing the biological
response of interest is computed by means of the relationship:
effective irradiance — ]J^E^ d'k Cl 0-3")
A
where 1A and EA are, respectively, the irradiance and its relative effectiveness at wavelength A.
Until 1980, it was generally thought that wavelengths <3 15 nm were responsible for the
most significant adverse UV radiation health effects; however, recent studies have found that the
longer wavelengths in the UV-A range also may produce adverse responses at substantially
higher doses (ICNIRP, 2004). As UV-A radiation is not absorbed by O3, health effects solely
induced by UV-A exposure are not relevant in an analysis of public health risks/benefits
associated with O3-related UV attenuation. Therefore, this section focuses on the latest available
information on the various adverse health effects associated with acute and chronic UV-B
radiation exposure.
10.2.3.1 Erythema
Association Between Ultraviolet Radiation Exposure and Erythema
The most conspicuous and well-recognized acute response to UV radiation is erythema, or
the reddening of the skin, which is likely caused by direct damage to DNA by UV-B and UV-A
radiation (Matsumura and Ananthaswamy, 2004). Indirect oxidative damage also may occur at
longer wavelengths (Matsumura and Ananthaswamy, 2004). Skin type appears to play a large
role in the sensitivity to UV radiation-induced erythema. The Fitzpatrick classifications for skin
types are: (1) skin type I - individuals with extremely sensitive skin that sunburns easily and
severely, and is not likely to tan (e.g., very fair skin, blue eyes, freckles); (2) skin type
II - individuals with very sensitive skin that usually sunburns easily and severely, and tans
minimally (e.g., fair skin, red or blond hair, blue, hazel or brown eyes); (3) skin type
III - individuals with sensitive skin that sunburns moderately and tans slowly (e.g., white skin,
dark hair); (4) skin type IV - individuals with moderately sensitive skin that sunburns minimally
and usually tans well (e.g., white or light brown skin, dark hair, dark eyes); (5) skin type
V - individuals with minimally sensitive skin that rarely sunburns and tans deeply (e.g., brown
skin); and (6) skin type VI - individuals with nonsensitive skin that never sunburns and tans
10-22
-------
profusely (e.g., dark skin). Harrison and Young (2002) found that the perceptible minimal
erythemal dose was approximately twofold greater for individuals with skin type IV compared to
skin type I, although there was considerable overlap in the minimal erythemal dose among the
four skin types. Waterston et al. (2004) further observed that within an individual, erythemal
response differed by body site (e.g., abdomen, chest, front upper arm, back of thigh). These
differences were likely attributable to body site-specific variations in melanin pigmentation.
Kollias et al. (2001) investigated the change in erythemal response following a previous
exposure to UV radiation. Body sites that received a second exposure to UV radiation always
showed a reduced erythemal response compared to body sites with a single exposure, especially
when the first exposure was at levels greater than the minimal erythemal dose. The suppression
of erythema was more pronounced when the second exposure was given 48 hours after the first.
These findings support the well established notion that repeated exposures to UV radiation
results in adaptation (e.g., stimulation of melanogenesis). Kaidbey and Kligman (1981)
examined individuals with skin types I, II, and III, and found that multiple exposures to
subthreshold doses of UV radiation at 24-hour intervals resulted in cumulative injury to the skin,
as indicated by a lowering of the minimal erythemal dose. These results suggest that a longer
time period than 24 hours may be necessary to repair damage from a single exposure to UV
radiation. Henriksen et al. (2004) also observed a lowering of the minimal erythemal dose with
repeated exposure at 24-hour intervals in 49 healthy volunteers with skin types II, III, IV, and V.
However, adaptation was reached after the 4th consecutive exposure. Henriksen et al. further
found that the change in threshold depended on skin type. After 4 days of repeated UV
radiation, there was little change (10 to 20%) in the erythemal threshold dose with repeated
exposure to UV radiation in the fair-skinned individuals. Among the darker-skinned individuals,
the minimal erythemal dose was lowered by 40 to 50%. However, both the initial UV dose and
the dose to erythema after four days of exposure was still higher in the dark-skinned persons.
A reference erythema action spectrum was adopted by the Commission Internationale de
1'Eclairage (International Commission on Illumination, CIE) in 1987 (McKinlay and Diffey,
1987). The CIE erythema action spectrum indicates that UV-B radiation is orders of magnitude
more effective per unit dose than UV-A radiation.
10-23
-------
Risk of Erythema from Changes in Tropospheric O3 Levels
There is no literature examining the risk of erythema associated with changes specifically
in tropospheric or ground-level O3 levels. The scientific studies, available to date, focus on the
effects of a reduction in stratospheric ozone. One such study has assessed the effects of
stratospheric O3 depletion on the risk of erythema (Longstreth et al., 1998). The analysis by
Longstreth et al. (1998) concluded that the risk of erythema would not appreciably increase with
depletion of the stratospheric O3 layer. This is due to the powerful adaptation of the skin to
different levels of UV radiation, as evidenced by its ability to cope with changes in UV radiation
by season (Van Der Leun and De Gruijl, 1993). Gradual exposure to increasing UV radiation
from the winter to summer leads to decreased sensitivity of the skin. In midlatitudes, the UV-B
radiation in the summer is 10-fold greater than in the winter. In contrast, the steady depletion of
the O3 layer has been estimated to result in an approximately 20% increase in UV-B over 10
years (Longstreth et al., 1998). The comparatively small increase in UV radiation throughout the
years, therefore, would not significantly increase the risk of erythema. Tropospheric O3
constitutes no more than 10% of total atmospheric O3. Given that stratospheric O3 depletion was
unlikely to increase the risk of erythema, one could reasonably conclude that small changes in
ground-level O3 that take place with attainment of the O3 NAAQS would also not result in
increased risk.
10.2.3.2 Skin Cancer
According to the American Academy of Dermatology, one in five Americans develop skin
cancer during their lifetime. The three main forms of skin cancer include basal cell carcinoma
and squamous cell carcinoma, which are both nonmelanoma skin cancers, and malignant
melanoma. Nonmelanoma skin cancers constitute more than one-third of all cancers in the U.S.
and -90% of all skin cancers, with basal cell carcinoma being approximately four times as
common as squamous cell carcinoma (Diepgen and Mahler, 2002; ICNIRP, 2004). The
incidence of malignant melanoma is much lower than nonmelanoma skin cancers. In 2004,
more than one million cases of basal and squamous cell skin cancer are expected to be newly
diagnosed, compared to 40,780 cases of melanoma (Jemal et al., 2004). However, melanoma
has great metastatic potential and accounts for the majority of skin cancer deaths.
10-24
-------
Exposure to UV radiation is considered to be a major risk factor for all three forms of skin
cancer (Gloster and Brodland,1996; Diepgen and Mahler, 2002; IARC, 1992). Ultraviolet
radiation is especially effective in inducing genetic mutations and acts as both a tumor initiator
and promoter. Keratinocytes have evolved DNA repair mechanisms to correct the damage
induced by UV; however, mutations can occur, leading to skin cancers that are appearing with
increasing frequency (Hildesheim and Fornace, 2004). The relationship between skin cancer and
chronic exposure to UV radiation is further explored below, followed by discussion of the
influence of O3 on the incidence of skin cancer.
10.2.3.3 Ultraviolet Radiation Exposure and the Incidence of Nonmelanoma Skin Cancers
The incidence of all three types of cancers has been shown to rise with increasing UV
radiation concentrations across the U.S. (De Gruijl, 1999); however, the most convincing
evidence for a causal relationship exists between UV radiation and squamous cell carcinoma.
Squamous cell carcinoma occurs almost exclusively on skin that is regularly exposed to the sun,
such as on the face, neck, arms, and hands. The incidence is higher among whites in areas of
lower latitudes, where solar flux is greater (Kricker et al., 1994). The risk of squamous cell
carcinoma was shown to increase with life-long accumulated exposure to UV radiation in one
cross-sectional study (Vitasa et al., 1990); however, increased risk was found to be associated
only with exposure 10 years prior to diagnosis in a case-control study (Gallagher et al., 1995a).
One of the major concerns with both types of studies is the potential for recall bias in reporting
past UV radiation exposure by individuals already aware of their disease status.
Ultraviolet radiation also has been linked to basal cell carcinoma. Basal cell carcinoma is
common on the face and neck (80-90%) but rarely occurs on the back of the hands (De Gruijl,
1999). While cumulative UV radiation exposure was not associated with an increased risk of
basal cell carcinoma (Vitasa et al., 1990), increased risk was observed in individuals with greater
recreational UV radiation exposure in adolescence and childhood (age <19 years) and
individuals with a history of severe erythema in childhood (Gallagher et al., 1995b). Once again,
consideration must be given to potential recall bias in assessing these results. Thus, there is
suggestive evidence that UV radiation also plays a role in the development of basal cell
carcinoma, but the etiologic mechanisms for squamous cell carcinoma and basal cell carcinoma
likely differ. In an Australian study conducted in a subtropical community, the factors of having
10-25
-------
fair skin, a history of repeated sunburns, and nonmalignant solar skin damage diagnosed by
dermatologists were strongly associated with both types of nonmelanoma skin cancer (Green
et al., 1996). The authors attributed the finding that outdoor occupation was not associated with
nonmelanoma skin cancer to self-selection. Individuals with fair or medium complexions and a
tendency to sunburn accounted for more than 80% of the community study sample; however,
they were systematically underrepresented among outdoor workers (Green et al., 1996). Such
self-selection bias might partly explain the lack of consistent quantitative evidence of a causal
link between UV radiation and skin cancer in humans.
De Gruijl et al. (1993) assessed the action spectrum for nonmelanoma skin cancers using
hairless albino mice. Human data are not available regarding wavelength dependence of the
carcinogenicity of UV radiation. After adjusting for species differences, the Skin Cancer
Utrecht-Philadelphia action spectrum indicated the highest effectiveness in the UV-B range with
a maximum at 293 nm, which dropped to 1CT4 of this maximum at the UV-A range above
340 nm (De Gruijl et al., 1993). The mutations commonly present in thep53 tumor suppressor
gene in individuals with squamous cell carcinoma and basal cell carcinoma are called the
"signature" mutations of UV-B radiation (De Gruijl, 2002). UV-B radiation is highly mutagenic,
because DNA is a chromophore for UV-B, but not for UV-A radiation (Ichihashi et al., 2003).
Nevertheless, other studies have found that UV-A radiation, in addition to UV-B radiation, can
induce DNA damage (Persson et al, 2002; Riinger et al., 2000). DNA damage by UV-A is
mediated by reactive oxygen species, making it indistinguishable from damage caused by other
agents that generate reactive oxygen species (De Gruijl, 2002). Epidemiologic evidence of a
carcinogenic effect of UV-A was found in a study of psoriasis patients receiving oral psoralen
and UV-A radiation treatment (Stern et al., 1998). High-dose exposure to oral psoralen and
UV-A radiation was associated with a persistent, dose-related increase in the risk of squamous
cell cancer. Risk of basal cell cancer also was increased in those patients exposed to very high
levels of UV-A radiation. Therefore, although UV-B radiation has long been considered the
main culprit for nonmelanoma skin cancer, limited evidence suggest that UV-A radiation may
also play a role.
Susceptible populations for nonmelanoma skin cancers include individuals with reduced
capacity for nucleotide excision repair, the primary repair mechanism for UV radiation-induced
DNA lesions (Ichihashi et al., 2003). At particular risk are individuals with xeroderma
10-26
-------
pigmentosum, as they have defective nucleotide excision repair in all tissues (Kraemer, 1997;
Sarasin, 1999). Skin type also largely affects susceptibility to skin cancer. Of the six skin
phenotypes, the most sensitive individuals are those with skin types I and II, who have a fair
complexion, blue or green eyes, and red or blond hair (Diepgen and Mahler, 2002). These
individuals tend to sunburn easily, tan poorly, and freckle with sun exposure. A history of
repeated sunburns also appears to increase the risk of both cancers, while sunburns during
childhood are more associated with increased basal cell carcinoma (Gallagher et al., 1995b;
Green etal., 1996).
Ultraviolet Radiation and the Incidence of Cutaneous Malignant Melanoma
From 1973 to 1994, the incidence rate of melanoma increased 120.5% along with an
increased mortality rate of 38.9% among whites in the United States (Hall et al., 1999). The
ICNIRP (2004) states that during the past 40 years or so, each decade has seen a twofold
increase in the incidence of malignant melanoma in white populations, with increased incidence
observed more prominently in individuals living in lower latitudes. Cutaneous malignant
melanoma has a mutifactorial etiology with environmental, genetic, and host factors (Lens and
Dawes, 2004). The major environmental factor of malignant melanoma has been identified as
UV radiation exposure (Diepgen and Mahler, 2002); therefore, the increased incidence of
melanoma throughout the years might be partially attributable to changes in human activity
patterns (e.g., increased outdoor activity) that influence UV exposure or increased UV radiation
at the ground level. The risk of melanoma appears to depend on the interaction between the
nature of the exposure and skin type (Lens and Dawes, 2004).
Fears et al. (2002) examined the association between invasive cutaneous melanoma and
UV radiation in non-Hispanic whites using a case-control study design. Lifetime residential
history was coupled with mid-range UV-B radiation flux measurements to reduce exposure
misclassification and recall bias. A 10% increase in the average annual UV-B flux was
significantly associated with a 19% (95% CI: 5, 35) increase in individual odds for melanoma in
men and a 16% (95% CI: 2, 32) increase in women. Whiteman et al. (2001) conducted a
systematic review of studies that examined the association between childhood UV radiation
exposure and risk of melanoma. Researchers found that ecological studies assessing ambient
sun exposure consistently reported higher risks of melanoma among people who resided in an
10-27
-------
environment with high UV radiation during their childhood (Whiteman et al., 2001). The lack of
consistency among the case-control studies was likely due to the varying methods used to assess
UV radiation dose.
While the evidence is generally suggestive of a causal relationship between UV radiation
and malignant melanoma, possibly conflicting data also has been observed. For example, the
highest occurrence of malignant melanoma is on men's backs and women's legs, areas that do
not have prolonged exposure to the sun (Rivers, 2004). This indicates that, unlike nonmelamona
skin cancers, malignant melanoma tends to occur in sites of intermittent, intense sun exposure
(trunk and legs), rather than in areas of cumulative sun damage (head, neck, and arms) (Swetter,
2003). A study by Whiteman et al. (2003) observed that individuals with melanomas of the
trunk had more melanocyte nevi and less solar keratoses compared to individuals with head and
neck melanomas, suggesting that cutaneous melanomas may arise through two pathways, one
associated with melanocyte proliferation and the other with chronic exposure to sunlight. Green
et al. (1999) also found that melanomas of the soles and palms resembled other cutaneous
melanomas in their association with sun exposure, but were distinguished from them by their
strong positive associations with nevi on the soles.
The available data conflict with regard to the relative importance of UV-A versus UV-B
in inducing melanomas. UV-A has a much higher flux rate at the Earth's surface, as it is not
absorbed by O3; and it is able to penetrate more deeply into the skin surface due to its longer
wavelength. However, UV-B, as mentioned earlier, is much more energetic and, therefore, more
effective in photochemically altering DNA. The individual roles of UV-A and UV-B in the
development of cutaneous malignant melanoma have been examined in several studies.
A case-control study of 571 patients and 913 matched controls found an elevated odds ratio of
1.8 (95% CI: 1.2, 2.7), after adjusting for skin type, hair color, raised nevi, and number of
sunburns, for developing malignant melanoma in individuals who regularly used tanning beds,
which typically are UV-A sources (Westerdahl et al., 2000). In a study by Setlow et al. (1993),
an action spectrum using the tropical fish Xiphophorus indicated that UV-A range wavelengths
were especially important in malignant melanoma induction. However, an action spectrum
using the opossum Monodelphis domestica found that the potency of UV-A for melanoma
induction was extremely low compared to that of UV-B (Robinson et al., 2000). A recent study
by De Fabo et al. (2004) examined the differences in wavelength effectiveness using a
10-28
-------
hepatocyte growth factor/scatter factor-transgenic mouse model. The epidermal tissue of these
transgenic mice behaves similar to the human epidermis in response to UV exposure. Given the
absence of a mammalian melanoma action spectrum, the standardized CIE erythema action
spectrum was used to deliver identical erythemally effective doses. Only UV-B radiation was
found to initiate mammalian cutaneous malignant melanoma. UV-A radiation, even at doses
considered physiologically relevant, were ineffective at inducing melanoma (De Fabo et al.,
2004). Overall, current evidence suggests that UV-B, and not UV-A, is the primary risk factor
for malignant melanoma (ICNIRP, 2004).
The populations susceptible for malignant melanoma are similar to those for nonmelanoma
skin cancers. Once again, individuals with xeroderma pigmentosum or a reduced capacity for
nucleotide excision repair are at increased risk (Tomescu et al., 2001; Wei et al., 2003).
Individuals with skin types I and II, or the fair-skin phenotype (blue or green eyes; blond or red
hair; skin that freckles, sunburns easily, and does not tan), have increased susceptibility to
malignant melanoma (Evans et al., 1988; Swetter, 2003; Veierad et al., 2003). However, the
incidence of melanoma was also positively associated with UV radiation in Hispanics and blacks
(Hu et al., 2004). Although the incidence of melanoma is much lower in Hispanics and blacks
compared to whites, melanomas in these populations are more likely to metastasize and have a
poorer prognosis (Black et al., 1987; Bellows et al., 2001). Among children, malignant
melanoma appears to have similar epidemiologic characteristics to the adult form of the disease
(Whiteman et al., 1997). Individuals with intermittent, intense sun exposure, particularly during
childhood, were found to have increased risk of melanoma (Whiteman et al., 2001), in contrast
to the association between cumulative exposure and increased risk of squamous cell carcinoma.
One study found that a personal history of nonmelanoma skin cancer or precancer, higher
socioeconomic status, and increased numbers of nevocytic nevi also were associated with
increased incidence of melanoma (Evans et al., 1988).
Effect of Changes in Tropospheric O3 Levels on Skin Cancer Incidence
The current evidence strongly suggests a causal link between exposure to UV radiation and
the incidence of both nonmelanoma and melanoma skin cancer. Genetic factors, including skin
phenotype and ability to repair DNA, affect an individual's susceptibility to skin cancer.
Quantifying the relationship between UV radiation and skin cancer is complicated by the
10-29
-------
uncertainties involved in the selection of an action spectrum and appropriate characterization of
dose (e.g., peak or cumulative levels of exposure, childhood or lifetime exposures). In addition,
there are multiple complexities in attempting to quantify the effect of tropospheric O3 levels on
UV-radiation exposure, as described in Section 10.2. The absence of published studies that
critically examine increased incidence of skin cancer attributable to decreased tropospheric O3
levels reflects the significant challenges in determining ground-level O3-related changes in UV
radiation exposure. An analysis by Lutter and Wolz (1997) attempted to estimate the effects of
a nationwide 10 ppb reduction in seasonal average tropospheric O3 on the incidence of
nonmelanoma and melanoma skin cancers and cataracts. Their estimate, however, depended
upon several simplifying assumptions, ranging from an assumed generalized 10 ppb reduction
in O3 column density, national annual average incidence rates for the two types of skin cancer,
and simple, linear biological amplification factors. Further, the methodologies used in this
analysis inherently have ignored area-specific factors that are important in estimating the extent
to which small, variable changes in ground-level O3 mediate long-term exposures to UV-B
radiation. More reasonable estimates of the human health impacts of enhanced UV-B
penetration following reduced surface O3 concentrations require both (a) solid understanding
of the multiple factors that define the extent of human exposure to UV-B at present and
(b) well-defined and quantifiable links between human disease and UV-B exposure. The reader
is referred to the U.S. EPA 2002 Final Response to Court Remand (Federal Register, 2003) for
detailed discussions of the data and scientific issues associated with the determination of public
health benefits resulting from the attenuation of UV-B by surface-level O3.
In the absence of studies specifically addressing the reduction of tropospheric O3 (by
assuming that the key variable is total column O3 density), inferences could be made concerning
the effects of reduced tropospheric O3-related increases in UV-B exposure on the basis of studies
focused on stratospheric O3 depletion. Several studies have examined the potential effect of
stratospheric O3 depletion on the incidence of skin cancer (De Gruijl, 1995; Longstreth et al.,
1995; Madronich and De Gruijl, 1993; Slaper et al., 1996; Urbach, 1997). Note that several of
the concerns expressed in relation to the Lutter and Wolz (1997) analysis are relevant here as
well. Stratospheric O3 depletion is likely to increase the ground-level UV-B flux, as O3 absorbs
radiation in that wavelength range with high efficiency. Because UV-B radiation is primarily
implicated in the induction of skin cancer, especially among persons with skin phenotypes I and
10-30
-------
II, there is concern that the depletion of the O3 layer would result in significantly increased
incidence of skin cancers.
Estimation of the increased risk in melanoma associated with stratospheric O3 depletion
cannot be done adequately due to the lack of a mammalian action spectrum for melanoma.
Furthermore, the complexity of the UV-related induction mechanism of melanoma adds an
additional layer of uncertainty to the calculations. The excess risk in nonmelanoma skin cancers
associated with a decrease in stratospheric O3 was estimated using the Skin Cancer Utrecht-
Philadelphia action spectrum based on hairless albino mice (Longstreth et al., 1995).
Quantification of how much more UV radiation would reach ground level with each percentage
decrease in O3 required several assumptions: (1) annual doses were an appropriate measure; (2)
personal doses were proportional to ambient doses; and, most notably, (3) each percentage
decrease in O3 was associated with a 1.2% increase in UV radiation. Next, the relationship
between UV radiation and nonmelanoma skin cancer incidence was determined: each percent
increase in annual UV radiation dose was estimated to cause a 2.5% increase in squamous cell
carcinoma and 1.4% increase in basal cell carcinoma over a human lifetime. Incorporating all
these factors, Longstreth et al. (1995) calculated that a sustained 10% decrease in stratospheric
O3 concentration would result in 250,000 additional nonmelanoma skin cancer cases per year.
Madronich and De Gruijl (1993) noted that the largest percent of O3-induced nonmelanoma skin
cancer increases would be at high latitudes, where baseline incidence of skin cancer is usually
small. Assuming a phaseout of primary O3-depleting substances by 1996, as established by the
Copenhagen Amendments in 1992, Slaper et al. (1996) estimated that the number of excess
nonmelanoma skin cancers in the U.S. caused by O3 depletion would exceed 33,000 per year (or
approximately 7 per 100,000) around the year 2050.
However, estimating the increase in nonmelanoma skin cancer incidence attributable to the
depletion of the stratospheric O3 layer is marred by uncertainty. The following statement by
Madronich and De Gruijl (1994) describes the uncertainty of estimating the effect of
stratospheric O3 depletion on the incidence of skin cancer:
Extrapolating trends and effects of UV into the future is very hypothetical due to uncertainties
that arise from atmospheric chemistry, epidemiology, and related disciplines. The values that we
calculated are one plausible measure of the magnitude of the 03-UV effects... The timescales for
atmospheric change and skin-cancer development are still far from certain: 03 reductions are
expected to continue well into next century, and the time between UV exposure and
development of skin cancer is essentially unknown.
10-31
-------
Therefore, much caution is necessary when assessing and interpreting the quantitative results of
excess nonmelanoma skin cancer incidence due to stratospheric O3 depletion. Although the
effect of reductions in tropospheric or ground-level O3 concentrations on skin cancer incidence
has not been assessed, it would be expected to be much less compared to the effect from the
depletion of the stratospheric O3 layer, given that tropospheric O3 makes up < 10% of the total
atmospheric O3.
10.2.3.4 Ocular Effects of Ultraviolet Radiation Exposure
Ultraviolet Radiation Exposure and Risk of Ocular Damage
Ocular damage from UV radiation exposures includes effects on the cornea, lens, iris, and
associated epithelial and conjunctival tissues. Absorption of UV radiation differs by
wavelength, with short wavelengths (<300 nm) being almost completely absorbed by the cornea,
whereas longer wavelengths are transmitted through the cornea and absorbed by the lens
(McCarty and Taylor, 2002). The most common acute ocular effect of environmental UV
exposure is photokeratitis, also known as snowblindness, caused by absorption of short
wavelength UV radiation by the cornea. The action spectrum indicated that maximum
sensitivity of the human eye was found to occur at 270 nm (ICNIRP, 2004; Pitts, 1993). The
threshold for photokeratitis in humans varied from 4 to 14 ml/cm2 for wavelengths 220 to
310nm.
Exposure to longer wavelengths has been shown to cause both transient and permanent
opacities of the lens, or cataracts. Extensive toxicologic and epidemiologic evidence supports
the causal association between UV radiation and cataracts (Hockwin et al., 1999; McCarty and
Taylor, 2002). Ultraviolet radiation-induced cataracts are hypothesized to be caused by
oxidative stress leading to increased reactive species in the lens, which then causes damage to
lens DNA and cross-linking of proteins. Exposure time to low-dose UV radiation was found to
strongly influence cataract formation (Ayala et al., 2000). An action spectrum determined using
young female rats indicated that the rat lens was most sensitive to 300 nm, correcting for corneal
transmittance (Merriam et al., 2000). Oriowo et al. (2001) examined the action spectrum for
cataract formation using whole cultured pig lenses. As pig lenses are similar in shape and size
to the human lens, some inferences may be made. Results indicated that the 270 to 315 nm
waveband was most effective in producing UV-induced cataracts in vitro. However, the
10-32
-------
threshold values varied widely within that range, from 0.02 J/cm2 for 285 nm to 0.74 J/cm2 for
315 nm (Oriowo et al., 2001). At wavelengths >325 nm, the threshold levels were orders of
magnitude larger, with a minimum threshold value of 18.7 J/cm2.
An epidemiologic study examined the effects of UV radiation on cataract formation in
watermen (e.g., commericial fishermen, boat workers) who worked on Chesapeake Bay, MD
(Taylor et al., 1988). Among the 838 individuals surveyed in this study, 111 had cortical
cataracts and 229 had nuclear cataracts. Results indicated that UV-B radiation was significantly
associated with cortical, but not nuclear, cataract formation. For a given age, a doubling of
cumulative UV-B exposure was associated with a 60% excess risk (95% CI: 1, 164) of cortical
cataracts. No association was observed between cataracts and UV-A radiation in this outdoor-
working population.
Risk of Ocular Damage from Changes in Tropospheric Ozone Levels
Cataracts are the most common cause of blindness in the world. McCarty et al. (2000)
calculated that ocular UV radiation exposure accounted for 10% of the cortical cataracts in an
Australian cohort of 4,744 individuals from both urban and rural areas. A study by Javitt and
Taylor (1994-1995) found that the probability of cataract surgery in the U.S. increased by 3% for
each 1° decrease in latitude. These results suggest that depletion of the stratospheric O3 layer
may increase UV radiation-induced cataract formation. After assuming a certain wavelength
dependency along with several additional assumptions, every 1% decrease in the stratospheric O3
layer was estimated to be associated with a 0.3 to 0.6% increase in cataracts (Longstreth et al.,
1995). Longstreth et al. (1995) noted that this estimate has a high degree of uncertainty due to
inadequate information on the action spectrum and dose-response relationships. Quantitative
estimates have not been possible for photokeratitis, pterygium, or other UV-related ocular effects
due to lack of epidemiologic and experimental data.
As is the case for all of the other UV-related health outcomes, there is no published
information on the potential effects on cataract formation due to any changes in surface-level
UV flux resulting from decreases in tropospheric O3.
10-33
-------
10.2.3.5 Ultraviolet Radiation and Immune System Suppression
Experimental studies have suggested that exposure to UV radiation may suppress local and
systemic immune responses to a variety of antigens (Clydesdale et al., 2001; Garssen and Van
Loveren, 2001; Belgrade et al., 1997). In rodent models, these effects have been shown to
worsen the course and outcome of some infectious diseases and cancers (Granstein and Matsui,
2004; Norval et al., 1999). Granstein and Matsui (2004) stated that exposure to UV-B radiation
caused immunosuppression in mice ultimately by releasing cytokines that prevent antigen-
presenting cells from performing their normal functions and causing direct damage to epidermal
Langerhans cells. Noonan et al. (2003) investigated UV skin cancer induction in two strains of
reciprocal Fl hybrid mice and found that genetically determined differences in susceptibility to
UV-induced immunosuppression was a risk factor for skin cancer. At high UV radiation doses,
mice with greater susceptibility to immune suppression had a larger proportion of skin tumors
compared to those with lower susceptibility (Noonan et al., 2003). In a study by Yoshikawa
et al. (1990), development of contact hypersensitivity to dinitrochlorobenzene on irradiated
buttock skin was examined. Individuals who failed to develop contact hypersensitivity were
considered to be susceptible to UV-B radiation. Virtually all skin cancer patients (92%) were
susceptible to UV-B radiation-induced suppression of contact hypersensitivity, compared to
approximately 40% of healthy volunteers. Others studies have observed increased skin cancer in
immune suppressed organ transplant patients (Caforio et al., 2000; Lindelof et al., 2000).
Collectively, results from these studies suggest that immune suppression induced by UV
radiation may be a risk factor contributing to skin cancer induction (Ullrich, 2005).
There is also some evidence that UV radiation has indirect involvement in viral
oncogenesis through the human papillomavirus (Pfister, 2003). Additional evidence of
UV-related immunosuppression comes from an epidemiologic study of 919 patients with rare
autoimmune muscle diseases from 15 cities on four continents with variable UV radiation
intensity (Okada et al., 2003). Ultraviolet radiation was strongly associated with the prevalence
of dermatomyositis, an autoimmune disease distinguished by the presence of photosensitive
pathognomonic rashes (Okada et al., 2003). In patients with the human immunodeficiency virus,
UV-B radiation lead to activation of the virus in their skin through the release of cytoplasmic
nuclear factor kappa B (Breuher-McHam et al., 2001). In a study by Belgrade et al. (2001),
UV-induced immunosuppression was examined in 185 subjects with different skin
10-34
-------
pigmentations. To assess immune suppression, dinitrochlorobenzene was applied to irradiated
buttock skin 72 hours after irradiation. Differences in sensitivity were unrelated to skin type
based on the Fitzpatrick classification or minimal erythemal dose (Belgrade et al., 2001).
However, erythemal reactivity, assessed by the steepness of the erythemal dose-response curve,
was shown to be significantly associated with UV-induced immunosuppression. Only subjects
with steep erythemal responses, which included individuals with skin types I through V, showed
a dose-response relationship between UV exposure and immune suppression (Belgrade et al.,
2001).
In other studies, UV radiation was associated with decreased autoimmune diseases.
Several ecologic studies observed a decreased prevalence of multiple sclerosis, insulin-
dependent diabetes mellitus, and rheumatoid arthritis in regions with lower latitude (i.e., higher
UV radiation exposure) (Ponsonby et al., 2002). These results may be attributable to UV
radiation-induced immunosuppression and UV-B-related production of vitamin D, which has
immunomodulatory effects (Cantorna, 2000). The protective effects of UV radiation resulting
from its active role in vitamin D production are further discussed in the next section.
Most action spectrum investigations have concluded that immunosuppression is caused
most effectively by the UV-B waveband (Garssen and Van Loveren, 2001). The effects of UV-
A on local and systemic immunosuppression have been unclear and inconsistent. There is some
evidence that high doses of UV-A are protective of immunosuppression induced by UV-B
exposure (Halliday et al., 2004). Given the variety of outcomes of immune suppression and
possible mechanisms of effect, little detailed information exists on UV radiation action spectra
and dose-response relationships. The available data are insufficient to conduct a critical risk
assessment of UV radiation-induced immunosuppression in humans.
10.2.3.6 Protective Effects of Ultraviolet Radiation—Production of Vitamin D
Any risk assessment that attempts to quantify the consequences of increased UV-B
exposure on humans due to reduced ground-level O3 must include consideration of both negative
and positive effects. A potential health benefit of increased UV-B exposure relates to the
production of vitamin D in humans. Most humans depend on sun exposure to satisfy their
requirements for vitamin D (Holick, 2004). UV-B photons are absorbed by
7-dehydrocholesterol in the skin, leading to its transformation to previtamin D3, which is rapidly
10-35
-------
converted to vitamin D3. Vitamin D3 is metabolized in the liver, then in the kidney to its
biologically active form of 1.25-dihydroxyvitamin D3. One minimal erythemal dose produces
vitamin D equivalent to an oral dose of 20,000 IU vitamin D, which is 100 times the
recommended dietary allowance for adults under 50 years of age (Giovannucci, 2005;
Holick, 2004).
Vitamin D deficiency can cause metabolic bone disease among children and adults, and
also may increase the risk of many common chronic diseases, including type I diabetes mellitus
and rheumatoid arthritis (Holick, 2004). Substantial in vitro and toxicologic evidence also
support a role for vitamin D activity against the incidence or progression of various cancers
(Giovannucci, 2005; Studzinski and Moore, 1995). Large geographical gradients in mortality
rates for a number of cancers in the U.S. are not explained by dietary or other risk factors;
therefore, it has been hypothesized that some carcinomas may be due to latitude-related
reduction in UV-B radiation exposure. Published literature indicates that solar UV-B radiation,
by increasing vitamin D production, is associated with a reduced risk of cancer. Most of these
studies used an ecologic study design, in which latitude gradient was examined in relation to
cancer rates. Kimlin and Schallhorn (2004) observed that latitude was a valid predictor of
vitamin D-producing UV radiation. The strongest evidence exists for an association between
UV radiation and reduced risk of colorectal cancer (Giovannucci, 2005; Grant and Garland,
2004; Freedman et al., 2002). Several other studies also have found an inverse relationship
between UV radiation and various other cancers, including cancer of the breast (Freedman et al.,
2002; Garland et al., 1990; Gorham et al., 1990; Grant, 2002a; John et al., 1999), ovary
(Freedman et al., 2002; Lefkowitz and Garland, 1994), and prostate (Freedman et al., 2002;
Hanchette and Schwartz, 1992), as well as non-Hodgkin lymphoma (Hughes et al, 2004; Hartge
et al., 1996). Eight other cancers (i.e., bladder, esophageal, kidney, lung, pancreatic, rectal,
stomach, and corpus uteri) have been found to exhibit an inverse correlation between mortality
rates and UV-B radiation (Grant, 2002b).
Using UV-B data from July 1992 and U.S. cancer mortality rates from 1970 to 1994,
premature cancer deaths attributable to insufficient UV-B exposure were analyzed in an ecologic
study (Grant, 2002b). The minimum mortality rate, which was determined as the value
corresponding to the maximum UV-B dose, was used to calculate the number of premature
deaths. This analysis observed that the annual number of premature deaths from various cancers
10-36
-------
due to latitude-related reduction in UV-B exposure was 21,700 (95% CI: 20,400, 23,400) for
white Americans; 1,400 (95% CI: 1,100, 1,600) for black Americans; and 500 (95% CI: 400,
600) for Asian Americans and other minorities. Uncertainty in the estimations of UV-B
exposure limits the confidence for the estimates of excess cancer deaths attributable to
insufficient exposure. Caution is required in interpreting results from ecologic data; however, no
strong alternative explanation is indicated in the association observed between UV radiation and
the decreased risk of cancer (Giovannucci, 2005). No study has assessed the decreased risk of
cancer mortality resulting from increased UV radiation attributable to decreased tropospheric O3
levels, but the change in risk is expected to be unappreciable.
In establishing guidelines on limits of exposure to UV radiation, the ICNIRP agreed that
some low-level exposure to UV radiation has health benefits (ICNIRP, 2004). However, the
adverse health effects of higher UV exposures necessitated the development of exposure limits
for UV radiation. The ICNIRP recognized the challenge in establishing exposure limits that
would achieve a realistic balance between beneficial and adverse health effects.
10.2.4 Summary and Conclusions for Ozone Effects on UV-B Flux
Latitude and altitude are primary variables in defining UV-B flux at the Earth's surface,
immediately followed in importance by clouds, surface albedo, PM concentration and
composition, and then by gas phase pollution. Of all of these, only latitude and altitude can be
defined with small uncertainty in any effort to develop a UV climatology for use in a public
health benefits analysis relevant to the areas not presently attaining the NAAQS for O3. Cloud
cover, and its effect on surface UV flux, continues to be extremely difficult to define and predict.
Particulate matter and gas-phase tropospheric pollutants are subject to similarly high degrees of
uncertainty in predicting their relative concentration distributions. Land cover and,
consequently, surface albedos are highly variable at the geographic scales relevant to NAAQS
attainment.
Within the uncertain context of presently available information on UV-B surface fluxes, a
risk assessment of UV-B-related health effects would need to factor in human habits (e.g., daily
activities, recreation, dress, and skin care) in order to adequately estimate UV-B exposure levels.
Little is known about the impact of variability in these human factors on individual exposure to
UV radiation. Furthermore, detailed information does not exist regarding the relevant type (e.g.,
10-37
-------
peak or cumulative) and time period (e.g., childhood, lifetime, or current) of exposure,
wavelength dependency of biological responses, and interindividual variability in UV resistance.
Recent reports of the necessity of UV-B in the production of vitamin D—a vitamin in which
many individuals are deficient—suggests that increased risks of human disease due to a slight
excess in UV-B radiation exposure may be offset by the benefits of enhanced vitamin D
production. However, as with other impacts of UV-B on human health, this beneficial effect of
UV-B has not been studied in sufficient detail to allow for a credible health benefits assessment.
In conclusion, the effect of changes in surface-level O3 concentrations on UV-induced health
outcomes cannot yet be critically assessed within reasonable uncertainty.
10.3 TROPOSPHERIC OZONE AND CLIMATE CHANGE
Water vapor, CO2, O3, N2O, CH4, CFCs, and other polyatomic gases present in the Earth's
troposphere, trap infrared radiation emitted by the Earth's surface, leading to surface warming.
This phenomenon is widely known as the "Greenhouse Effect" (Arrhenius, 1896), and the gases
involved are known as "greenhouse gases" (GHGs). The term used for the role a particular
atmospheric component, or any other component of the greater climate system, plays in altering
the Earth's radiative balance is "forcing." In the past decade, the global atmospheric sciences
and climate communities have made significant progress in determining the specific role that
atmospheric O3 plays in forcing climate.
The Intergovernmental Panel on Climate Change (IPCC) was founded in 1988 by the
World Meteorological Society (WMO) and the United Nations Environmental Program (UNEP)
to support the work of the Conference of Parties (COP) to the United Nations Framework
Convention on Climate Change (UNFCCC). Drawing from the global climate and atmospheric
sciences community for its authors and reviewers, the IPCC produces reports containing
thorough assessments of the available peer-reviewed science regarding the physical climate
system, past and present climate, and evidence of human-induced climate change. This section
summarizes the reviews of the available information on the forcing properties of tropospheric O3
as provided by the IPCC Third Assessment Report (IPCC, 200 la) and also describes some of the
more recent developments on the subject.
10-38
-------
The projected effects of global climate change will be briefly explained to provide the
context within which O3 serves as a regional, and possibly global, anthropogenic pollutant.
The concept of climate forcing is also explained, along with the factors that influence the extent
of climate forcing by O3. The section concludes with a summary of the various estimates that
have been placed on the amount of globally averaged forcing due to O3.
10.3.1 The Projected Impacts of Global Climate Change
The study of the atmospheric processes involved in global climate change, and its potential
consequences for human health and global ecosystems, is an area of active research. The IPCC
Third Assessment Report (TAR) is the most thorough evaluation currently available of the
science concerning climate change. In addition to the first and second IPCC assessments in
1990 and 1995, along with other IPCC reports, earlier assessments included those conducted by
the UNEP (1986), the WMO (1988), the U.S. Environmental Protection Agency (1987), and
others (e.g., Patz et al., 2000a,b). The reader is referred to those documents for a complete
discussion of climate change science. An abbreviated list of the IPCC conclusions to date and a
short discussion of the potential impacts of climate change on human health and welfare is
provided here to serve as the context for the discussion of the role of the increasing tropospheric
O3 concentration in climate change.
According to various historic and modern measurement records, atmospheric GHG
concentrations have increased dramatically in the past century and have been attributed to human
activities. The IPCC TAR describes the scientific theory and evidence linking increases
in GHGs to human activities (IPCC, 200la).
An increasing body of geophysical observations shows that the Earth is warming and that
other climate changes are underway. These observations include the global surface temperature
record assembled since the year 1860, the satellite temperature record begun in 1979, recorded
changes in snow and ice cover since the 1950s, sea level measurements taken throughout the
20th century, and sea surface temperature observations recorded since the 1950s.
Observations (Levitus et al., 2005) show that -84% of the total heating of the Earth System
(oceans, atmosphere, continents, and cryosphere) over the last 40 years has gone into warming
the oceans. Barnett et al. (2005) have reported the emergence of a clear pattern of ocean surface
warming associated with anthropogenic GHGs. The authors constructed a model-based
10-39
-------
fingerprint (i.e., a map of predicted changes in the vertical temperature profile of the Earth's six
major oceans) and compared this map to the newly upgraded and expanded ocean temperature
data set (Levitus et al., 2005). They concluded that the warming signal far exceeds what would
be expected from natural variability, a finding that was in compelling agreement with GHG-
forced model profiles. Other evidence of ocean warming includes a marked increase in the
frequency, intensity, and persistence of the zonal atmospheric circulation shifts known as the
El Nino-Southern Oscillation (ENSO) over the past 100 years. ENSO events occur when the
tropical Pacific Ocean has accumulated a large, localized mass of warm water that interrupts
cold surface currents along South America, altering precipitation and temperature patterns in the
tropics, subtropics, and the midlatitudes.
IPCC (1998, 200la) reports also describe the results of general circulation model (GCM)
studies predicting that human activities will alter the climate system in a manner likely to lead to
marked global and regional changes in temperature, precipitation, and other climate properties.
These changes are expected to increase the global mean sea level as well as increase the number
of extreme weather events such as floods and droughts, increased wind speeds and precipitation
intensity of tropical cyclones, and changes in soil moisture. These predicted changes can be
expected to directly impact human health, ecosystems, and global economic sectors (e.g.,
hydrology and water resources, food and fiber production) (IPCC, 1998, 2001b). Table 10-1
summarizes these projected impacts.
Wide variations in the course and net impacts of climate change in different geographic
areas are expected. In general, the projected changes constitute additional stressors on natural
ecosystems and human societal systems already impacted by increasing resource demands,
unsustainable resource management practices, and pollution. Some of the predicted changes
include alterations in ecological balances; in the availability of adequate food, water, clean air;
and in human health and safety. Poorer nations can be expected to suffer the most, given their
limited adaptive capabilities.
Although many regions are predicted to experience severe, possibly irreversible, adverse
effects due to climate change, beneficial changes may also take place. For example, certain
regions may benefit from warmer temperatures or increased CO2 fertilization, e.g., U.S. West
Coast coniferous forests, and some Western rangelands. Specific benefits may include reduced
energy costs for heating, reduced road salting and snow-clearance costs, longer open-water
10-40
-------
Table 10-1. Examples of Impacts Resulting From Projected Changes in Extreme
Climate Events
Projected changes during the 21st Century
in Extreme Climate Phenomena and their
Likelihood"
Representative Examples of Projected Impacts'"
(all high confidence of occurrence in some areasc)
Simple Extremes
Higher maximum temperatures; more hot
days and heat waves'1 over nearly all land areas
(very likely")
Higher (increasing) minimum temperatures;
fewer cold days, frost days, and cold waves'1
over nearly all land areas (very likely")
More intense precipitation events
(very likely" over many years)
Complex Extremes
Increased summer drying over most
midlatitude continental interiors and
associated risk of drought (likely")
Increase in tropical cyclone peak wind
intensities, mean and peak precipitation
intensities (likely" over some areas)e
Intensified droughts and floods associated
with El Nino events in many different regions
(likely") (see also under droughts and intense
precipitation events)
Increased Asian summer monsoon precipitation
variability (likely")
Increased intensity of midlatitude storms
(little agreement between current models/
• Increased incidence of death and serious illness in older age groups
and urban poor
• Increased heat stress in livestock and wildlife
• Shift in tourist destinations
• Increased risk of damage to a number of crops
• Increased electric cooling demand and reduced energy supply reliability
• Decreased cold-related human morbidity and mortality
• Decreased risk of damage to a number of crops, and increased risk
to others
• Extended range and activity of some pest and disease vectors
• Reduced heating energy demand
• Increased flood, landslide, avalanche, and mudslide damage
• Increased soil erosion
• Increased flood runoff could increase recharge of some floodplain
aquifers
• Increased pressure on government and private flood insurance systems
and disaster relief
Decreased crop yields
Increased damage to building foundations caused by ground shrinkage
Decreased water resource quantity and quality
Increased risk of forest fire
Increased risk to human life, risk of infections, disease epidemics,
and many other risks
Increased coastal erosion and damage to coastal buildings and
infrastructure
Increased damage to coastal ecosystems such as coral reefs and
mangroves
Decreased agricultural and rangeland productivity in drought- and
flood-prone regions
Decreased hydropower potential in drought-prone regions
Increased flood and drought magnitude and damages in temperate
and tropical Asia
Increased risks to human life and health
Increased property and infrastructure losses
Increased damage to coastal ecosystems
"Likelihood refers to judgmental estimates of confidence used by TAR WGI: very likely (90-99% chance); likely
(66-90% chance). Unless otherwise stated, information on climate phenomena is taken from the Summary for
Policymakers, TAR WGI. TAR WGI = Third Assessment Report of Working Group 1 (IPCC, 2001a).
bThese impacts can be lessened by appropriate response measures.
GHigh confidence refers to probabilities between 67 and 95%.
dlnformation from TAR WGI, Technical Summary.
'Changes in regional distribution of tropical cyclones are possible but have not been established.
Source: IPCC (2001 b).
10-41
-------
seasons in northern channels and ports, and improved agricultural opportunities in the northern
latitudes, as well as in the Western interior and coastal areas. For further details about the
projected effects of climate change on a U.S.-regional scale, the reader is also referred to several
regionally-focused reports (MARAT, 2000; Yarnal et al., 2000; NERAG, 2001; GLRAG, 2000),
as well as a report on potential impacts of climate change on human health (Bernard et al.,
2001a,b). The IPCC report, "The Regional Impacts of Climate Change," (IPCC, 1998) describes
the projected effects of human-induced climate change on various regions of the globe, including
Africa, the Arctic and Antarctic, the Middle East and arid Asia, Australasia, Europe, Latin
America, North America, the small island nations, temperate Asia, and tropical Asia.
While current climate models can successfully simulate the present-day annual mean
global climate and the seasonal cycles on a continental scale, they have been less successful on a
regional scale. Clouds and humidity, essential factors in defining local and regional (sub-grid
scale) climate, are significantly uncertain (IPCC, 200la). Due to modeling uncertainties, both in
reproducing regional climate and in predicting the future economic activity that will govern
future GHG emissions, the projected impacts discussed above are also uncertain.
Findings from the U.S. Global Change Research Program (USGCRP) (NAST, 2000) and
related reports illustrate the considerable uncertainties and difficulties in projecting likely
climate change impacts at the regional or local scale. The USGCRP findings also reflect the
mixed nature of projected potential climate change impacts, i.e., combinations of deleterious and
beneficial effects, for U.S. regions and the variation of projected impacts across different
regions. Difficulties in projecting region-specific climate change impacts are complicated by the
need to evaluate the potential effects of regional- or local-scale changes in key air pollutants not
only on global-scale temperature trends, but also on regional- or local-scale temperature and
precipitation patterns. The EPA is currently leading a research effort that uses regional-scale
climate models with the goal of identifying changes to O3 and PM concentrations that may occur
in a warming climate. An assessment of the results of this effort will be available by the next
review of the O3 NAAQS. This focused effort to determine the impact of a warming climate on
criteria air pollution requires regional-scale models with improved skill in reproducing climate
history and predicting change. Among the innovations being employed in this effort is the
downscaling of global circulation model outputs to provide boundary conditions for model
calculations at the regional scale (Liang et al., 2005). While focusing on projecting the impact of
10-42
-------
a warming climate on regional O3 concentrations, the effort applied to improving regional-scale
modeling will also lead to improved estimates of current and projected future impacts of
tropospheric O3 on climate.
10.3.2 Solar Energy Transformation and the Components of the Earth's
Climate System
Mass, in any form, has the capacity to interact with solar and terrestrial radiation, but the
manner in which it interacts with radiation is governed by its particular physical form and/or
molecular properties. Water provides one of the most interesting examples of how physical form
affects radiative properties. In its gaseous form, water is the most important GHG present in the
climate system, because of its ability to absorb long-wave terrestrial radiation. Conversely, in its
frozen form as snow or sea ice, water plays a very important role in the climate system by
scattering UV and visible solar radiation back to space, i.e., decreasing the Earth's net solar
radiation receipts by increasing the Earth's reflective properties (albedo). In its liquid aerosol
form as clouds, water also greatly increases the Earth's albedo. In its bulk liquid form as ocean
water, it absorbs terrestrial radiation and represents the Earth's most important reservoir of
heat energy.
The atomic composition and molecular structure of a gas determines the wavelengths of
light it can absorb and, therefore, its role in defining the heat capacity of the atmosphere. Ozone
and O2 provide examples of the relative importance of these molecular properties. While these
molecules are both composed solely of oxygen atoms, their bond structures are distinct. Ozone
has a three-atom, bent molecular structure, giving it the capacity to absorb terrestrial (infrared)
wavelengths - making it a GHG. At any altitude, i.e., in the stratosphere or troposphere, O3 has
the capacity to absorb UV radiation of 320 nm and shorter, further increasing the energy-
absorbing capacity of the troposphere. Conversely, O2, due to its diatomic, linear structure, is
limited to absorbing very short-wave UV light—and does so at altitudes too high to influence the
climate system significantly.
Each component of the climate system plays a role in absorbing, transforming, storing,
dispersing, or scattering solar radiation. Weather is a tangible consequence of the transformation
and dispersion of terrestrial radiation within the atmosphere. The term "weather" refers to the
condition of the Earth's atmosphere at a specific time and place. It is defined by several specific
10-43
-------
variables: the air temperature, air pressure, humidity, clouds, precipitation, visibility, and wind
speed. The "climate" for a given place on the Earth's surface is a long-term average of these
variables accounting for daily and seasonal weather events. The frequency of extreme weather
events is used to distinguish among climates that have similar averages (Ahrens, 1994).
Climate components (including GHGs, land, oceans, sea ice, land ice and snow,
atmospheric particles, vegetation, clouds, etc.) all contribute to the Earth's heat capacity, i.e., its
ability to absorb and retain solar energy. Changes in the properties (or mass) of these
components will "force" the climate system in one direction or the other, i.e., warmer versus
cooler. The transformation of atmospheric O2 into O3 by way of air pollution chemistry,
enhances the heat capacity of the atmosphere. The principles behind the important concept of
climate forcing are further described below.
10.3.3 The Composition of the Atmosphere and the Earth's
Radiative Equilibrium
The "greenhouse effect" is the term given to the decreased rate of reemission of absorbed
solar energy due to the heat-retaining properties of the Earth's atmosphere. According to simple
radiative transfer theory, at thermal equilibrium, the Earth's temperature should be near -15 °C.
This is the temperature of a theoretical "black body" that is receiving and then reemitting
342.5 WnT2, i.e., the globally averaged amount of full-spectrum solar energy absorbed and then
reemitted by the Earth as infrared terrestrial radiation per square meter. In fact, satellite
observations well above the atmosphere indicate that the Earth's average planetary temperature
is remarkably close to its theoretical black body value at -18 °C, a temperature at which liquid
water ordinarily does not exist.
At its surface, however, the Earth's average temperature is +15 °C. The +33 °C
temperature differential between the Earth's planetary and surface temperatures is due to the
presence of infrared (IR) radiation-absorbing components in the atmosphere such as water
vapor, CO2, CH4, several other trace gases, and some types of particles and clouds.
The atmosphere, when cloud-free, is largely transparent in the solar wavelength range.
A small fraction of this radiation is absorbed and reemitted as black body radiation by dark
atmospheric particles (IPCC, 2001a). However, the majority of clouds and particles, in part,
10-44
-------
offset the greenhouse effect by increasing the Earth's albedo, thereby decreasing the overall
amount of solar radiation absorbed by the Earth system.
Ozone, SO2 and NO2 also absorb ultraviolet and near ultraviolet wavelengths, in addition to
infrared radiation. Once absorbed by a gas molecule, the energy introduced by a photon may
induce a photochemical reaction with the residual energy thermally exciting (heating) the
products of the reaction. Alternatively, the energy introduced into the molecule by the photon
may be dispersed amongst neighboring molecules via intermolecular collisions, or it may be
reemitted in part as a lower energy (i.e., IR) photon.
Radiation from the sun or the Earth's surface that is absorbed by gases and particles is
reemitted isotropically, i.e., it is equally likely to be emitted in all directions. Therefore, to a
first approximation, half of the radiation trapped by the Earth's atmosphere is reflected back to
its surface. A portion of this radiation is transformed into the heat energy that drives the
atmospheric processes that form the basis of weather and climate. Radiation that is not absorbed
by gases and aerosols reaches the Earth's surface where it is scattered (reflected) or absorbed,
depending on the surface albedo.
Successful modeling of the Earth's climate and, therefore, the assessment of the extent of
human-induced climate change and development of appropriate policy depend on the quality of
available information on the relative efficiencies, amounts, and spatial and temporal distributions
of the various radiatively active components of the atmosphere that absorb and/or reflect solar
and terrestrial radiation, along with all the other nonatmospheric components of the Earth
system.
10.3.3.1 Forcing of the Earth's Radiative Balance
As mentioned earlier, the commonly used measure of the relative influence of a given
component of the climate system on the Earth's radiative balance is its radiative forcing (IPCC,
2001a; Houghton et al., 1990). Radiative forcing, in WnT2, is a quantity that was developed by
the climate modeling community as a first-order-only means of estimating relative effects of
individual anthropogenic and natural processes on the energy balance within the climate system.
When the effect of a particular component of the climate system is to reduce the amount of
solar energy absorbed, usually by increasing the Earth's albedo, this component is said to
provide a "negative" forcing, measured in WnT2. The convention assigns a positive value to the
10-45
-------
forcing induced by climate system components that enhance the greenhouse effect or otherwise
act to increase the heat-absorbing capacity of the Earth system. Purely reflective atmospheric
aerosol, clouds, white rooftops, snow-covered land surfaces, and dense sea ice provide a
negative forcing, while highly absorbing dark-colored atmospheric aerosols, GHGs, and
increases in dark ocean surface area, due to the melting of sea ice sheets, positively force the
climate system.
Global and regional climate are roughly defined by the balance between the large number
of positive and negative forcings induced by the many different components of the Earth system
and any changes in the properties of those components due to natural processes or anthropogenic
activities. Following a perturbation or added forcing, such as an increase in GHG concentrations
or modification the Earth's albedo through changes in land use, this balance is re-established via
a complex redistribution of energy within the Earth system. Feedback mechanisms that are
theorized but difficult to resolve at the quantitative level further complicate the prediction of the
sensitivity of variables, such as surface temperature, to changes in forcing.
A simple example of a positive feedback would be melting sea ice. As sea ice melts with
increasing ocean temperatures, the dark ocean surface that is revealed is more efficient at
absorbing IR radiation, further increasing the rate of warming. A negative feedback would be
the formation of clouds over a moist, warming surface. As clouds form, less radiation is
available to warm the surface, leading to cooling. The role of feedbacks in determining the
sensitivity of climate to changes in radiative forcing is described in detail in the IPCC TAR
(IPCC, 200la).
Discussions are presently underway within the climate community regarding a metric to
replace forcing as the standard measure of climate impact—one that will account for more of the
factors that determine the effectiveness of a specific change in altering climate. However,
forcing remains the current standard (NRC, 2005).
The IPCC has reported estimated values for forcing by individual radiatively active gases,
and by particle-phase components of the atmosphere that were derived primarily through expert
judgment incorporating the results of peer-reviewed modeling studies. The forcing estimates,
shown in Figure 10-6, are global averages attributed to known GHGs, including O3, particles,
anthropogenic cirrus clouds, land-use change, and natural solar flux variations. Uncertainty
ranges are assigned to reflect the range of modeled values reported in those studies. The current
10-46
-------
£ o,
re .E
I I
fi
ng
S i>
o =
£
re
2-
.
i -
-1-
-2-
_L_
1
Halocarbons
N2O
CH4 f
Aerosols
Rlark
^
__ carbon from
CO2
Tropospheric
Ozone
pri
I T 1 I I
^^ \ T 1
Stratospheric
ozone 1
Sulphate
High Medium Medium Low
'fuel' Mineral Aviation-induced
burning Dust , , Solar
r-J-|
Organic 1
carbon Biomass
from, burning
fossil a
fuel
burning
Very Very Very
Low Low Low
Contrails Cirrus
HH
Land-
^erosol use
ndirect (albedo)
effect only
1
Very Very Very Very
Low Low Low Low
-±1
Very
Low
Level of Scientific Understanding
Figure 10-6. Estimated global mean radiative forcing exerted by gas and various particle
phase species for the year 2000, relative to 1750.
Source: IPCC(2001a).
estimate of forcing due to long-lived, well-mixed, GHGs accumulated in the atmosphere from
the preindustrial era (ca., 1750) through the year 2000 is +2.4 WnT2 ± 10% (IPCC, 2001a).
An indication of the level of confidence in each of these estimates is given along the bottom of
this figure, again reflecting the expert judgment of the IPCC.
The IPCC reported a global average forcing value of 0.35 ± 0.15 WnT2 for tropospheric O3,
based on model calculations constrained by climatological observations. The considerations and
studies used to estimate this value are outlined below. Hansen and Sato (2001), accounting for
uncertainties in pre-industrial emissions levels, more recently estimated a value of 0.5 ± 0.2
WnT2 for forcing by O3.
10.3.4 Factors Affecting the Magnitude of Climate Forcing by Ozone
The radiative properties of O3 are distinct from those of other important GHGs in that it is
capable of absorbing both UV and IR radiation. Furthermore, it is able to absorb long-wave
radiation in a portion of the IR spectrum where water vapor does not absorb, i.e., the 9 tolO mm
10-47
-------
wavelength range, meaning that its ability to trap heat and force climate are unchanged by
variations in humidity. Given its relatively short atmospheric lifetime in comparison to other
GHGs, the distribution of tropospheric O3 is highly variable in geographic extent and time.
These properties enhance the prospect of attributing a unique geographic and time-dependent
pattern or fingerprint to forcing by O3.
Due to human activities, tropospheric O3 is estimated to have provided the third largest
increase in direct forcing since preindustrial times. It may also play a role in indirect forcing
through its participation in the oxidative removal of other radiatively active trace gases, such as
CH4 and the HCFCs. The direct and indirect forcing that tropospheric O3 imposes on the climate
system depends upon its geospatial and temporal distribution, but it also depends upon its
vertical position (altitude) in the atmosphere and the albedo of the underlying surface.
10.3.4.1 The Global Burden of Tropospheric Ozone
Little historical data exist that may be used to estimate the global O3 burden prior to
industrialization, although a few late 19th-century measurements suggest that O3 has more than
doubled in Europe during the 20th century. The insufficient data record on preindustrial
tropospheric O3 distributions introduces a major uncertainty in the estimation of the change
in O3-induced forcing since that period (IPCC, 2001a; Mickley et al., 2004a; Mickley et al.,
2001; Shindell and Faluvegi, 2002).
Ozone reacts photochemically at time scales that are generally shorter than those for large-
scale mixing processes in the atmosphere. Concentrated O3 plumes evolve downwind of strong
sources of its precursor pollutants: reactive nitrogen, CO, and non-methane hydrocarbons
(NMHCs). The most important of these sources are midlatitude industrialized areas and tropical
biomass burning. When viewed from above the atmosphere by satellite-borne spectrometers, O3
enhancements appear as relatively localized air masses or regional-scale plumes, usually
originating from industrialized areas or areas in which active biomass burning is underway. The
IPCC (200la) describes the efforts of several research teams who have analyzed data supplied by
the satellite-borne Total Ozone Mapping Spectrometer (TOMS) and other remote-sensing
instruments to map the global distribution of tropospheric O3 and to attempt to identify processes
that influence the global tropospheric O3 budget (IPCC, 200la). More recently, coincident
observations of total O3 by TOMS and the Solar Backscattered UV (SBUV) instrument were
10-48
-------
used by Fishman et al. (2003) to construct well-resolved spatial and temporal maps of the
regional distribution of tropospheric O3. Their results were consistent with those reported by
others, but with higher regional-scale resolution. They reported large O3 enhancements in the
southern tropics in austral spring and in the northern temperate latitudes in the summer. The
regional nature of high O3 concentrations was clearly visible in northeastern India, the eastern
United States, eastern China, and west and southern Africa, each coincident with high population
densities. Fishman et al. (2003) noted, as have the other groups cited above, significant
interannual variability in the concentrations observed over these regions. In situ measurements
of tropospheric O3 concentrations range from 10 ppb over remote oceans to 100 ppb in both the
upper troposphere and in plumes downwind from polluted metropolitan regions (IPCC, 200la).
Ground-level concentrations in urban areas are often >100 ppb. In the Southern Hemisphere,
one of the largest sources of O3 precursors is biomass burning. Biomass burning elevates O3on
large spatial scales, particularly in the tropical Atlantic west of the coast of Africa and in
Indonesia.
In its third assessment report, the IPPC estimates placed the global burden of tropospheric
O3 at a highly uncertain 370 Tg, equivalent to an average column density of 34 Dobson Units (1
DU = 2.687 x 1016 molecules/cm"2) or a mean concentration of 50 ppb (IPCC, 2001a).
Accounting for differences in levels of industrialization between the hemispheres, the average
column burden in the Northern Hemisphere is estimated to be 36 DU, with the Southern
Hemisphere estimated to average 32 DU. Due to its rapid photochemistry, individual surface
measurements of tropospheric O3 cannot capture large-scale concentrations, nor will they
represent the higher altitude concentrations. Dense surface and vertical measurements
(ozonesondes) would be required to supplement available output from remote sensing
instruments to provide the complete set of observations needed to derive a credible global O3
budget. Such a measurement program appears, at present, to be impractical.
10.3.4.2 Background Concentrations versus Regionally-Oriented Ozone Enhancements
Vingarzan (2004) surveyed the air quality literature and reported that annual average
background O3 concentrations at ground level in the Northern Hemisphere appear to range
between 20 and 45 ppb, depending upon geographic location, elevation, and the influence of
local sources. Fiore et al. (2003) modeled the U.S. continental O3 concentrations and found that
10-49
-------
surface background levels overlap the lower end of the range reported by Vingarzan (2004), e.g.,
15 to 35 ppb, with higher levels (40 to 50 ppb) arising at high-elevation sites due to the influence
of the upper troposphere (See Chapter 3 and its associated annexes for a complete discussion of
"policy relevant background [PRB]). Local- and regional-scale enhancements in O3 may be
thought of as roughly superimposed upon these background levels, with the exception of longer
stagnation events in which preexisting background O3 reacts away or is deposited as fresh O3 is
produced from locally-emitted precursors.
Lin et al. (2001) analyzed the EPA AIRS database for the 1980-1998 period and noted
that O3 concentrations have declined at the high end of the probability distribution, consistent
with the effects of emissions controls, but had increased at the low end of the distribution by 3 to
5 ppb. They divided the monitoring data for the continental U.S. into 4 quadrants by geography
and noted a pattern of increase for the Western states that might be attributed to the long-range
transport of O3 precursors from Asia. They found, however, that the Northeastern quadrant had
the highest increase in the low end of the concentration probability distribution, which could not
be reasonably attributed to transboundary transport of O3 precursors.
While not representing an ideal source of information for assessing the climatic effects
of O3 within the continental United States, data from the large air-quality-focused ground-based
monitoring network may be used to identify boundary-layer geospatial and temporal patterns
in O3 concentrations for comparison to regional-scale chemistry/climate models. Extensive
analysis of data available within the EPA AQS database can be found in Chapter 3 of this
document, including an analysis showing the diurnal O3 concentration patterns for several large
metropolitan areas with peak values ranging up to!60 ppb (Los Angeles). Lehman et al. (2004)
analyzed the AQS database of daily 8-h maximum O3 concentrations collected for 1,090 stations
in the eastern half of the United States for the 1993 to 2002 period. They applied a rotated
principle component analysis to a reasonably complete, spatially representative, nonurban subset
of the database in order to identify coherent, regionally oriented patterns in O3 concentrations.
Five spatially homogenous regions were identified: the U.S. Northeast, Great Lakes, Mid-
Atlantic, Southwest (including Alabama, Louisiana, Texas, Oklahoma), and Florida. The
Mid-Atlantic region displayed the highest mean concentration (52 ppb) of all of the regions
analyzed, followed by the Great Lakes, Southwest, and Northeast regions with around 47 ppb.
The average concentration derived for Florida was 41 ppb. The authors found strong
10-50
-------
correlations in measured concentrations among stations within the same region, suggesting that
the geospatial patterns of pollutant emissions and meteorological activity may also have a
regional orientation. These results suggest that these regions may define natural domains for
regional scale modeling studies of the influence of O3 (as well as PM) on climate.
10.3.4.3 Ozone Trends: Globally and in North America
For the Northern Hemisphere, weekly continuous data are available from 1970 for only
nine stations in the latitude range 36° N to 59° N (IPCC, 2001a). Available tropospheric O3
measurements do not reveal a clear trend in concentration, while trends in the stratosphere are
more readily identified. Different trends are seen at different locations for different periods,
consistent with regional changes in pollutant emission, especially NOX. Logan et al. (1999)
analyzed the composite record of mid-tropospheric O3 abundance from the nine-station network.
A plot of data is shown in Figure 10-7. While no clear trend appeared for 1980 through 1996,
the average level for second half of this record (about 57 ppb) is clearly greater than for the first
half (about 53 ppb). The trend may be consistent with changes in regional NOX emission rates
occurring due to pollution reduction efforts in developed countries being offset by increasing
emissions in rapidly growing economies in Asia. The measurements shown in Figure 10-7 are
concentrations observed at 4 to 7 km (mid-troposphere). Fewer locations have measured
changes in the concentrations of O3 as a function of altitude. Fewer still are locations that have
collected and maintained data records prior to 1970. The absence of historical data on the
vertical distribution of O3 adds to the difficulty in estimating historical atmospheric burdens and
trends in O3-related climate forcing.
The IPCC (200la) surveyed the results of published chemistry transport model (CTM)
modelling studies (see Table 10-2) that estimated the global average increases in total column O3
since the preindustrial era. Model estimates ranged from +7 to +12 DU. On the basis of these
estimates, available measurements, and other analyses, the IPCC estimated that total column O3
has increased by 9 DU, with a 67% confidence range of+6 to +13 DU. In some of the modelling
studies, emissions scenarios predicted a further increase in column O3 due to growing emissions
of O3 precursors. Fusco and Logan (2003) stated that, according to the models, increased NOX
emissions from fossil fuel combustion have had the greatest effect on O3 in the lower
troposphere since the 1970s. In addition, increases in background CH4 have also contributed as
10-51
-------
£1
a.
70
65
60
50
45
40 -
400 to 630 hPa
36° N to 59° N
1970
1975
1980
1985
1990
1995
Figure 10-7. Mid-tropospheric O3 abundance (ppb) in northern midlatitudes
(36 °N-59 °N) for the years 1970 to 1996. Observations between 630 and
400 hPa are averaged from nine ozonesonde stations (four in North America,
three in Europe, two in Japan), following the data analysis of Logan et al.
(1999). Values are derived from the residuals of the trend fit, with the trend
added back to allow for discontinuities in the instruments. Monthly data
(points) are shown with a smoothed 12-month mean (line).
Source: IPCC(2001a).
Table 10-2. CTM Studies Assessed by the IPCC for its Estimate of the Change in
Global and Total Column O3 Since the Preindustrial Era
Estimated Change in
Column O3 in DU
7.9
8.9
8.4
9.5
12
7.2
8.7
9.6
8
Model Used
GFDL
MOZART-1
NCAR/2D
GFDL-scaled
Harvard/GISS
ECHAM4
UKMO
UIO
MOGUNTIA
References
Haywoodetal. (1998)
Hauglustaine et al. (1998)
Kiehletal. (1999)
Levy etal. (1997)
Mickley et al. (1999)
Roelofs etal. (1997)
Stevenson et al. (2000)
Berntsen and Isaksen (1999)
Van Dorland etal. (1997)
Source: IPCC (200la)
10-52
-------
much as 20% to the increase in tropospheric O3 in the northern latitudes. Given its longer
atmospheric residence time, CH4 can serve as an O3 precursor at much longer distances from its
source than can other O3 precursors and, therefore, has a more uniform effect across the globe.
Fusco and Logan (2003) found a 10% increase in O3 concentrations year-round over
Canada, Europe, and Japan and a 20% increase for Japan and Europe during spring and summer.
It was expected—but not found—that O3 concentrations over Japan would increase in line with
emissions from China. The authors suggested that convective activity over Asia is stronger than
that seen over other industrialized areas of the globe. Such a meteorological characteristic would
result in an injection of pollutants into the free troposphere, allowing long-range transport to
North America. Their suggestion is supported by evidence of increasing background
concentrations within the United States (Fiore et al., 1998).
NARSTO (2000), in its assessment of the available information on O3 pollution in North
America, stated that no single pattern for trends in O3 over North America can be found in the
available monitoring data. In the United States, the average 1-h concentration at surface
monitoring sites decreased by 15% between 1986 and 1996, with most of the observed
declines occurring in urban and urban-influenced locations. The largest declines occurred in
Los Angeles, New York, and Chicago. Free tropospheric O3 concentrations appeared to hold
steady, or only declined slightly, from the 1980s forward.
In preparation for the IPCC TAR (IPCC, 200la), research groups engaged in modeling
global-scale tropospheric chemistry were invited to participate in a model intercomparison
focusing on potential changes in the oxidative capacity of the atmosphere (OxComp), which
included O3 concentrations for the 2000 to 2100 period. Participating groups employed the
IPCC A2p scenario, i.e., including the highest emissions levels, to calculate the geospatial
distribution of O3 up to 20 km. The predicted spatial distributions of O3 where quite variable,
but the predictions for total column O3 density change fell within 9 DU of each other (11.4 to
20.5 DU) in all cases and that was considered to be encouraging by the authors. Fusco and
Logan (2003) pointed out that several unresolved issues may limit the ability of models in
reproducing observed trends in tropospheric O3. Among these are the use of different
meteorological inputs, photochemical reaction schemes, and predicted cloud cover—each
contributing to different predictions in O3 production and loss rates.
10-53
-------
10.3.4.4 The Sensitivity of Ozone-Related Forcing Surface to Albedo
The characteristics of the surface underlying an O3 enhancement play a role in the O3
forcing effect. Highly reflective surfaces, such as light-colored deserts, sea ice and snow, scatter
solar short wave (UV and visible) radiation. UV and visible radiation can then be absorbed,
transformed into long-wave radiation, and reemitted in part back to the surface by tropospheric
O3. Studies by two groups, Hauglustaine et al. (1998) and Mickley et al. (1999), have shown
that industrial pollution that has been transported to the Arctic induces a high, regional O3-
related forcing due to the highly reflective underlying ice and snow surface.
Liao et al. (2004) calculated that the maximum change in O3-related top-of-the-atmosphere
forcing occurs over high albedo regions in high northern latitudes. Surface forcing was
calculated to be greatest at high northern latitudes as well as at dust-source regions, which also
tend to have high surface albedos.
10.3.4.5 The Altitude Dependence of Forcing by Tropospheric Ozone
Altitude plays an important role in the forcing effect of tropospheric O3 (IPCC, 1992;
Gauss et al., 2003). The efficiency of IR absorption by O3 depends upon its temperature-at
atmospheric temperatures that are low, relative to the Earth's surface, it has the capacity to
absorb more IR radiation than O3 at temperatures close to that of the surface. While this
temperature effect applies to all GHGs, it introduces a complication for estimating forcing by O3,
because O3 is not homogeneously mixed within the troposphere. Ozone forcing estimates must
account for these difficult-to-predict vertical inhomogeneities. However, as part of the OxCornp
modeling intercomparison, Gauss et al. (2003) found that the overall forcing by O3 can be
calculated within reasonable uncertainty simply on the basis of total column density.
10.3.4.6 Co-occurrence of Ozone with Particulate Matter
Analysis of the 2001 data from the AQS database showed infrequent co-occurrence of
high PM25 and O3 concentrations (Chapter 3 of this document). For those cases when O3
production is high, in combination with high PM concentrations, there is a suggestion in the
literature that heterogeneous chemistry on PM surfaces may lead to reduced gas-phase O3.
Liao et al. (2004) modeled heterogeneous chemistry taking place on PM, and found a significant
10-54
-------
titration of O3 and its NOX precursors. The importance of this titration effect remains an open
question, given the difficulty in obtaining in situ measurements to validate model calculations.
Liao et al. (2004) also estimated that forcing by BC, mineral dust, and organic carbon
aerosols substantially offsets forcing by tropospheric O3, yielding an overall negative globally
averaged forcing at both the top of the atmosphere and at the Earth's surface. However, such
estimates neglect the regional aspects of forcing by these individual pollutants. Elevated
concentrations of these very different types of pollutants often appear independently of the
others, such as with biomass burning plumes, Saharan dust, and organic aerosols associated
with biogenic terpene emissions by forests. It is unlikely that a global average of the forcing
effects of these individual pollutants will adequately capture their impacts on climate at the
regional scale.
10.3.5 Estimated Forcing by Tropospheric Ozone
10.3.5.1 Direct Climate Forcing Due to Ozone
The inhomogeneous distribution of O3 within the troposphere, coupled with the large
uncertainly in the global O3 budget, significantly complicates the matter of estimating the global
average direct forcing due to O3. The IPCC Third Assessment Risk (200la) lists the results of
several modeling studies that estimated the annual change in the relative forcing by O3 from
preindustrial times. It was noted that the differences among the estimates were most likely due
to differences in predicted O3 chemistry, including the emissions inventories used and the
chemical process and transport mechanisms incorporated into the models, rather than by factors
relating to radiative transfer. The IPCC intercomparison of the models and their results indicate
that the uncertainties in estimated forcings due to O3 have decreased since the IPCC Second
Assessment Report (1996).
The O3-related forcings estimated by studies considered by the IPCC (200la) are listed in
Table 10-3. Ten of the listed estimates are based on global CTM calculations. One study was
constrained by a climatology derived from observations. Given the differences in calculated
total column O3 among the models, a normalized forcing (WnT2 per Dob son Unit of
tropospheric O3 change) is listed in addition to the absolute forcing (WnT2) estimated by each
model. Both clear sky (cloud-free) and total sky (including clouds) forcing estimates are listed.
10-55
-------
Table 10-3. Tropospheric O3 Change (O3) in Dobson Units (DU) Since Preindustrial
Times, and the Accompanying Net (SW plus LW) Radiative Forcings (Wm~2),
After Accounting for Stratospheric Temperature Adjustment (using the Fixed
Dynamical Heating Method). Estimates are Taken From the Published Literature.
Normalized Forcings (norm.) Refer to Radiative Forcing per O3 Change (Wm~2 per DU)
Estimated Global Average Forcing Due to TronosDheric Ozone
Clear sky conditions
Reference
Berntsen et al. (1997) - [Reading model]
Stevenson etal. (1998)
Berntsen et al. (1997) - [Oslo model]
Haywood etal. (1998)
Kiehl etal. (1999)
Berntsen et al. (2000)
Brasseur etal. (1998)
Van Dorland etal. (1997)
Roelofs etal. (1997)
Lelieveld and Dentener (2000)
Hauglustaine et al. (1998)
Mean
AO3
7.600
8.700
7.600
7.900
8.400
9.600
—
8.070
7.200
—
8.940
8.224
Net
0.310
0.391
0.390
0.380
0.379
0.428
0.443
0.397
—
0.511
0.403
Net (norm.)
0.041
0.045
0.051
0.048
0.045
0.045
—
0.055
0.055
—
0.057
0.049
Total sky conditions
Net
0.280
0.289
0.310
0.310
0.320
0.342
0.370
0.380
0.404
0.420
0.426
0.343
Net (norm.)
0.037
0.033
0.041
0.039
0.038
0.036
—
0.047
0.056
—
0.048
0.042
Source: IPCC(2001a).
The largest O3-related forcings coincide with the strongest sources of tropospheric O3,
which the models predict occur in the northern midlatitude regions (40° to 50° N) and reach as
much as 1 WnT2 in the summer as well as in the tropics, and are related to biomass burning.
In general, the estimates are comparable in magnitude and show similarity in geographic
distribution. For total sky conditions, the range in globally and annually averaged
tropospheric O3 forcing from all of these models is from 0.28 to 0.43 WnT2, while the
normalized forcing is 0.033 to 0.056 WnT2 per DU. As expected, they are opposite in sign to
the forcing estimated for sulfate aerosols, which scatter radiation. The range in normalized
forcings emphasizes the differences in assumptions used by the different models. The
10-56
-------
tropospheric O3 forcing constrained by the observational climatology is 0.32 WnT2 for globally
averaged, total sky conditions. As shown in Figure 10-6, the IPCC (200la) concluded that
0.35 ± 0.15 WnT2represents the most likely value for annually and globally-averaged forcing
by tropospheric O3. Not included here is the study by Hansen and Sato (2001), that evaluated
forcing by O3 with corrections made to the assumptions concerning pre-industrial O3
concentrations and the effects of natural O3 precursors, especially NOx generated by lightning.
Hansen and Sato (2001) concluded that a more likely range for globally averaged forcing by O3
is 0.4 to 0.8 WnT2, with 0.5 WnT2 as their best estimate. Since the publication of the IPCC
Third Assessment Report (200 la), new studies have been published that illuminate some of the
regionally-relevant details associated with direct forcing by O3 (Mickley et al., 2004a; Liao et al.
2004). Forcing by O3, due to its capacity for absorbing solar UV as well as solar and terrestrial
IR radiation, can be divided into "shortwave" forcing and "long-wave" forcing. These forcings
occur under different conditions. Shortwave forcing can only take place during daytime, while
long-wave forcing can occur at all hours as a function of the diurnally varying concentration of
atmospheric O3. As noted, earlier, unlike CO2, the absorption spectrum for O3 is distinct from
that of water vapor—meaning that O3 will absorb and reemit long-wave radiation without
interference by water under high humidity conditions. Mickley et al. (2004a) reported that,
according to their modeling study, surface temperature response to the predicted O3
enhancement since the preindustrial period differs greatly from that of the CO2 response, and
that this difference can only be explained by the geographical distribution and absorption
properties of O3. Liao et al. (2004) estimated globally averaged top-of-the-atmosphere separate
short- and long-wave forcings to O3 to be 0.21 W/m2 and 0.32 W/m2, respectively.
10.3.5.2 Indirect Forcing Due to Ozone
Ozone has an indirect climate forcing effect due to its role in the oxidative removal of
other reactive GHGs, including CH4, hydrofluorocarbons (FIFCs), and other reactive NMHCs.
The primary actor in this effect is a second generation product of the photolysis of O3, the
hydroxyl (OH) radical. Hydroxyl radicals are produced by way of a pair of reactions that start
with the photodissociation of O3 by solar UV.
10-57
-------
O3 + hv -» O(1D) + O2 (10-4)
+ H2O ->> OH + OH (10-5)
Reactions with OH are the primary removal mechanism for CH4 and NMHCs as well as the
pollutants NOX and CO. Methane and CO are in especially high abundance in the global
atmosphere. OH is estimated to react with these two gases within 1 second of its formation.
In addition to CH4, NOX, CO, and the NMHCs, OH concentrations are controlled by local
concentrations of H2O (i.e., humidity) and the intensity of solar UV. Different atmospheric
concentrations of the required precursors suggest that preindustrial OH concentrations were
likely to have been different from present-day concentrations, but there is no consensus on the
magnitude of this difference. Observations of global atmospheric concentrations of
methylchloroform (CH3CC13), a well-mixed tropospheric species that also reacts with OH, have
been used to estimate OH abundances. Independent studies have shown overlapping trends for
the period 1978 to 1994, but none of the trends are outside the given uncertainty ranges (0.5 ±
0.6%/year) (Prinn et al., 1995; Krol et al., 1998). The IPCC (2001a) reported a range of +5% to
-20% for predicted changes in global OH abundances.
Given the difficulty in estimating global OH abundances in the past, present, and future,
estimates of indirect forcing due to O3 have been difficult to obtain and are highly uncertain.
10.3.5.3 Predictions for Future Climate Forcing by Anthropogenic Ozone
Not surprisingly, CTM modeling attempts to predict future precursor emissions and
resulting O3 abundances indicate that the largest future O3-related forcings will be related to
population growth and economic development in Asia (Van Borland et al., 1997; Brasseur et al.,
1998). The results of these modeling studies predict that the globally averaged total radiative
forcing due to O3 from preindustrial times 0.66 WnT2 will rise to 0.63 WnT2 by 2050. Chalita
et al. (1996) predicted a change in the globally averaged radiative forcing from preindustrial
times to 2050 of 0.43 WnT2. Stevenson et al. (1998) predicted an O3-related forcing of 0.48
WnT2 in 2100. Applying the SRES scenario projecting the highest emissions out to the year
2100 (IPCC, 2000), the OxComp model intercomparison study yielded a projected O3-induced
forcing ranging from 0.40 to 0.78 WnT2. The authors concluded, given their prediction for
10-58
-------
forcing by well-mixed GHGs of 5.6 WrrT2, that O3 would remain an important contributor to
overall anthropogenic forcing well into the future. However, all of these predictions must be
viewed with caution given the considerable uncertainties associated with the long-term economic
activity projections required for such estimates.
10.3.6 The Impact of a Warming Climate on Atmospheric Ozone
Concentrations
Evaluation of the potential impact of climate warming on U.S. air quality is currently
underway. Initial modeling results reported by Mickley et al. (2004b) suggest that reduced
cyclone frequency in a warmer climate will lead to increases in the severity of summertime
pollution episodes. Cyclonic weather patterns are known to play an important role in ventilating
pollution away from the surface. They note that compelling evidence is accumulating that the
frequency of these cyclones has decreased over the past few decades. An early study by Jacob
et al. (1993) found a correlation between O3 concentrations and temperature was due to the effect
of temperature on atmospheric chemistry, biogenic emissions, and stagnation.
10.3.7 Conclusion
The general consensus within the atmospheric sciences community, as represented by the
United Nations Intergovernmental Panel on Climate Change (IPCC), is that human activities
have a discernable effect on the Earth's climate. However, quantifying the extent of human-
induced forcing on climate requires detailed information about human-induced change on the
components of the Earth System that govern climate. Tropospheric O3 is a well-known GHG,
but information regarding its historical trends in concentration, its current and future
atmospheric burden, and other critical details needed for estimating its direct and indirect
forcing effects on the climate system are highly uncertain.
The IPCC has estimated that the globally averaged forcing due to O3 is approximately
0.35 ± 0.15 WnT2, with an updated value of 0.5 ± 0.2 WnT2 provided by Hansen and Sato
(2001). However, the most important role of O3 in climate is likely to be at the regional scale,
adjacent to the sources of its chemical precursors. This expectation is consistent with satellite
observations of high regional scale column O3 densities near large urban areas and large-scale
biomass burning activity. Modeling studies evaluated by the IPCC have estimated that regional-
10-59
-------
scale forcing due to O3 can approach 1 WnT2, or as much as 40% of the globally averaged
forcing due to the well-mixed GHGs. While more certain estimates of the overall importance of
global-scale forcing due to tropospheric O3 await further advances in monitoring and chemical
transport modeling, the overall body of scientific evidence suggests that high concentrations
of O3 on the regional scale could have a discernable influence on climate, leading to surface
temperature and hydrological cycle changes. Confirming this effect requires improvement in
regional-scale modeling—an activity that is currently underway.
10-60
-------
REFERENCES
Ahrens, D. C. (1994) Meteorology today: an introduction to weather, climate and the environment. Minneapolis,
MN: West Publishing Co., p. 17.
Airey, D. K.; Wong, J. C.; Fleming, R. A.; Meldrum, L. R. (1997) An estimate of the UV-B exposure for outdoor
workers during a south-east Queensland summer. Health Phys. 72: 544-549.
Alados-Arboledas, L.; Alados, L; Foyo-Moreno, L; Olmo, F. J.; Alcantara, A. (2003) The influence of clouds on
surface UV erythema! irradiance. Atmos. Res. 66: 273-290.
Ancellet, G.; Beekmann, M. (1997) Evidence for changes in the ozone concentrations in the free troposphere over
southern France from 1976 to 1995. Atmos. Environ. 31: 2835-2851.
Arrhenius, S. (1896) On the influence of carbonic acid in the air upon the temperature of the ground. Philos. Mag.
41:237-276.
Autier, P.; Dore, J. F.; Negrier, S.; Lienard, D.; Panizzon, R.; Lejeune, F. J.; Guggisberg, D.; Eggermont, A. M.
(1999) Sunscreen use and duration of sun exposure: a double-blind, randomized trial. J. Natl. Cancer Inst.
91: 1304-1309.
Autier, P.; Dore, J.-F.; Reis, A. C.; Grivegnee, A.; Ollivaud, L.; Truchetet, F.; Chamoun, E.; Rotmensz, N.; Severi,
G.; Cesarini, J. P.; EORTC Melanoma Co-operative Group. (2000) Sunscreen use and intentional exposure to
ultraviolet A and B radiation: a double blind randomized trial using personal dosimeters. Br. J. Cancer
83: 1243-1248.
Ayala, M. N.; Michael, R.; Soderberg, P. G. (2000) Influence of exposure time forUV radiation-induced cataract.
Invest. Ophthalmol. Visual Sci. 41: 3539-3543.
Balis, D. S.; Zerefos, C. S.; Kourtidis, K.; Bais, A. F.; Hofzumahaus, A.; Kraus, A.; Schmitt, R.; Blumthaler, M.;
Gobbi, G. P. (2002) Measurements and modeling of photolysis rates during the photochemical activity and
ultraviolet radiation (PAUR) II campaign. J. Geophys. Res. (Atmos.) 107(D18): 10.1029/2000JD000136.
Barnard, W. F.; Saxena, V. K.; Wenny, B. N.; DeLuisi, J. J. (2003) Daily surface UV exposure and its relationship
to surface pollutant measurements. J. Air Waste Manage. Assoc. 53: 237-245.
Barnett, T. P.; Pierce, D. W.; AchutaRao, K. M.; Gleckler, P. J.; Santer, B. D.; Gregory, J. M.; Washington, W. M.
(2005) Penetration of human-induced warming into the world's oceans. Science (Washington, DC, U.S.)
309: 284-287.
Bellows, C. F.; Belafsky, P.; Fortgang, I. S.; Beech, D. J. (2001) Melanoma in African-Americans: trends in
biological behavior and clinical characteristics over two decades. J. Surg. Oncol. 78: 10-16.
Bernard, S. M.; McGeehin, M. G.; Patz, J. A., eds. (2001a) Health: the potential consequences of climate variability
and change. Environ. Health Perspect. 109(suppl. 2): 175-233.
Bernard, S. M.; Samet, J. M.; Grambsch, A.; Ebi, K. L.; Romieu, I. (2001b) The potential impacts of climate
variability and change on air pollution-related health effects in the United States. Environ. Health Perspect.
109(suppl. 2): 199-209.
Berntsen, T. K.; Isaksen, I. S. A. (1999) Effects of lightning and convection on changes in tropospheric ozone due to
NOX emissions from aircraft. Tellus 5IB: 766-788.
Berntsen, T. K.; Isaksen, I. S. A.; Myhre, G.; Fuglestvedt, J. S.; Stordal, F.; Larsen, T. A.; Freckleton, R. S.; Shine,
K. P. (1997) Effects of anthropogenic emissions on tropospheric ozone and its radiative forcing. J. Geophys.
Res. [Atmos.] 102: 28,101-28,126.
Berntsen, T. K.; Myhre, G.; Stordal, F.; Isaksen, I. S. A. (2000) Time evolution of tropospheric ozone and its
radiative forcing. J. Geophys. Res. [Atmos.] 105: 8915-8930.
Berret, J.; Liardet, S.; Scaletta, C.; Panizzon, R.; Hohlfeld, P.; Applegate, L. A. (2002) Use of sunscreens in families
living in Switzerland. Dermatology 204: 202-208.
Black, W. C.; Goldhahn, R. T., Jr.; Wiggins, C. (1987) Melanoma within a southwestern Hispanic population. Arch.
Dermatol. 123: 1331-1334.
Brasseur, G. P.; Kiehl, J. T.; Mtiller, J.-F.; Schneider, T.; Granier, C.; Tie, X.-X.; Hauglustaine, D. (1998) Past and
future changes in global tropospheric ozone: impact on radiative forcing. Geophys. Res. Lett. 25: 3807-3810.
Breuer-McHam, J.; Simpson, E.; Dougherty, L; Bonkobara, M.; Ariizumi, K.; Lewis, D. E.; Dawson, D. B.; Duvic,
M.; Cruz, P. D., Jr. (2001) Activation of HIV in human skin by ultraviolet B radiation and its inhibition by
NFKB blocking agents. Photochem. Photobiol. 74: 805-810.
Bruhl, C.; Crutzen, P. J. (1989) On the disproportionate role of tropospheric ozone as a filter against solar UV-B
radiation. Geophys. Res. Lett. 16: 703-706.
10-61
-------
Caforio, A. L. P.; Fortina, A. B.; piaserico, S.; Alaibac, M; Tona, F.; Feltrin, G.; Pompei, E.; Testolin, L.; Gambino,
A.; Volta, S. D.; Thiene, G.; Casarotto, D.; Peserico, A. (2000) Skin cancer in heart transplant recipients: risk
factor analysis and relevance of immunosuppressive therapy. Circulation 102(suppl. Ill): 111-222 -111-221.
Cantorna, M. T. (2000) Vitamin D and autoimmunity: is vitamin D status an environmental factor affecting
autoimmune disease prevalence? Proc. Soc. Exp. Biol. Med. 223: 230-233.
Chalita, S.; Hauglustaine, D. A.; Le Treut, H.; Miiller, J. F.; Penkett, S., eds. (1996) Radiative forcing due to
increased tropospheric ozone concentrations. Atmos. Environ. 30: 1641-1646.
Clydesdale, G. J.; Dandie, G. W.; Muller, H. K. (2001) Ultraviolet light induced injury: immunological and
inflammatory effects. Immunol. Cell Biol. 79: 547-568.
Cockell, C. S. (2001) A photobiological history of Earth. In: Cockell, C. S.; Blaustein, A. R., eds. Ecosystems,
evolution, and ultraviolet radiation. New York, NY: Springer-Verlag; pp. 1-35.
Cutchis, P. (1974) Stratospheric ozone depletion and solar ultraviolet radiation on Earth. Science (Washington, DC)
184: 13-19.
Davis, K. J.; Cokkinides, V. E.; Weinstock, M. A.; O'Connell, M. C.; Wingo, P. A. (2002) Summer sunburn and sun
exposure among US youths ages 11 to 18: national prevalence and associated factors. Pediatrics 110: 27-35.
De Fabo, E. C.; Noonan, F. P.; Fears, T.; Merlino, G. (2004) Ultraviolet B but not ultraviolet A radiation initiates
melanoma. Cancer Res. 64: 6372-6376.
De Gruijl, F. R. (1995) Action spectrum for photocarcinogenesis. Recent Results Cancer Res. 139: 21-30.
De Gruijl, F. R. (1999) Skin cancer and solar UV radiation. Eur. J. Cancer 35: 2003-2009.
De Gruijl, F. R. (2002) Photocarcinogenesis: UVA vs. UVB radiation. Skin Pharmacol. Appl. SkinPhysiol.
15: 316-320.
De Gruijl, F. R.; Henricus, J. C. M.; Sterenborg, H.; Forbes, P. D.; Davies, R. E.; Cole, C.; Kelfkens, G.; Van
Weelden, H.; Slaper, H.; VanderLeun, J. C. (1993) Wavelength dependence of skin cancer induction by
ultraviolet irradiation of albino hairless mice. Cancer Res. 53: 53-60.
Diepgen, T. L.; Mahler, V (2002) The epidemiology of skin cancer. Br. J. Dermatol. 146(suppl. 61): 1-6.
Diffey, B. L. (1990) Human exposure to ultraviolet radiation. Semin. Dermatol. 9: 2-10.
Diffey, B. L.; Gibson, C. J.; Haylock, R.; McKinlay, A. F. (1996) Outdoor ultraviolet exposure of children and
adolescents. Br. J. Dermatol. 134: 1030-1034.
Evans, R. D.; Kopf, A. W.; Lew, R. A.; Rigel, D. S.; Bart, R. S.; Friedman, R. J.; Rivers, J. K. (1988) Risk factors
for the development of malignant melanoma—I: review of case-control studies. J. Dermatol. Surg. Oncol.
14: 393-408.
Fears, T. R.; Bird, C. C.; Guerry, D., IV; Sagebiel, R. W.; Gail, M. H.; Elder, D. E.; Halpern, A.; Holly, E. A.;
Hartge, P.; Tucker M. A. (2002) Average midrange ultraviolet radiation flux and time outdoors predict
melanoma risk. Cancer Res. 62: 3992-3996.
Federal Register. (2003) National ambient air quality standards for ozone: final response to remand; final rule. F. R.
(January 6) 68: 614-645.
Finlayson-Pitts, B. J.; Pitts, J. N., Jr. (1986) Atmospheric chemistry: fundamentals and experimental techniques.
New York, NY: John Wiley & Sons.
Fiore, A. M.; Jacob, D. J.; Logan, J. A.; Yin, J. H. (1998) Long-term trends in ground level ozone over the
contiguous United States, 1980-1995. J. Geophys. Res. (Atmos.) 103: 1471-1480.
Fiore, A.; Jacob, D. J.; Liu, H.; Yantosca, R. M.; Fairlie, T. D.; Li, Q. (2003) Variability in surface ozone
background over the United States: implications for air quality policy. J. Geophys. Res. (Atmos.)
108(D24): 10.1029/2003 JD003855.
Fishman, J.; Wozniak, A. E.; Creilson, J. K. (2003) Global distribution of tropospheric ozone from satellite
measurements using the empirically corrected tropospheric ozone residual technique: identification of the
regional aspects of air pollution. Atmos. Chem. Phys. 3: 893-907.
Frederick, J. E.; Koob, A. E.; Weatherhead, E. C. (1993) Empirical studies of tropospheric transmission in the
ultraviolet: roadband measurements. J. Appl. Meteorol. 32: 1883-1892.
Freedman, D. M.; Dosemeci, M.; McGlynn, K. (2002) Sunlight and mortality from breast, ovarian, colon, prostate,
and non-melanoma skin cancer: a composite death certificate based case-control study. Occup. Environ. Med.
59: 257-262.
Fusco, A. C.; Logan, J. A. (2003) Analysis of 1970-1995 trends in tropospheric ozone at Northern Hemisphere
midlatitudes with the GEOS-CHEM model. J. Geophys. Res. (Atmos.) 108(D15): 10.1029/2002JD002742.
Gallagher, R. P.; Hill, G. B.; Bajdik, C. D.; Coldman, A. J.; Fincham, S.; McLean, D. I.; Threlfall, W. J. (1995a)
Sunlight exposure, pigmentation factors and risk of nonmelanocytic skin cancer. II. Squamous cell carcinoma.
Arch. Dermatol. 131: 164-169.
10-62
-------
Gallagher, R. P.; Hill, G. B.; Bajdik, C. D.; Fincham, S.; Goldman, A. J.; McLean, D. I.; Threlfall, W. J. (1995b)
Sunlight exposure, pigmentary factors, and risk of nonmelanocytic skin cancer. I. Basal cell carcinoma. Arch.
Dermatol. 131: 157-163.
Garland, F. C.; Garland, C. F.; Gorham, E. D.; Young, J. F. (1990) Geographic variation in breast cancer mortality in
the United States: a hypothesis involving exposure to solar radiation. Prev. Med. 19: 614-622.
Garssen, J.; Van Loveren, H. (2001) Effects of ultraviolet exposure on the immune system. Crit. Rev. Immunol.
21: 359-397.
Gauss, M; Myhre, G.; Pitari, G.; Prather, M. J.; Isaksen, I. S. A.; Berntsen, T. K.; Brasseur, G. P.; Dentener, F. J.;
Derwent, R. G.; Hauglustaine, D. A.; Horowitz, L. W.; Jacob, D. J.; Johnson, M.; Law, K. S.; Mickley, L. J.;
Mtiller, J.-F.; Plantevin, P. H.; Pyle, J. A.; Rogers, H. L.; Stevenson, D. S.; Sundet, J. K.; Van Weele, M.;
Wild, O. (2003) Radiative forcing in the 21st century due to ozone changes in the troposphere and the lower
stratosphere. J. Geophys. Res. (Atmos.) 108(D9): 10.1029/2002JD002624.
Gies, P.; Wright, J. (2003) Measured solar ultraviolet radiation exposures of outdoor workers in Queensland in the
building and construction industry. Photochem. Photobiol. 68: 78-83.
Gies, P.; Roy, C.; Toomey, S.; MacLennan, R.; Watson, M. (1998) Solar UVR exposures of primary school children
at three locations in Queensland. Photochem. Photobiol. 68: 78-83.
Giovannucci, E. (2005) The epidemiology of vitamin D and cancer incidence and mortality: a review (United
States). Cancer Causes Control 16: 83-95.
Gloster, H. M., Jr.; Brodland, D. G. (1996) The epidemiology of skin cancer. Dermatol. Surg. 22: 217-226.
Godar, D. E. (2001) UV doses of American children and adolescents. Photochem. Photobiol. 74: 787-793.
Godar, D. E.; Wengraitis, S. P.; Shreffler, J.; Sliney, D. H. (2001) UV doses of Americans. Photochem. Photobiol.
73:621-629.
Godar, D. E.; Urbach, F.; Gasparro, F. P.; Van Der Leun, J. C. (2003) UV doses of young adults. Photochem.
Photobiol. 77: 453-457.
Gonzalez, C. R.; Veefkind, J. P.; De Leeuw, G. (2000) Aerosol optical depth over Europe in August 1997 derived
from ATSR-2 data. Geophys. Res. Lett. 27: 955-958.
Gorham, E. D.; Garland, F. C.; Garland, C. F. (1990) Sunlight and breast cancer incidence in the USSR. Int. J.
Epidemiol. 19: 820-824.
Granstein, R. D.; Matsui, M. S. (2004) UV radiation-induced immunosuppression and skin cancer. Cutis 74
(suppl. 5): 4-9.
Grant, W. B. (2002a) An ecologic study of dietary and solar ultraviolet-B links to breast carcinoma mortality rates.
Cancer 94: 272-281.
Grant, W. B. (2002b) An estimate of premature cancer mortality in the U.S. due to inadequate doses of solar
ultraviolet-B radiation. Cancer 94: 1867-1875.
Grant, W. B.; Garland, C. F. (2004) A critical review of studies on vitamin D in relation to colorectal cancer. Nutr.
Cancer 48: 115-123.
Great Lakes Regional Assessment Group (GLRAG). (2000) Preparing for a changing climate: the potential
consequences of climate variability and change. Great Lakes overview. Washington, DC: U.S. Environmental
Protection Agency; Office of Research and Development; U.S. Global Change Research Program (USGCRP)
and Ann Arbor, MI: University of Michigan; Atmospheric, Oceanic and Space Sciences Department.
Available: http://www.geo.msu.edu/glra/assessment/assessment.html [17 April 2002].
Green, A.; Battistutta, D.; Hart, V.; Leslie, D.; Weedon, D. (1996) Skin cancer in a subtropical Australian
population: incidence and lack of association with occupation. The Nambour Study Group. Am. J. Epidemiol.
144: 1034-1040.
Green, A.; McCredie, M.; MacKie, R.; Giles, G.; Young, P.; Morton, C.; Jackman, L.; Thursfield, V (1999) A
case-control study of melanomas of the soles and palms (Australia and Scotland). Cancer Causes Control
10: 21-25.
Hall, H. L; Miller, D. R.; Rogers, J. D.; Bewerse, B. J. (1999) Update on the incidence and mortality from melanoma
in the United States. J. Am. Acad. Dermatol. 40: 35-42.
Halliday, G. M.; Byrne, S. N.; Kuchel, J. M.; Poon, T. S.; Barnetson, R. S. (2004) The suppression of immunity by
ultraviolet radiation: UVA, nitric oxide and DNA damage. Photochem. Photobiol. Sci. 3: 736-740.
Hanchette, C. L.; Schwartz, G. G. (1992) Geographic patterns of prostate cancer mortality. Cancer 70: 2861-2869.
Hansen, J. E.; Sato, M. (2001) Trends of measured climate forcing agents. Proc. Natl. Acad. Sci. 98: 14778-14783.
Harrison, G. I.; Young, A. R. (2002) Ultraviolet radiation-induced erythema in human skin. Methods (San Diego,
CA,U.S.)28: 14-19.
10-63
-------
Hartge, P.; Devesa, S. S.; Grauman, D.; Fears, T. R.; Fraumeni, J. F., Jr. (1996) Non-Hodgkin's lymphoma and
sunlight. J. Natl. Cancer Inst. 88: 298-300.
Hartmann, D. L. (1994) Global physical climatology. New York, NY: Academic Press. (International geophysics:
v. 56).
Hauglustaine, D. A.; Brasseur, G. P.; Walters, S.; Rasch, P. J.; Mtiller, J.-F.; Emmons, L. K.; Carroll, M. A. (1998)
MOZART, a global chemical transport model for ozone and related chemical tracers 2. model results and
evaluation. J. Geophys. Res. (Atmos.) 103: 28291-28335.
Haywood, J. M.; Schwarzkopf, M. D.; Ramaswamy, V. (1998) Estimates of radiative forcing due to modeled
increases in tropospheric ozone. J. Geophys. Res. [Atmos.] 103: 16,999-17,007.
Henriksen, M.; Na, R.; Agren, M. S.; Wulf, H. C. (2004) Minimal erythema dose after umltiple UV exposures
depends on pre-exposure skin pigmentation. Photodermatol. Photoimmunol. Photomed. 20: 163-169.
Herman, J. R.; Bhartia, P. K.; Torres, O.; Hsu, C.; Seftor, C.; Celarier, E. (1997) Global distribution of
UV-absorbing aerosols from Nimbus 7/TOMS data. J. Geophys. Res. [Atmos.] 102: 16,911-16,922.
Hildesheim, J.; Fornace, A. J., Jr. (2004) The dark side of light: the damaging effects of UV rays and the protective
efforts of MAP kinase signaling in the epidermis. DNA Repair 3: 567-580.
Hockwin, O.; Kojima, M.; Sakamoto, Y.; Wegener, A.; Shui, Y. B.; Sasaki, K. (1999) UV damage to the eye lens:
further results from animal model studies: a review. J. Epidemiol. 9(suppl. 6): S39-S47.
Holick, M. F. (2004) Sunlight and vitamin D for bone health and prevention of autoimmune diseases, cancers, and
cardiovascular disease. Am. J. Clin. Nutr. 80(suppl.): 1678S-1688S.
Houghton, J. T.; Jenkins, G. J.; Ephraums, J. J., eds. (1990) Climate change: the IPCC scientific assessment.
Cambridge, MA: Cambridge University Press; p. 55.
Hu, S.; Ma, F.; Collado-Mesa, F.; Kirsner, R. S. (2004) UV radiation, latitude, and melanoma in US Hispanics and
blacks. Arch. Dermatol. 140: 819-824.
Hughes, A. M.; Armstrong, B. K.; Vajdic, C. M.; Turner, J.; Grulich, A. E.; Fritschi, L.; Milliken, S.; Kaldor, J.;
Benke, G.; Kricker, A. (2004) Sun exposure may protect against non-Hodgkin lymphoma: a case-control
study. Int. J. Cancer 112: 865-871.
Ichihashi, M.; Ueda, M.; Budiyanto, A.; Bito, T.; Oka, M.; Fukunaga, M.; Tsuru, K.; Horikawa, T. (2003)
UV-induced skin damage. Toxicology 189: 21-39.
Intergovernmental Panel on Climate Change (IPCC). (1992) Climate change 1992 - the supplementary report to the
IPCC scientific assessment. Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change. (1996) Climate change 1995: the science of climate change. Summary
for policymakers and technical summary of the working group I report. Presented at: the fifth session of the
IPCC Working Group I; November; Madrid, Spain. Geneva, Switzerland: World Meteorological Association.
Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
of vulnerability, a special report of IPCC Working Group II. Watson, R. T.; Zinyowera, M. C.; Moss, R. H.,
eds. Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (2000) Emissions scenarios: a special report of working group
III of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom: Cambridge University
Press. Available: http://www.grida.no/climate/ipcc/emission/ [26 August, 2005].
Intergovernmental Panel on Climate Change (IPCC). (2001a) Climate change 2001: the scientific basis. Contribution
of working group I to the third assessment report of the Intergovernmental Panel on Climate Change.
Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (200 Ib) Climate change 2001: impacts, adaptation, and
vulnerability. Contribution of working group II to the third assessment report of the Intergovernmental Panel
on Climate Change. Cambridge, United Kingdom: Cambridge University Press.
International Agency for Research on Cancer (IARC). (1992) Solar and ultraviolet radiation. Lyon, France:
International Agency for Research on Cancer. (IARC monographs on the evaluation of carcinogenic risks to
humans: v. 55).
International Commission on Non-Ionizing Radiation Protection (ICNIRP). (2004) Guidelines on limits of exposure
to ultraviolet radiation of wavelengths between 180 nm and 400 nm (incoherent optical radiation). Health
Phys. 87: 171-186.
Jacob, D. J.; Logan, J. A.; Yevich, R. M.; Gardner, G. M.; Spivakovsky, C. M.; Wofsy, S. C.; Munger, J. W.;
Sillman, S.; Prather, M. J.; Rodgers, M. O.; Westberg, H.; Zimmerman, P. R. (1993) Simulation of
summertime ozone over North America. J. Geophys. Res. [Atmos.] 98: 14,797-14,816.
Javitt, J. C.; Taylor, H. R. (1994-95) Cataract and latitude. Doc. Ophthalmol. 88: 307-325.
10-64
-------
Jemal, A.; Tiwari, R. C.; Murray, T.; Ghafoor, A.; Samuels, A.; Ward, E.; Feuer, E. I; Thun, M. J. (2004) Cancer
statistics, 2004. CA-Cancer J. Clin. 54: 8-29.
John, E. M.; Schwartz, G. G.; Dreon, D. M.; Koo, J. (1999) Vitamin D and breast cancer risk: the NHANES I
Epidemiologic Follow-up Study, 1971-1975 to 1992. Cancer Epidemiol. Biomarkers Prev. 8: 399-406.
Kaidbey, K. H.; Kligman, A. M. (1981) Cumulative effects from repeated exposures to ultraviolet radiation. J.
Invest. Dermatol. 76: 352-355.
Kerr, J. B.; McElroy, C. T. (1993) Evidence for large upward trends of ultraviolet-B radiation linked to ozone
depletion. Science (Washington, DC) 262: 1032-1034.
Kiehl, J. T.; Schneider, T. L.; Portmann, R. W.; Solomon, S. (1999) Climate forcing due to tropospheric and
stratospheric ozone. J. Geophys. Res. [Atmos.] 104: 31,239-31,254.
Kimlin, M. G.; Schallhorn, K. A. (2004) Estimations of the human Vitamin D' UV exposure in the USA. Photochem.
Photobiol. Sci. 3: 1067-1070.
Kimlin, M. G.; Parisi, A. V.; Wong, J. C. (1998a) Quantification of personal solar UV exposure of outdoor workers,
indoor workers and adolescents at two locations in Southeast Queensland. Photodermatol. Photoimmunol.
Photomed. 14:7-11.
Kimlin, M. G.; Wong, J. C.; Parisi, A. V. (1998b) Simultaneous comparison of the personal UV exposure of two
human groups at different altitudes. Health Phys. 74: 429-434.
Kollias, N.; Stamatas, G. N.; Youn, J. I. (2001) Suppression of UVB-induced cutaneous erythema by a previous
UVB exposure. Photochem. Photobiol. 74: 471-476.
Kraemer, K. H. (1997) Sunlight and skin cancer: another link revealed. Proc. Natl. Acad. Sci. U. S. A. 94: 11-14.
Kricker, A.; Armstrong, B. K.; English, D. R. (1994) Sun exposure and non-melanocytic skin cancer. Cancer Causes
Control 5: 367-392.
Krol, M.; Van Leeuwen, P. J.; Lelieveld, J. (1998) Global OH trend inferred from methylchloroform measurements.
J. Geophys. Res. [Atmos.] 103: 10,697-10,711.
Krotkov, N. A.; Bhartia, P. K.; Herman, J. R.; Fioletov, V; Kerr, J. (1998) Satellite estimation of spectral surface
UVirradiance in the presence of tropospheric aerosols: 1. Cloud-free case. J. Geophys. Res. (Atmos.)
103: 8779-8793.
Lefkowitz, E. S.; Garland, C. F. (1994) Sunlight, vitamin D, and ovarian cancer mortality rates in U.S. women. Int.
J. Epidemiol. 23: 1133-1136.
Lehman, J.; Swinton, K.; Bortnick, S.; Hamilton, C.; Baldridge, E.; Ender, B.; Cox, B. (2004) Spatio-temporal
characterization of tropospheric ozone across the eastern United States. Atmos. Environ. 38: 4357-4369.
Lelieveld, J.; Dentener, F. J. (2000) What controls tropospheric ozone? J. Geophys. Res. [Atmos.] 105: 3531-3551.
Lens, M. B.; Dawes, M. (2004) Global perspectives of contemporary epidemiological trends of cutaneous malignant
melanoma. Br. J. Dermatol. 150: 179-185.
Levine, R. D.; Bernstein, R. B. (1987) Molecular reaction dynamics and chemical reactivity. New York, NY: Oxford
University Press.
Levitus, S.; Antonov, J.; Boyer, T. (2005) Warming of the world ocean, 1955-2003. Geophys. Res. Lett.
32(L02604): 10.1029/2004GL021592.
Levy, H., II; Kasibhatla, P. S.; Moxim, W. J.; Klonecki, A. A.; Hirsch, A. I.; Oltmans, S. J.; Chameides, W. L.
(1997) The global impact of human activity on tropospheric ozone. Geophys. Res. Lett. 24: 791-794.
Liang, X.-Z.; Pan, J. P.; Zhu, J. H.; Kunkel, K. E.; Dai, A.; Meehl, J. (2005) Regional climate model downscaling of
U.S. climate. J. Geophys. Res.: submitted.
Liao, H.; Seinfeld, J. H.; Adams, P. J.; Mickley, L. J. (2004) Global radiative forcing of coupled tropospheric ozone
and aerosols in a unified general circulation model. J. Geophys. Res. [Atmos.]
109(D16): 10.1029/2003 JD004456.
Lin, C.-Y. C.; Jacob, D. J.; Fiore, A. M. (2001) Trends in exceedances of the ozone air quality standard in the
continental United States, 1980-1998. Atmos. Environ. 35: 3217-3228.
Lindelof, B.; Sigurgeirsson, B.; Gabel, H.; Stern, R. S. (2000) Incidence of skin cancer in 5356 patients following
organ transplantation. Br. J. Dermatol. 143: 513-519.
Liu, S. C.; McKeen, S. A.; Madronich, S. (1991) Effect of anthropogenic aerosols on biologically active ultraviolet
radiation. Geophys. Res. Lett. 18: 2265-2268.
Logan, J. A. (1999) An analysis of ozonesonde data for the troposphere: recommendations for testing 3-D models
and development of a gridded climatology for tropospheric ozone. J. Geophys. Res. 104: 16,115-16,149.
Longstreth, J. D.; De Gruijl; Kripke, M. L.; Takizawa, Y.; VanderLeun, J. C. (1995) Effects of increased solar
ultraviolet radiation on human health. Ambio 24: 153-165.
10-65
-------
Longstreth, I; De Gruijl, F. R.; Kripke, M. L.; Abseck, S.; Arnold, F.; Slaper, H. I.; Velders, G.; Takizawa, Y.; Van
Der Leun, J. C. (1998) Health risks. J. Photochem. Photobiol. B 46: 20-39.
Lovato, C. Y.; Shoveller, J. A.; Peters, L.; Rivers, J. K. (1998a) Canadian National Survey on Sun Exposure &
Protective Behaviours: parent's reports on children. Cancer Prev. Control. 2: 123-128.
Lovato, C. Y.; Shoveller, J. A.; Peters, L.; Rivers, J. K. (1998b) Canadian National Survey on Sun Exposure &
Protective Behaviours: youth at leisure. Cancer Prev. Control 2: 117-122.
Lutter, R.; Wolz, C. (1997) UV-B screening by tropospheric ozone: implications for the national ambient air quality
standard. Environ. Sci. Technol. 31: 142A-146A.
Madronich, S.; De Gruijl, F. (1993) Skin cancer and UV radiation. Nature (London, UK) 366(6450): 23.
Madronich, S.; De Gruijl, F. (1994) Reply to "Kricker, A.; Armstrong, B. K.; McMichael, A. J. (1994) Skin cancer
and ultraviolet. Nature (London, UK) 368: 594." Nature (London, UK) 368(6472): 594.
Madronich, S.; McKenzie, R. L.; Caldwell, M. M.; Bjorn, L. O. (1995) Changes in ultraviolet radiation reaching the
Earth's surface. Ambio 24: 143-152.
Matsumura, Y.; Ananthaswamy, H. N. (2004) Toxic effects of ultraviolet radiation on the skin. Toxicol. Appl.
Pharmacol. 195: 298-308.
Matthijsen, J.; Slaper, H.; Reinen, H. A. J. M.; Velders, G. J. M. (2000) Reduction of solar UV by clouds: a
comparison between satellite-derived cloud effects and ground-based radiation measurements. J. Geophys.
Res. [Atmos.] 105: 5069-5080.
McCarty, C. A.; Taylor, H. R. (2002) A review of the epidemiologic evidence linking ultraviolet radiation and
cataracts. Dev. Ophthalmol. 35: 21-31.
McCarty, C. A.; Nanjan, M. B.; Taylor, H. R. (2000) Attributable risk estimates for cataract to prioritize medical and
public health action. Invest. Ophthalmol. Vis. Sci. 41: 3720-3725.
McKenzie, R. L.; Seckmeyer, G.; Bais, A. F.; Kerr, J. B.; Madronich, S. (2001) Satellite retrievals of erythemal UV
dose compared with ground-based measurements at northern and southern midlatitudes. J. Geophys. Res.
[Atmos.] 106: 24,051-24,062.
McKenzie, R.; Smale, D.; Bodeker, G.; Claude, H. (2003) Ozone profile differences between Europe and New
Zealand: effects on surface UV irradiance and its estimation from satellite sensors. J. Geophys. Res. [Atmos.]
108(D6): 10.1029/2002JD002770.
McKinlay, A. F.; Diffey, B. L. (1987) A reference action spectrum for ultraviolet induced erythema in human skin.
In: Passchier, W. F.; Bosnijakovic, B. F. M., eds. Human exposure to ultraviolet radiation: risks and
regulations. Amsterdam, The Netherlands: Elsevier Science Publishers; pp. 83-87.
Merriam, J. C.; Lofgren, S.; Michael, R.; Soderberg, P.; Dillon, J.; Zheng, L.; Ayala, M. (2000) An action spectrum
for UV-B radiation and the rat lens. Invest. Ophthalmol. Vis. Sci. 41: 2642-2647.
Mickley, L. J.; Murti, P. P.; Jacob, D. J.; Logan, J. A.; Koch, D. M.; Rind, D. (1999) Radiative forcing from
tropospheric ozone calculated with a unified chemistry-climate model. J. Geophys. Res. [Atmos.]
104: 30,153-30,172.
Mickley, L. J.; Jacob, D. J.; Rind, D. (2001) Uncertainty in preindustrial abundance of tropospheric ozone:
implications for radiative forcing calculations. J. Geophys. Res. [Atmos.] 106: 3389-3399.
Mickley, L. J.; Jacob, D. J.; Field, B. D.; Rind, D. (2004a) Climate response to the increase in tropospheric ozone
since preindustrial times: a comparison between ozone and equivalent CO2 forcings. J. Geophys. Res.
[Atmos.] 109(05): 10.1029/2003 JD003653.
Mickley, L. J.; Jacob, D. J.; Field, B. D.; Rind, D. (2004b) Effects of future climate change on regional air pollution
episodes in the United States. Geophys. Res. Lett. 31(L24103): 10.1029/2004GL021216.
Mid-Atlantic Regional Assessment Team (MARAT). (2000) Preparing for a changing climate: the potential
consequences of climate variability and change. Mid-Atlantic overview. Washington, DC: U.S.
Environmental Protection Agency; Office of Research and Development; U.S. Global Change Research
Program (USGCRP) and University Park, PA: Pennsylvania State University. Available:
http://www.usgcrp.gov/usgcrp/nacc/midatlantic.htm [17 April 2002].
Minis, F. M., Ill; Frederick, J. E. (1994) Cumulus clouds and UV-B. Nature (London) 371: 291.
Moehrle, M. (2001) Ultraviolet exposure in the Ironman triathlon. Med. Sci. Sports Exercise 33: 1385-1386.
Moehrle, M.; Heinrich, L.; Schmid, A.; Garbe, C. (2000) Extreme UV exposure of professional cyclists.
Dermatology 201: 44-45.
Moise, A. F.; Btittner, P. G.; Harrison, S. L. (1999) Sun exposure at school. Photochem. Photobiol. 70: 269-274.
10-66
-------
National Assessment Synthesis Team (NAST). (2000) Climate change impacts on the United States: the potential
consequences of climate variability and change. Overview. Washington, DC: U.S. Global Change Research
Program. Available at: http://www.usgcrp.gov/usgcrp/Library/nationalassessment/overview.htm (17 January
2003).
National Research Council. (2005) Radiative forcing of climate change: expanding the concept and addressing
uncertainties. Washington, DC: The National Academies Press.
New England Regional Assessment Group (NERAG). (2001) Preparing for a changing climate: the potential
consequences of climate variability and change. New England regional overview. Washington, DC: U.S.
Environmental Protection Agency; Office of Research and Development; U.S. Global Change Research
Program (USGCRP) and Durham, NH: University of New Hampshire; Institute for the Study of Earth,
Oceans, and Space. Available: http://www.necci.sr.unh.edu/2001-NERA-report.html [17 April 2002].
Newchurch, M. I; Ayoub, M. A.; Oltmans, S.; Johnson, B.; Schmidlin, F. J. (2003) Vertical distribution of ozone at
four sites in the United States. J. Geophys. Res. 108(D1): 10.1029/2002JD002059.
Nole, G.; Johnson, A. W. (2004) An analysis of cumulative lifetime solar ultraviolet radiation exposure and the
benefits of daily sun protection. Dermatol. Ther. 17(suppl. 1): 57-62.
Noonan, F. P.; Muller, H. K.; Fears, T. R.; Kusewitt, D. F.; Johnson, T. M.; De Fabo, E. D. (2003) Mice with
genetically determined high susceptibility to ultraviolet (UV)-induced immunosuppression show enhanced
UV carcinogenesis. J. Invest. Dermatol. 121: 1175-1181.
North American Research Strategy for Tropospheric Ozone (NARSTO) Synthesis Team. (2000) An assessment of
tropospheric ozone pollution: a North American perspective. Palo Alto, CA: Electric Power Research
Institute.
Norval, M.; Garssen, J.; VanLoveren, H.; El-Ghorr, A. A. (1999) UV-induced changes in the immune response to
microbial infections in human subjects and animal models. J. Epidemiol. 9(suppl. 6): S84-S92.
Okada, S.; Weatherhead, E.; Targoff, I. N.; Wesley, R.; Miller, F. W.; International Myositis Collaborative Study
Group. (2003) Global surface ultraviolet radiation intensity may modulate the clinical and imunologic
expression of autoimmune muscle disease. Arthritis Rheum. 48: 2285-2293.
Oriowo, O. M.; Cullen, A. P.; Chou, B. R.; Sivak, J. G. (2001) Action spectrum and recovery for in vitro
UV-induced cataract using whole lenses. Invest. Ophthalmol. Visual Sci. 42: 2596-2602.
Palancar, G. G.; Toselli, B. M. (2004) Effects of meteorology on the annual and interannual cycle of UV-B and total
radiation in Cordoba City, Argentina. Atmos. Environ. 38: 1073-1082.
Parisi, A. V.; Willey, A.; Kimlin, M. G.; Wong, J. C. (1999) Penetration of solar erythema! UV radiation in the
shade of two common Australian trees. Health Phys. 76: 682-686.
Patz, J. A.; Engelberg, D.; Last, J. (2000a) The effects of changing weather on public health. Annu. Rev. Public
Health 21: 271-307.
Patz, J. A.; McGeehin, M. A.; Bernard, S. M.; Ebi, K. L.; Epstein, P. R.; Grambsch, A.; Gubler, D. J.; Reiter, P.;
Romieu, I.; Rose, J. B.; Samet, J. M.; Trtanj, J. (2000b) The potential health impacts of climate variability and
change for the United States: executive summary of the report of the health sector of the U.S. national
assessment. Environ. Health Perspect. 108: 367-376.
Persson, A. E.; Edstrom, D. W.; Backvall, H.; Lundeberg, J.; Ponten, F.; Ros, A. M.; Williams, C. (2002) The
mutagenic effect of ultraviolet-A1 on human skin demonstrated by sequencing the p53 gene in single
keratinocytes. Photodermatol. Photoimmunol. Photomed. 18: 287-293.
Pfister, H. (2003) Human papillomavirus and skin cancer. J. Natl. Cancer Inst. Monogr. 31: 52-56.
Pitts, D. G. (1993) Ocular effects of radiant energy. In: Pitts, D. G.; Kleinstein, R. N., eds. Environmental vision.
Stoneham, MA: Butterworth-Heinemann, pp. 151-220.
Ponsonby, A.-L.; McMichael, A.; Van derMei, I. (2002) Ultraviolet radiation and autoimmune disease: insights
from epidemiological research. Toxicology 181/182: 71-78.
Prinn, R. G.; Weiss, R. F.; Miller, B. R.; Huang, J.; Alyea, F. N.; Cunnold, D. M.; Fraser, P. J.; Hartley, D. E.;
Simmonds, P. G. (1995) Atmospheric trends and lifetime of CH3CC13 and global OH concentrations. Science
(Washington, DC) 269: 187-192.
Rigel, D. S.; Rigel, E. G.; Rigel, A. C. (1999) Effects of altitude and latitude on ambient UVB radiation. J. Am.
Acad. Dermatol. 40: 114-116.
Rivers, J. K. (2004) Is there more than one road to melanoma? Lancet 363: 728-730.
Robinson, E. S.; Hill, R. H., Jr.; Kripke, M. L.; Setlow, R. B. (2000) TheMonodelphis melanoma model: initial
report on large ultraviolet A exposures of suckling young. Photochem. Photobiol. 71: 743-746.
10-67
-------
Roelofs, G. I; Lelieveld, I; Van Borland, R. (1997) A three-dimensional chemistry/general circulation model
simulation of anthropogenically derived ozone in the troposphere and its radiative climate forcing. J.
Geophys. Res. [Atmos.] 102: 23,389-23,401.
Rosenthal, F. S.; Phoon, C.; Bakalian, A. E.; Taylor, H. R. (1988) The ocular dose of ultraviolet radiation to outdoor
workers. Invest. Ophthalmol. Visual Sci. 29: 649-656.
Rozema, J.; Manetas, Y.; Bjorn, L. O., eds. (2001) Responses of plants to UV-B radiation. Dordrecht, The
Netherlands: Kluwer Academic. (Advances in vegetation science: v. 1)].
Rtinger, T. M; Moller, K.; Jung, T.; Dekant, B. (2000) DNA damage formation, DNA repair, and survival after
exposure of DNA repair-proficient and nucleotide excision repair-deficient human lymphoblasts to UVA1
and UVB. Int. J. Radiat. Biol. 76: 789-797.
Salby, M. L. (1996) Fundamentals of atmospheric physics. New York, NY: Academic Press. (Dmowska, R.; Holton,
J. R., eds. International geophysics series: v. 61).
Sarasin, A. (1999) The molecular pathways of ultraviolet-induced carcinogenesis. Mutat. Res. 428: 5-10.
Sarkar, A. K. (2004) An evaluation of UV protection imparted by cotton fabrics dyed with natural colorants. BMC
Dermatol. 4: 15.
Schafer, J. S.; Saxena, V K.; Wenny, B. N.; Barnard, W.; DeLuisi, J. J. (1996) Observed influence of clouds on
ultraviolet-B radiation. Geophys. Res. Lett. 23: 2625-2628.
Schenker, M. B.; Orenstein, M. R.; Samuels, S. J. (2002) Use of protective equipment among California farmers.
Am. J. Ind. Med. 42: 455-464.
Selgrade, M. K.; Repacholi, M. H.; Koren, H. S. (1997) Ultraviolet radiation-induced immune modulation: potential
consequences for infectious, allergic, and autoimmune disease. Environ. Health Perspect. 105: 332-334.
Selgrade, M. K.; Smith, M. V.; Oberhelman-Bragg, L. J.; LeVee, G. J.; Koren, H. S.; Cooper, K. D. (2001) Dose
response for UV-induced immune suppression in people of color: differences based on erythema! reactivity
rather than skin pigmentation. Photochem. Photobiol. 74: 88-95.
Setlow, R. B.; Grist, E.; Thompson, K.; Woodhead, A. D. (1993) Wavelengths effective in induction of malignant
melanoma. Proc. Natl. Acad. Sci. U. S. A. 90: 6666.6670.
Shindell, D. T.; Faluvegi, G. (2002) An exploration of ozone changes and their radiative forcing prior to the
chlorofluorocarbon era. Atmos. Chem. Phys. 2: 363-374.
Shoveller, J. A.; Lovato, C. Y.; Peters, L.; Rivers, J. K. (1998) Canadian National Survey on Sun Exposure &
Protective Behaviours: adults at leisure. Cancer Prev. Control 2: 111-116.
Slaper, H.; Velders, G. J. M.; Daniel, J. S.; De Gruijl, F. R.; VanderLeun, J. C. (1996) Estimates of ozone depletion
and skin cancer incidence to examine the Vienna Convention achievements. Nature (London, U.K.) 384:
256-258.
Solanki, S. K.; Fligge, M. (2000) Reconstruction of past solar irradiance. Space Sci. Rev. 94: 127-138.
Stern, R. S.; Liebman, E. J.; Vakeva, L. (1998) Oral psoralen and ultraviolet-A light (PUVA) treatment of psoriasis
and persistent risk of nonmelanoma skin cancer. PUVA Follow-up Study. J. Natl. Cancer Inst. 90: 1278-1284.
Stevenson, D. S.; Johnson, C. E.; Collins, W. J.; Derwent, R. G.; Shine, K. P.; Edwards, J. M. (1998) Evolution of
tropospheric ozone radiative forcing. Geophys. Res. Lett. 25: 3819-3822.
Stevenson, D. S.; Johnson, C. E.; Collins, W. J.; Derwent, R. G.; Edwards, J. M. (2000) Future estimates of
tropospheric ozone radiative forcing and methane turnover - the impact of climate change. Geophys. Res.
Lett. 27: 2073-2076.
Studzinski, G. P.; Moore, D. C. (1995) Sunlight—can it prevent as well as cause cancer? Cancer Res.
55: 4014-4022.
Swelter, S. M. (2003) Dermatological perspectives of malignant melanoma. Surg. Clin. N. Am. 83: 77-95.
Taylor, H. R.; West, S. K.; Rosenthal, F. S.; Munoz, B.; Newland, H. S.; Abbey, H.; Emmett, E. A. (1988) Effect of
ultraviolet radiation on cataract formation. N. Engl. J. Med. 319: 1429-1433.
Thieden, E.; Agren, M. S.; Wulf, H. C. (2001) Solar UVR exposures of indoor workers in a working and a holiday
period assessed by personal dosimeters and sun exposure diaries. Photodermatol. Photoimmunol. Photomed.
17: 249-255.
Thieden, E.; Philipsen, P. A.; Heydenreich, J.; Wulf, H. C. (2004a) UV radiation exposure related to age, sex,
occupation, and sun behavior based on time-stamped personal dosimeter readings. Arch. Dermatol.
140: 197-203.
Thieden, E.; Philipsen, P. A.; Sandby-Meller, J.; Heydenreich, J.; Wulf, H. C. (2004b) Proportion of lifetime UV
dose received by children, teenagers and adults based on time-stamped personal dosimetry. J. Invest.
Dermatol. 123: 1147-1150.
10-68
-------
Tomescu, D.; Kavanagh, G.; Ha, T.; Campbell, H.; Melton, D. W. (2001) Nucleotide excision repair geneXPD
polymorphisms and genetic predisposition to melanoma. Carcinogenesis 22: 403-408.
Trepte, S.; Winkler, P. (2004) Reconstruction of erythema! UV irradiance and dose at Hohenpeissenberg
(1968-2001) considering trends of total ozone, cloudiness and turbidity. Theor. Appl. Climatol. 77: 159-171.
Ullrich, S. E. (2005) Mechanisms underlying UV-induced immune suppression. Mutat. Res. 571: 185-205.
U.S. Environmental Protection Agency. (1987) Assessing the risks of trace gases that can modify the stratosphere,
v. 1-8. Washington, DC: Office of Air and Radiation; report no. EPA 400/1-87/001A-H.
U.S. Environmental Protection Agency. (2004) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
http://cfpub.epa.gov/ncea/ [9 November, 2004].
United Nations Environment Programme (UNEP). (1986) Report of the international conference on the assessment
of the role of carbon dioxide and of other greenhouse gases in climate variations and associated impacts;
October 1985; Villach, Austria. Geneva, Switzerland: World Meteorological Organization; WMO no. 661.
Urbach, F. (1997) Ultraviolet radiation and skin cancer of humans. J. Photochem. Photobiol. B 40: 3-7.
Van der Leun, J. C.; De Gruijl, F. R. (1993) Influences of ozone depletion on human and animal health. In: Tevini,
M., ed. UV-B radiation and ozone depletion: Effects on humans, animals, plants, microorganisms, and
materials. Ann Arbor: Lewis Publisher; pp. 95-123.
Van Borland, R.; Dentener, F. J.; Lelieveld, J. (1997) Radiative forcing due to tropospheric ozone and sulfate
aerosols. J. Geophys. Res. [Atmos.] 102: 28,079-28,100.
Veiered, M. B.; Weiderpass, E.; Thorn, M.; Hansson, J.; Lund, E.; Armstrong, B.; Adami, H.-O. (2003) A
prospective study of pigmentation, sun exposure, and risk of cutaneous malignant melanoma in women. J.
Natl. Cancer Inst. 95: 1530-1538.
Vingarzan, R. (2004) A review of surface ozone background levels and trends. Atmos. Environ. 38: 3431-3442.
Vishvakarman, D.; Wong, J. C.; Boreham, B. W. (2001) Annual occupational exposure to ultraviolet radiation in
central Queensland. Health Phys. 81: 536-544.
Vitasa, B. C.; Taylor, H. R.; Strickland, P. T.; Rosenthal, F. S.; West, S.; Abbey, H.; Ng, S. K.; Munoz, B.; Emmett,
E. A. (1990) Association on nonmelanoma skin cancer and actinic keratosis with cumulative solar ultraviolet
exposure in Maryland watermen. Cancer 65: 2811-2817.
Waterston, K.; Nay smith, L.; Rees. J. L. (2004) Physiological variation in the erythemal response to ultraviolet
radiation and photoadaptation. J. Invest. Dermatol. 123: 958-964.
Wayne, R. P. (2000) Chemistry of Atmospheres: an introduction to the chemistry of the atmospheres of Earth, the
planets, and their satellites. 3rd ed. New York, NY: Oxford University Press, Inc.
Wei, Q.; Lee, J. E.; Gershenwald, J. E.; Ross, M. I.; Mansfield, P. F.; Strom, S. S.; Wang, L. E.; Guo, Z.; Qiao, Y.;
Amos, C. I.; Spitz, M. R.; Duvic, M. (2003) Repair of UV light-induced DNA damage and risk of cutaneous
malignant melanoma. J. Natl. Cancer Inst. 95: 308-315.
Wen, G. Y.; Frederick, J. E. (1995) The effects of horizontally extended clouds on backscattered ultraviolet sunlight.
J. Geophys. Res. [Atmos.] 100: 16387-16393.
Wendisch, M.; Mayer, B. (2003) Vertical distribution of spectral solar irradiance in the cloudless sky: a case study.
Geophys. Res. Lett. 30: 10.1029/2002GL016529.
Westerdahl, J.; Ingvar, C.; Masback, A.; Jonsson, N.; Olsson, H. (2000) Risk of cutaneous malignant melanoma in
relation to use of sunbeds: further evidence forUV-A carcinogenicity. Br. J. Cancer 82: 1593-1599.
Whiteman, D. C.; Valery, P.; McWhirter, W.; Green, A. C. (1997) Risk factors for childhood malanoma in
Queensland, Australia. Int. J. Cancer 70: 26-31.
Whiteman, D. C.; Whiteman, C. A.; Green, A. C. (2001) Childhood sun exposure as a risk factor for melanoma: a
systematic review of epidemiologic studies. Cancer Causes Control 12: 69-82.
Whiteman, D. C.; Watt, P.; Purdie, D. M.; Hughes, M. C.; Hayward, N. K.; Green, A. C. (2003) Melanocytic nevi,
solar keratoses, and divergent pathways to cutaneous melanoma. J. Natl. Cancer Inst.
Wong, J. C.; Airey, D. K.; Fleming, R. A. (1996) Annual reduction of solar UV exposure to the facial aea of outdoor
workers in southeast Queensland by wearing a hat. Photodermatol. Photoimmunol. Photomed. 12: 131-135.
World Meteorological Organization (WMO). (1988) Developing policies for responding to climatic change: a
summary of the discussions and recommendations of workshops; September-October 1987; Villach, Austria;
and November 1987; Bellagio, Austria. Geneva, Switzerland: World Meteorological Organization; report no.
WMO/TD; no. 225. [World Climate Impact Programme series report no. WCIP-1].
World Meteorological Organization (WMO). (1999) Scientific assessment of ozone depletion: 1998. Geneva,
Switzerland: World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
10-69
-------
World Meteorological Organization (WMO). (2002) Scientific assessment of ozone depletion: 2002. Geneva,
Switzerland: United Nations Environment Programme. [Global Ozone and Monitoring Project; report no. 47].
Xenopoulos, M. A.; Schindler, D. W. (2001) Physical factors determining ultraviolet radiation flux into ecosystems.
In: Cockell, C.; Blaustein, A. R. Ecosystems, evolution, and ultraviolet radiation. New York, NY: Springer;
pp. 36-62.
Yarnal, B.; Kalkstein, L. S.; Scheraga, J. D. (2000) Mid-Atlantic regional assessment of climate change impacts.
Clim. Res. (CR Special 7) 14: 153-269.
Yoshikawa, T.; Rae, V.; Bruins-Slot, W.; Van den Berg, J. W.; Taylor, J. R.; Streilein, J. W. (1990) Susceptibility to
effects of UVB radiation on induction of contact hypersensitivity as a risk factor for skin cancer in humans. J.
Invest. Dermatol. 95: 530-536.
Zerefos, C. S.; Kourtidis, K. A.; Melas, D.; Balis, D.; Zanis, P.; Katsaros, L.; Mantis, H. T.; Repapis, C.; Isaksen, L;
Sundet, J.; Herman, J.; Bhartia, P. K.; Calpini, B. (2002) Photochemical activity and solar ultraviolet radiation
(PAUR) modulation factors: an overview of the project. J. Geophys. Res. [Atmos.] 107(D18):
10/1029/2000JD000134.
10-70
-------
11. OZONE EFFECTS ON MAN-MADE MATERIALS
Ozone (O3) and other photochemical oxidants react with many economically important
man-made materials, decreasing their useful life and aesthetic appearance. Some materials
known to be damaged by O3 include elastomers, fibers, dyes, and paints. This chapter provides a
brief discussion of O3 effects on man-made materials, including denoting of damage mechanisms
and, where possible, concentration-response relationships. Much of what is known about O3
effects on man-made materials is derived from research conducted in the 1970's, 1980's, and
early 1990's, with very little new research on the subject having been conducted since then.
Since only very limited new information has been published on O3 effects on materials, this
chapter mainly summarizes key information assessed in the previous 1996 Air Quality Criteria
Document for Ozone and other photochemical oxidants (1996 O3 AQCD) (U.S. Environmental
Protection Agency, 1996) and provides detailed discussion of the very limited new information
that has become available since then. In the ensuing sections, discussion is focused on O3 effects
on: elastomers (Sect 11.1); textiles and fabrics (11.2); dyes, pigments, and inks (11.3); artist's
pigments (11.4); and surface coatings (11.5). Evaluation of the relevance and economic
importance of O3 materials damage information, as it affects productivity or cultural resources
(such as museums), is beyond the scope of this chapter. The reader is referred to the previous
criteria document (1996 O3 AQCD) for more detailed discussion of the earlier studies
summarized below.
11.1 ELASTOMERS
The elastomeric compounds, natural rubber and synthetic polymers and copolymers of
butadiene, isoprene, and styrene, are particularly susceptible to even low levels of O3.
Elastomeric compounds are long chain unsaturated organic molecules. Ozone damages these
compounds by breaking the molecular chain at the carbon-carbon double bond; a chain of three
oxygen atoms is added directly across the double bond, forming a five-membered ring structure
(Mueller and Stickney, 1970). The change in structure promotes the characteristic cracking of
11-1
-------
stressed/stretched rubber called "weathering." A 5% tensile strain will produce cracks on the
surface of the rubber that increase in number with increased stress/stretching. The rate of crack
growth is dependent on the degree of stress, the type of rubber compound, concentration, time of
exposure, velocity, and temperature (Bradley and Haagen-Smit, 1951; Lake and Mente, 1992)
(Gent and McGrath, 1965). Once cracking occurs, there is further penetration, additional
cracking, and eventually mechanical weakening or stress relaxation (U.S. Environmental
Protection Agency, 1996). Razumovskii et al. (1988) demonstrated the effect of O3 on stress
relaxation of polyisoprene vulcanizates. A decrease in stress (stress relaxation) is caused
by O3-induced cracks in exposed elastomers resulting in irreversible changes in the elastomer
dimensions and decreased tensile strength.
To counteract O3 effects on elastomers, antiozonants and wax are often added to
elastomeric formulations during processing. An antiozonant is an additive used to protect a
polymer against the effects of O3-induced degradation and, hence, is used mainly in diene
rubbers. Antiozonant protection works either (a) by providing a physical barrier to O3
penetration via forming a thin surface film of an O3-resisting wax or (b) by chemically reacting
with O3 or polymer ozonolysis products, as do aromatic diamines such as p-phenylene diamine
derivatives. The antiozonant diffuses to the surface of the elastomeric material, where it reacts
with O3 faster than O3 reacts to break the molecular chain and the carbon-carbon double bond,
or the antiozonant diffuses to the surface of the material but is not reactive with O3 and serves as
a protective coating against O3 attack. The antiozonant may also serve to scavenge O3 while also
providing protective film against O3 attack (Andries et al., 1979; Lattimer et al., 1984).
Most studies of O3 effects on elastomers were designed to evaluate the effectiveness of
antiozonants in counteracting the rubber cracking produced by O3 exposure. Consequently,
many of the studies were conducted using O3 concentrations notably higher than those
typically found in the ambient air. Natural rubber strips exposed to high concentrations of O3
(20,000 ppm) under stressed conditions cracked almost instantaneously and were broken within
1 sec. When the O3 concentration was lowered (0.02 to 0.46 ppm), the time to required to
produce cracks in the exposed rubber material was increased (Bradley and Haagen-Smit, 1951).
Lake and Mente (1992) studied the effect of temperature on O3-induced elastomer cracking and
antiozonant protection on natural rubber, epoxidised natural rubber, and two acrylonitrile-
butadiene copolymers under constant strain. Temperatures ranged from -20 °C to +70 °C.
11-2
-------
The elastomers were exposed to 0.05 to 1,000 ppm O3 for 16 h. Ozone cracking decreased at
lower ambient temperatures; however, diffusing of both chemical and wax antiozonants also
slowed at the lower temperatures. Cracking was slightly increased at the higher temperatures,
but the antiozonants offered more protection.
Serrano et al. (1993) evaluated the appropriateness of using O3-induced elastomer cracking
to estimate ambient O3 concentrations. Two vulcanized natural rubber compounds were exposed
for 24 h to varying O3 concentrations under stressed conditions. Ozone concentrations were
60, 80, 90, 100, and 120 ppb for durations of 2, 4, or 6 h. The 24 h average O3 concentrations
ranged from 31 to 57.5 ppb. There was a clear relationship between the 24-h average O3
concentration and the distribution of crack length frequencies on the rubber surface. Table 11-1
gives the average 24-h O3 concentration and lengths for two vulcanized natural rubber strips.
Table 11-1. Average 24-h Ozone Concentrations Producing the Highest Frequency of
Cracks of a Certain Length in the Middle and Central Zones of the Rubber Test Strips
1% Antiozonant 4010NA #
Crack
0.
0.
0.
0.
05-
.10-
15-
.20-
Length (mm)
0
0
.10
.15
0.20
0
.40
Middle
37
45
48
>57
Zones
.5
.0
.0
.5
Central Zones
37
48
>57
>57
.5
.0
.5
.5
0.5% Antiozonant 4010NA
Middle
40
48
Zones
.0
.0
>57.5
>57
.5
Central
42
53
>57
>57
Zones
.5
.0
.5
.5
Ozone concentrations given in ppb.
Adapted from Serrano et al. (1993).
11.2 TEXTILES AND FABRICS
Ozone can damage textiles and fabrics by methods similar to those associated with
elastomers. Generally, synthetic fibers are less affected by O3 than natural fibers; however, O3
contribution to the degradation of textiles and fabrics is not considered significant (U.S.
Environmental Protection Agency, 1996). A study reported by Bogaty et al. (1952) showed
that O3 affects moistened cloth more than dry cloth. Scoured cotton duck cloth and
-------
commercially bleached cotton print cloth were exposed to 20 to 60 ppb O3 for 1,200 h (50 days).
The rate of deterioration was measured by the changes in cuprammonium fluidity values and the
fabric breaking strength. At the end of the 1,200-h exposure, there was a 20% loss in breaking
strength. Table 11-2 list the changes in cuprammonium fluidity values for both fabrics.
Table 11-2. Cuprammonium Fluidity of Moist Cotton Cloth Exposed
to 20 to 60 ppb Ozone
Duration of Exposure (h) Cuprammonium Fluidity (rhes)
Duck Cloth 0 2.6
200 2.8
680 4.0
960 6.8
1200 9.5
Bleached Print Cloth 0 8.2
200 8.7
510 9.4
650 12.0
865 12.7
1500 16.5
Adapted from Bogaty et al. (1952).
11.3 DYES, PIGMENTS, AND INKS
Ozone fading of textile dyes is diffusion-controlled; the rate of fading is controlled by the
diffusion of the dye to the fiber surface. Many textile dyes react with O3; however, the rate and
severity of the O3 attack is influenced by the chemical nature of the textile fiber and the manner
in which the dye is applied. Ozone molecules break the aromatic ring portion of the dye
molecule, oxidizing the dye (U.S. Environmental Protection Agency, 1996). In case of aromatic
azo dyes, O3 attacks the aromatic rings and electron rich nitrogen atoms (Matsui et al., 1988).
Grosjean et al. (1987; 1988a,b) proposed a mechanism for reactions of O3 with indigo,
thioindigo, and dibromoindigo, alazarin, and curcumin dyes under dark conditions.
Ozone attaches to the dye molecule at the unsaturated carbon = carbon bond. An O3 adduct is
formed (1,2,3-trioxolane), followed by scission of the carbon-carbon bond and the subsequent
11-4
-------
formation of the corresponding Criegee biradical. A similar mechanism was proposed for the
reaction of O3 with triphenylmethane colorant Basic Violet 14. Ozone attacked Basic Violet 14
at the carbon=carbon unsaturated bond and at the carbon-nitrogen unsaturated bond under dark
conditions. Other members of the group of triphenylmethane colorants with unsaturated
carbon-carbon bonds also are expected to be subject to O3 fading. Tripheylmethane colorants
that are expected to be O3-fugitive include the amino-substituted cationic dyes (Malachite Green,
Brilliant Green, Crystal Violet, Pararosaniline Chloride, Methyl Green, and others) (Grosjean
etal., 1989).
An indication that O3 caused textile dye fading was first reported by Salvin and Walker
(1955). The researchers found that the fading was primarily the result of the destruction of the
blue dye molecule. Drapes made of acetate, Arnel, and Dacron and dyed with anthraquinone
blue dye exhibited a decrease in shade that was not accompanied by the characteristic reddening
caused by NOX. Figures 11-1 and 11-2 demonstrate the effect of O3 exposure on nylon 6 yarn
colored with several blue dyes. Nylon samples inside the home were located on a wall away
from sunlight. Outside nylon samples were placed on a covered patio or under the eaves of the
house to minimize exposure to sunlight and rain. Ozone concentrations ranged from 2 to 5 ppb
outside and 0 to 2 ppb inside. The percent change in dye color was determined monthly by
extraction and analysis of the remaining dye or by instrumental measurement of the color change
(Haylock and Rush, 1978).
11.4 ARTISTS'PIGMENTS
Several artists' pigments are sensitive to fading and oxidation by O3 when exposed to
concentrations found in urban areas (Shaver et al., 1983; Drisko et al., 1985; Whitmore et al.,
1987; Whitmore and Cass, 1988; Grosjean et al., 1993). The organic pigments that are O3
fugitive include alizarin red pigments containing lakes of the poly cyclic aromatic compound
1,2-dihydroxyanthraquinone, blue-violet pigments containing substituted triphenylmethane
lakes, indigo, and yellow coloring agents containing polyfunctional, polyunsaturated compounds
such as curcumin (Grosjean et al., 1987). Because of the potential of O3 to damage works of art,
recommended limits on O3 concentrations in museums, libraries, and archives are relatively low,
ranging from 0.013 to 0.01 ppm.
11-5
-------
J F M A M J J ASOND J
Months
Figure 11-1. In-service fading of nylon 6 yarn inside house. • = C.I. Disperse Blue 3;
O = C.I. Basic Blue 22; A = C.I. Acid Blue 27; x = C.I. Disperse Blue 56;
A = C.I. Acid Blue 232.
Source: Haylock and Rush (1978).
Experimental studies demonstrate a concentration x time (C x T) relationship between O3
concentration, exposure time, and pigment fading. Cass et al. (1991) summarized some of the
earlier research on the effects of O3 on artists' pigments. In studies evaluating the effect of O3 on
organic and inorganic watercolors and traditional organic pigments, only the traditional organic
pigments showed measurable fading from O3 exposure. Of the inorganic pigments tested,
only the arsenic sulfides showed O3-related changes. The pigments were exposed to 0.3 to
0.4 ppm O3 for 3 mo in the absence of light, at 22 °C and 50% RH. The authors equated this
exposure to a C x T of 6 to 8 years inside a Los Angeles museum with air conditioning but
without a pollutant removal system.
11-6
-------
100
90
80
70
>, 60
o
(fl
*
50
Q 40
30
20
10
J FMAMJ J ASOND J
Months
Figure 11-2. In-service fading of nylon 6 yarn outside house. • = C.I. Disperse Blue 3;
O = C.I. Basic Blue 22; A = C.I. Acid Blue 27; x = C.I. Disperse Blue 56;
A = C.I. Acid Blue 232.
Source: Haylock and Rush (1978).
Whitmore and Cass (1988) studied the effect of O3 on traditional Japanese colorants. Most
of these compounds are insoluble metal salts that are stable in light and air. Suspensions or
solutions of the colorants were airbrushed on hot-pressed watercolor paper or silk cloths.
A sample of Japanese woodblock print also was included in the analysis. Samples were exposed
to 0.4 ppm O3 at 22 °C, 50% relative humidity, in the absence of light for 12 wk. Changes in
reflectance spectra were used to evaluate the effect of O3 exposure on colorant fading. Among
the colorants tested on paper, curmin, indigo, madder lake, and lac lake were the most sensitive
to O3 exposure. Gamboge was relatively insensitive to O3. The blue and green areas of the
sample from the woodblock print was very reactive due to the indigo dye O3 sensitivity. The
11-7
-------
other colorants, red, yellow, and purple, showed very little sensitivity to O3. The textiles dyes
that reacted with O3 were indigo, alone or in combination with several yellow dyes.
Ye et al. (2000) reported the rate of O3 fading of traditional Chinese plant dyes. Twelve
different colorants were applied to watercolor paper and silk and exposed to 0.4 ppm O3 at
25 °C, at 50% RH, in the absence of light for 22 wks. Dye fading was greater when the colorant
was applied to the watercolor paper compared to the silk cloth due to the darker initial depth of
the shade, the greater saturation of the colorant throughout the cloth. Turn eric, gromwell, and
violet on paper was particularly reactive. Tangerine peel was moderately reactive and sappan
wood, dalbergia wood, Chinese gall, indigo, and Chinese yellow cork tree were slightly reactive
to O3. Black tea was not reactive to O3. The colorants on silk samples showing color changes
were gromwell, sappan wood, gardenia, tummeric, and violet. Figures 11-3 and 11-4
demonstrate the color change of the various colorants on watercolor paper and silk.
Artists' pigments also have exhibited fading when exposed to a mixture of photochemical
oxidants. Grosjean et al. (1993) exposed 35 artists' pigments to a mixture of photochemical
oxidants consisting of O3, nitrogen dioxide (NO2), and peroxyacetyl nitrate (PAN) for 12 wks.
Weekly average photochemical concentrations were 200 ppb for O3, 56 ± 12 to 99 ± 24 for NO2,
and Il±3tol8±2 for PAN. All exposures were carried out at room temperature in the
absence of light. To determine the effect of humidity on pigment fading, the relative humidity
was increased from 46% after 8 weeks of exposure to 83% for a 2 week period and then returned
to 46% for the remainder of the exposure.
Table 11-3 lists the artists' pigment and degree of fading. Eleven of the pigments tested
exhibited negligible color change, 12 had small color changes, 3 had modest color changes, and
9 exhibited substantial color changes. Fading of Disperse Blue 3 and Reactive Blue 2 were
likely the result of NO2 exposure, and the fading of triphenylmethanes is consistent with
exposure to nitric acid formed under high humidity conditions. Fading of the indigos was
dominated by O3 exposure and curcumin was faded by all of the photochemicals studied.
Increasing the relative humidity resulted in a substantial color change for all of the pigments,
with the exception of curcumin and indigo.
11-8
-------
35
30
25
0)
o 20
15
o
o
o
10
zi cao (gromwe
zi ding cao (violet)
5 10 15 20
Weeks of Exposure to 0.40 ppm Ozone
very reactive colorants
e 3
o
o
o
-0- su mu (sappan wood)
-*- Jiang xiang (dalbergia wood)
-Ar hong chaye (black tea)
. -Oh wu bei zi (Chinese gall)
-O- ban Ian gen (indigo)
"V~ ju zi pi (tangerine peel)
"A- huang bai (Chinese yellow cork tree)
0 5 10 15 20
Weeks of Exposure to 0.40 ppm Ozone
moderately reactive colorants
Figure 11-3. Observed color changes for natural colorant-on-paper systems during
exposure to 0.40 ppm O3 at 25 °C ± 1 °C, 50% RH, in the absence of light.
Source: Ye et al. (2000).
11-9
-------
a)
1
D
o 2
o
O
. su mu (sappan wood)
- zi cao (gromwell)
zi ding cao (violet)
- Jiang huang (tumeric)
- huang zhi zi (gardenia)
J =
10
15
20
Weeks of Exposure to 0.40 ppm Ozone
reactive colorants
2.0
1.5
HI
-------
Table 11-3. Color Change After 12 Weeks of Exposure to a Mixture of
Photochemical Oxidants
Colorant*
Color Change (AE1 units)f
Chemical Functionality or
Chemical Composition
Acid Red 37 (17045)*
Acid Yellow 65
Alizarin Carmine
Alizarin Crimson
(Pigment Red 83)
Aurora Yellow (77199)
Basic Fuschin (42510)*
Brilliant Green (42040)*
Brown Madder
Cadmium Yellow (77199)
Carmine
Chrome Yellow (77600)*
Copper phthalocyamne
(Pigment Blue 15)
Crimson Lake
Curcumin (Natural Yellow
Disperse Blue 3
French Ultramarine Blue
Gamboge (Natural Yellow
Hooker's Green Light
Indigo (a formulation)
Indigo carmine *
Indigo (73000)*
Mauve
New Gamboge
3)
24)
11.7 ±0.5
1.8 ±0.5
1.8 ±0.2
1.4 ±0.2
0.5 ±0.1
33.4 ±3.0
20.6 ±2.1
1.7±0.1
0.4 ±0.1
1.8 ±0.2
1.7 ± 1.2
l.OiO.l
3.5 ±0.3
15.2 ±2.6
10.8 ±0.1
0.8 ±0.3
0.4 ±0.1
1.5 ±0.4
l.liO.l
14.0 ± 1.9
64.1 ±4.5
3.6 ±0.5
0.9 ±0.1
Aminophenyl-substituted azo dye,
sulfonate salt
Nitro- and phenyl-substituted azo dye,
sulfonate salt
Alizarin lake
Alizarin lake
Cadmium sulfide
Amino-substituted triphenylmethane
Amino-substituted triphenylmethane
Alizarin lake
Cadmium sulfide
Lake of cochineal (substituted
anthraquinone)
Lead chromate
Copper phthalocyanine
Alizarin lake
1,7 bis (4-hydroxy-3-methoxyphenyl)-
1,6-heptadiene-3,5-dione
Amino-substituted anthraquinone
Gambogic acid
Chlorinated copper phthalocyanine plus
ferrous beta naphthol derivative
Alizarin lake plus lampblack plus copper
plthalocyanine
5,5-indigo disulfonic acid, sodium salt
Lake of triphenyl methane (basic fuschin)
plus copper phthalyocyanine
Arylamide yellow (CI11680) plus
toluidine red
11-11
-------
Table 11-3 (cont'd). Color Change After 12 Weeks of Exposure to a
Mixture of Photochemical Oxidants
Colorant*
Color Change (A£" units)f
Chemical Functionality or
Chemical Composition
Pararosaniline base (42500)*
Payne's Grey
Permanent Magenta
Permanent Rose
Prussian Blue
Prussian Green
Purple Lake
Reactive Blue 2 (61211)*
Rose Carthane (12467)
Rose Dore
Thioindigo Violet (73312)*
Winsor Yellow (11680)
25.6 ±4.7
l.OiO.l
l.liO.l
2.0 ±0.1
0.7 ±0.2
1.6 ±0.3
0.9 ±0.2
2.3 ±0.3
14.4 ±1.1
0.8 ±0.2
2.0 ±0.2
1.9 ±1.2
0.5 ±0.2
Amino-substituted triphenylmethane
Alizarin lake plus prussian blue plus
lampblack plus ultramarine blue
Quinacridone
Quinacridone
Ferric ferrocyanide
Arylamide yellow plus prussian blue
Alizarin lake
Amino-substiruted anthraquinone,
sulfonate salt
Arylamide (Pigment Red 10) plus
xanthene (Pigment Red 90)
Quinacridone plus Yellow 3
Chlorinated thioindigo
Arylamide yellow
* On watercolor paper unless otherwise indicated. Color Index (CI) names or CI numbers are given in
parentheses.
•f Mean ± one standard deviation for triplicate samples calculated from the parameters L*, a*, and b* measured
with the color analyzer.
{ On Whatman 41 paper.
Source: Grosjeanetal. (1993).
11.5 SURFACE COATINGS
Ozone will act to erode some surface coatings (paints, varnishes, and lacquers). However,
many of the available studies on O3 degradation of surface coatings do not separate the effects
of O3 from those of other pollutants or environmental factors such as weather, humidity, and
temperature. Campbell et al. (1974) attempted to demonstrate an O3-related effect on oil house
paint, acrylic latex coating, alkyd industrial maintenance coating, urea alkyd coil coating,
and nitrocellulose/acrylic automotive paint. Painted test panels were exposed to 100 and
11-12
-------
1,000 ppb O3 in a xenon arc accelerated weathering chamber for up to 1,000 h. Using weight
loss as a measure of O3-induced erosion, the researchers concluded that all of the paints tested
suffered degradation in the presence of O3 and that the automotive finish suffered the most O3-
induced degradation. When O3 degradation was measured using scanning electron microscopy,
the oil house paint and latex coating samples showed erosion above that seen with clean air but
only at the highest exposure level. No effects were noted for the automotive paint. The other
painted surfaces were not evaluated.
Spence et al. (1975) studied the effect of air pollutants and relative humidity on oil based
house paint, acrylic latex house paint, acrylic coil coating, and vinyl coil coating under
laboratory conditions. Test panels were exposed in weathering chambers equipped with a xenon
are light for simulating sunlight to low and high levels of O3 (0.08 and 0.5 ppm), sulfur dioxide
(0.03 and 0.5 ppm), and nitrogen dioxide (0.05 and 0.5 ppm) and relative humidity (50 and
90%). Samples were exposed for a total of 1000 h. The exposure cycle consisted of 20 min of
dew and 20 min of light. The effects of the exposure on the painted surfaces were measured by
weight loss and loss in film thickness. The acrylic coil coating had the lowest erosion rate of the
surface coatings tested. However, O3 was the only pollutant that had a significant effect on the
surface erosion. Sulfur dioxide and relative humidity were significant factors in the erosion of
oil base house paints and vinyl coil coating. The findings for acrylic latex house paint were not
reported.
11.6 CONCLUSIONS
Ozone and other photochemical oxidants react with many economically important
man-made materials, decreasing their useful life and aesthetic appearance. Some materials
known to be damaged by O3 include elastomers, fibers and dyes, and paints. Most studies have
been on single compounds rather than complex materials.
The elastomeric compounds, natural rubber and synthetic polymers and copolymers of
butadiene, isoprene, and styrene, are particularly susceptible to even low concentrations of O3.
Ozone damages these compounds by breaking the molecular chain at the carbon-carbon double
bond; a chain of three oxygen atoms is added directly across the double bond. The change in
structure promotes the characteristic cracking of stressed/stretched rubber called "weathering."
11-13
-------
Tensile strain produces cracks on the surface of the rubber that increase in number with
increased stress/stretching.
The rate of crack growth is dependent on the degree of stress, the type of rubber
compound, O3 concentration, time of exposure, O3 velocity, and temperature. After initial
cracking, there is further O3 penetration, resulting in additional cracking and, eventually,
mechanical weakening or stress relaxation.
Ozone can damage textiles and fabrics by methods similar to those associated with
elastomers. Generally, synthetic fibers are less affected by O3 than natural fibers; however, O3
contribution to the degradation of textiles and fabrics is not considered significant.
Ozone fading of textile dyes is a diffusion-controlled process; the rate of fading is
controlled by the diffusion of the dye to the fiber surface. Many textile dyes react with O3. The
rate and severity of the O3 attack is influenced by the chemical nature of the textile fiber and the
manner in which the dye is applied.
Several artists' pigments are also sensitive to fading and oxidation by O3 when exposed to
concentrations found in urban areas.
11-14
-------
REFERENCES
Andries, J. C.; Rhee, C. K.; Smith, R. W.; Ross, D. B.; Diem, H. E. (1979) A surface study of ozone attack and
antiozonant protection of carbon black loaded natural rubber compounds. Rubber Chem. Technol.
52: 823-837.
Bogaty, H.; Campbell, K. S.; Appel, W. D. (1952) The oxidation of cellulose by ozone in small concentrations.
Text. Res. J. 22: 81-83.
Bradley, C. E.; Haagen-Smit, A. J. (1951) The application of rubber in the quantitative determination of ozone.
Rubber Chem. Technol. 24: 750-755.
Campbell, G. G.; Schurr, G. G.; Slawikowski, D. E.; Spence, J. W. (1974) Assessing air pollution damage to
coatings. J. Paint Technol. 46: 59-71.
Cass, G. R.; Nazaroff, W. W.; Tiller, C.; Whitmore, P. M. (1991) Protection of works of art from damage due to
atmospheric ozone. Atmos. Environ. Part A 25: 441-451.
Drisko, K.; Cass, G. R.; Whitmore, P. M.; Druzik, J. R. (1985) Fading of artists' pigments due to atmospheric ozone.
In: Vendl, A.; Pichler, B.; Weber, J.; Banik, G., eds. Wiener Berichte tiber Naturwissenschaft in der Kunst:
Doppelband 2/3. Vienna, Austria: Verlag ORAC; pp. 66-87.
Gent, A. N.; McGrath, J. E. (1965) Effect of temperature on ozone cracking of rubbers. J. Polymer Sci. A
3: 1473-1482.
Grosjean, D.; Whitmore, P. M.; De Moor, C. P.; Cass, G. R.; Druzik, J. R. (1987) Fading of alizarin and related
artists' pigments by atmospheric ozone: reaction products and mechanisms. Environ. Sci. Technol.
21:635-643.
Grosjean, D.; Whitmore, P. M.; Cass, G. R.; Druzik, J. R. (1988a) Ozone fading of natural organic colorants:
mechanisms and products of the reaction of ozone with indigos. Environ. Sci. Technol. 22: 292-298.
Grosjean, D.; Whitmore, P. M.; De Moor, C. P.; Cass, G. R.; Druzik, J. R. (1988b) Ozone fading of organic
colorants: products and mechanism of the reaction of ozone with curcumin. Environ. Sci. Technol.
22: 1357-1361.
Grosjean, D.; Whitmore, P. M.; Cass, G. R.; Druzik, J. R. (1989) Ozone fading of triphenylmethane colorants:
reaction products and mechanisms. Environ. Sci. Technol. 23: 1164-1167.
Grosjean, D.; Grosjean, E.; Williams, E. L., II. (1993) Fading of artists' colorants by a mixture of photochemical
oxidants. Atmos. Environ. Part A 27: 765-772.
Haylock, J. C.; Rush, J. L. (1978) Studies on the ozone fading of anthraquinone dyes on nylon fibers. Part II:
In-service performance. Text. Res. J. 48: 143-149.
Lake, G. J.; Mente, P. G. (1992) Ozone cracking and protection of elastomers at high and low temperatures. J. Nat.
Rubber Res. 7: 1-13.
Lattimer, R. P.; Layer, R. W.; Rhee, C. K. (1984) Mechanisms of antiozonant protection: antiozonant-rubber
reactions during ozone exposure. In: 32nd annual conference on mass spectrometry and allied topics.
May/June; San Antonio, TX. Bethesda, MD: American Society for Mass Spectrometry; pp. 357-358.
Matsui, M.; Koike, T.; Shibata, K. (1988) Ozone fading of phenolphthalein and aurin. J. Soc. Dyers Colour
104: 482-486.
Mueller, W. J.; Stickney, P. B. (1970) A survey and economic assessment of the effects of air pollution on
elastomers: final report. Columbus, OH: Battelle Memorial Institute, Columbus Laboratories; National Air
Pollution Control Administration contract no. CPA-22-69-146.
Razumovskii, S. D.; Podmasteriev, V. V.; Zaikov, G. E. (1988) Kinetics and mechanism of stress relaxation of
polyisoprene vulcanizates under ozone ageing. Polym. Degrad. Stab. 20: 37-47.
Salvin, V. S.; Walker, R. A. (1955) Service fading of disperse dyestuffs by chemical agents other than the oxides of
nitrogen. Text. Res. J. 25: 571-585.
Serrano, E.; Castro, M.; Macias, A. (1993) An improved direct method of rubber cracking analysis for estimating
24-hour ozone levels. Atmos. Environ. Part A 27: 431-442.
Shaver, C. L.; Cass, G. R.; Druzik, J. R. (1983) Ozone and the deterioration of works of art. Environ. Sci. Technol.
17: 748-752.
Spence, J. W.; Haynie, F.; Upham, J. B. (1975) Effects of gaseous pollutants on paints: a chamber study. J. Paint
Technol. 47: 57-63.
11-15
-------
U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
http://cfpub2.epa.gov/ncea/.
Whitmore, P. M; Cass, G. R. (1988) The ozone fading of traditional Japanese colorants. Stud. Conserv. 33: 29-40.
Whitmore, P. M.; Cass, G. R.; Druzik, J. R. (1987) The ozone fading of traditional natural organic colorants on
paper. J. Am. Inst. Conserv. 26: 45-58.
Ye, Y.; Salmon, L. G.; Cass, G. R. (2000) The ozone fading of traditional Chinese plant dyes. J. Am. Inst. Conserv.
39: 245-257.
11-16
-------
wEPA
Please make all necessary changes in the below label, PRESORTED STANDARD
detach copy or copy, and return to the address in the upper POSTAGE & FEES PAID
left-hand corner. EPA
United States •-, PERMIT No. G-35
Environmental Protection 'f y°uhdo not wish to ™*Jve these reports CHECK HERE |J:
detach copy or copy, and return to the address in the
Agency upper left-hand corner.
National Center for
Environmental Assessment
Research Triangle Park, NC 27711
Official Business
Penalty for Private Use
$300
EPA 600/R-05/004aF
February 2006
------- |