^^ ^^ P^ A
» EPA
     United States
          Air Quality Criteria for
          Ozone and Related
          Photochemical Oxidants
          (Second External Review
          Draft)
          Volume I of

-------
                                                  EPA/600/R-05/004aB
                                                      August 2005
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 is a second external review draft for review purposes only and does not
constitute U.S. Environmental Protection Agency policy. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
                                     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 ozone document.
                                          I-ii

-------
     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 then released in August 1986; and a supplement, Summary of
Selected 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 some newer scientific data that became available after completion of the  1986
criteria document.  Such literature was assessed in the next periodic revision of the O3 air quality
criteria document (completed in 1996) and provided scientific bases supporting the setting by
EPA in  1997 of an  8-h O3 NAAQS that is currently in force together with the 1-h O3 standard.
     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 Ozone Air Quality Criteria Document (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,  some other scientific data are also
presented and evaluated 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
or accepted for publication through 2004.
     The present Second Draft O3 AQCD (dated August 2005) is being released for public
comment and review by the Clean Air Scientific Advisory Committee (CASAC) to obtain
comments on the organization and structure of the document, the issues addressed, the
approaches employed in assessing and interpreting the newly available information on O3
exposures and effects, and the key findings and conclusions arrived at as a consequence of this
assessment.  Public comments and recommendations will be taken into account making any
appropriate further revisions to this document for incorporation into the final version of the
document to be completed and issued by February 28, 2006. Evaluations contained in the
present document will be drawn on to provide inputs to associated PM Staff Paper analyses
                                          I-iii

-------
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
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. The document also evaluates ambient O3 environmental
effects on vegetation and ecosystems, man-made materials, and surface level solar UV radiation
flux and global climate change, 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 draft document.
                                          I-iv

-------
           Air Quality Criteria for Ozone and Related
                   Photochemical Oxidants
                (Second External Review Draft)


                         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.  EFFECT OF OZONE ON MAN-MADE MATERIALS	11-1
                             I-v

-------
           Air Quality Criteria for Ozone and Related
                   Photochemical Oxidants
                (Second External Review Draft)
                           (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

-------
                                 Table of Contents

                                                                                Page

List of Tables	  I-xviii
List of Figures 	I-xx
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-xxxviii
U.S. Environmental Protection Agency Science Advisory Board (SAB) Staff Office
      Clean Air Scientific Advisory Committee (CASAC) Ozone Review Panel  	I-xli
Abbreviations and Acronyms	I-xliv

EXECUTIVE SUMMARY	  E-l
     E.I    INTRODUCTION	  E-l
           E.I.I     Clean Air Act Legal Requirements  	  E-l
           E. 1.2     Chronology of Ozone NAAQS Revisions	  E-2
           E.I.3     Document Organization and Structure	  E-2
     E.2    ATMOSPHERIC CHEMISTRY AND PHYSICS OF TROPOSPHERIC
           OZONE FORMATION	  E-3
     E.3    ENVIRONMENTAL DISPERSAL, AMBIENT CONCENTRATIONS,
           AND HUMAN EXPOSURE TO OZONE	  E-6
     E.4    DOSIMETRIC STUDIES	  E-9
     E.5    ANIMAL TOXICOLOGY ASPECTS	  E-10
           E.5.1     Respiratory Tract Effects of Short-Term Exposures to Ozone	  E-ll
           E.5.2     Respiratory Tract Effects of Chronic (Long-Term) Exposures
                    to Ozone	  E-l5
           E5.3     Other Types of Ozone Exposure Effects Observed in Laboratory
                    Animal Models	  E-l5
     E.6    CONTROLLED HUMAN EXPOSURE STUDIES	  E-17
     E.7    EPIDEMIOLOGIC STUDIES 	  E-18
           E.7.1     Health Effects Associated with Acute Ozone Exposures	  E-18
           E.7.2     Issues Potentially Affecting Interpretation of Acute Exposure Studies ...  E-20
           E.7.3     Health Effects Associated with Chronic Ozone Exposure	  E-22
     E.8    INTEGRATIVE SYNTHESIS 	  E-22
     E.9    VEGETATION AND ECOLOGICAL EFFECTS	  E-28
     E. 10  TROPOSPHERIC OZONE EFFECTS ON UV-B FLUX AND ITS ROLE
           IN CLIMATE CHANGE 	  E-32
     E.ll  MATERIALS DAMAGE	  E-35

1.    INTRODUCTION	1-1
     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
                                        I-vii

-------
                             Table of Contents
                                   (cont'd)
                                                                        Pas
          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	2-1
     2.1   INTRODUCTION	2-1
     2.2   CHEMICAL PROCESSES INVOLVED IN OZONE FORMATION
          AND DESTRUCTION	2-2
     2.3   METEOROLOGICAL PROCESSES AFFECTING OZONE	2-7
     2.4   RELATIONS OF OZONE TO ITS PRECURSORS 	2-13
     2.5   THE ROLE OF CHEMISTRY-TRANSPORT MODELS IN
          UNDERSTANDING ATMOSPHERIC OZONE	2-17
     2.6   TECHNIQUES FOR MEASURING OZONE AND ITS PRECURSORS	2-21
     2.7   SUMMARY 	2-23
     REFERENCES 	2-25

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
          3.3.2    Small-scale Horizontal and Spatial Variability in
                  Ozone Concentrations 	3-14
     3.4   DIURNAL AND SEASONAL VARIABILITY OF OZONE 	3-17
     3.5   TRENDS IN OZONE CONCENTRATIONS  	3-33
     3.6   RELATIONSHIPS BETWEEN OZONE AND OTHER SPECIES	3-36
     3.7   POLICY RELEVANT BACKGROUND OZONE CONCENTRATIONS 	3-44
     3.8   OZONE EXPOSURE IN VARIOUS MICROENVIRONMENTS  	3-52
     3.9   SUMMARY OF KEY POINTS 	3-72
     REFERENCES 	3-77

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	4-4
          4.2.2    General Uptake Studies 	4-7
          4.2.3    Dosimetry Modeling	4-9
          4.2.4    Summary  and Conclusions - Dosimetry	4-12
     4.3   SPECIES HOMOLOGY, SENSITIVITY, AND ANIMAL-TO-
          HUMAN EXTRAPOLATION	4-14
          4.3.1    Summary  and Conclusions:  Species Homology, Sensitivity,
                  and Animal-to-Human Extrapolation	4-20
     REFERENCES 	4-21


                                    I-viii

-------
                                   Table of Contents
                                          (cont'd)
                                                                                      Page
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 O3 Interaction	5-2
                     5.2.1.2     Monooxygenases	5-4
                     5.2.1.3     Antioxidants, Antioxidant Metabolism, and Mitochondrial
                                Oxygen Consumption	5-5
                     5.2.1.4     Lipid Metabolism and Content of the Lung 	5-7
                     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
            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	5-27
                     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	5-31
            5.2.4     Morphological Effects 	5-34
                     5.2.4.1     Acute and Subchronic Exposure Effects	5-35
                     5.2.4.2     Summary of Acute and Subchronic Morphological Effects .  . . 5-38
                     5.2.4.3     Subchronic and Chronic Exposure Effects	5-39
                     5.2.4.4     Summary and Conclusions - Subchronic and Chronic
                                Morphological Effects	5-42
            5.2.5     Effects on Pulmonary Function  	5-43
                     5.2.5.1     Acute and Subchronic Exposure Effects on
                                Pulmonary Function  	5-43
                     5.2.5.2     Summary and Conclusions - Acute and Subchronic
                                Effects on Pulmonary Function	5-45
                     5.2.5.3     Ozone Effects on Airway Responsiveness 	5-45
                     5.2.5.4     Summary and Conclusions - Effects on Airway
                                Responsiveness	5-53


                                            I-ix

-------
                                  Table of Contents
                                        (cont'd)
                                                                                  Page
            5.2.6    Genotoxicity Potential of Ozone  	5-54
                    5.2.6.1    Summary and Conclusions - Genotoxicity Potential
                              of Ozone  	5-55
     5.3    SYSTEMIC EFFECTS OF OZONE EXPOSURE	5-55
            5.3.1    Neurobehavioral Effects	5-55
            5.3.2    Neuroendocrine Effects 	5-57
            5.3.3    Cardiovascular Effects	5-57
            5.3.4    Reproductive and Developmental Effects  	5-59
            5.3.5    Effects on the Liver, Spleen, and Thymus	5-61
            5.3.6    Effects on Cutaneous and Ocular Tissues  	5-61
            5.3.7    Summary and Conclusions - Systemic Effects of Ozone	5-62
     5.4    INTERACTIONS OF OZONE WITH OTHER CO-OCCURRING
            POLLUTANTS	5-63
            5.4.1    Ozone and Nitrogen Oxides	5-64
            5.4.2    Ozone and Other Copollutants	5-65
            5.4.3    Complex (Multicomponent) Mixtures Containing Ozone	5-67
            5.4.4    Summary and Conclusions - Interactions of Ozone with other
                    Co-occurring Pollutants	5-74
     5.5    EFFECTS OF OTHER PHOTOCHEMICAL OXIDANTS	5-76
            5.5.1    Summary and Conclusions - Effects of Other Photochemical Oxidants  . . . 5-77
            REFERENCES 	5-78

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	6-3
            6.2.3    Prolonged Ozone Exposures 	6-5
                    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  	6-6
                    6.2.3.3    Comparison of 2 h IE to 6.6 h O3 Exposure Effects
                              on Pulmonary Function	6-7
            6.2.4    Triangular Ozone Exposures 	6-7
            6.2.5    Mechanisms of Pulmonary Function Responses 	6-8
                    6.2.5.1    Pathophysiologic Mechanisms  	6-10
                    6.2.5.2    Mechanisms at a Cellular and Molecular Level 	6-13
     6.3    SUBJECTS WITH PREEXISTING DISEASE  	6-14
            6.3.1    Subjects with Chronic Obstructive Pulmonary Disease	6-14
            6.3.2    Subjects with Asthma	6-15
            6.3.3    Subjects with Allergic Rhinitis	6-17


                                          I-x

-------
                                  Table of Contents
                                        (cont'd)
                                                                                  Page
            6.3.4    Subjects with Cardiovascular Disease  	6-19
     6.4    INTERSUBJECT VARIABILITY AND REPRODUCIBILITY OF RESPONSE . .  . 6-19
     6.5    FACTORS MODIFYING RESPONSIVENESS TO OZONE	6-21
            6.5.1    Influence of Age	6-21
            6.5.2    Gender and Hormonal Influences	6-23
            6.5.3    Racial, Ethnic, and Socioeconomic Status Factors	6-24
            6.5.4    Influence of Physical Activity 	6-24
            6.5.5    Environmental Factors	6-24
            6.5.6    Oxidant-Antioxidant Balance	6-26
            6.5.7    Genetic Factors	6-26
     6.6    REPEATED O3 EXPOSURE EFFECTS 	6-28
     6.7    EFFECTS ON EXERCISE PERFORMANCE	6-28
     6.8    EFFECTS ON AIRWAY RESPONSIVENESS	6-29
     6.9    EFFECTS ON INFLAMMATION AND HOST DEFENSE	6-30
            6.9.1    Introduction  	6-30
            6.9.2    Inflammatory Responses in the Upper Respiratory Tract  	6-31
            6.9.3    Inflammatory Response in the Lower Respiratory Tract	6-32
            6.9.4    Adaptation of Inflammatory Responses	6-36
            6.9.5    Effect of Anti-Inflammatory and Other Mitigating Agents	6-37
            6.9.6    Changes in Host Defense Capability Following Ozone Exposures	6-38
     6.10   EXTRAPULMONARY EFFECTS OF OZONE  	6-40
     6.11   EFFECTS OF OZONE MIXED WITH OTHER POLLUTANTS	6-41
     6.12   CONTROLLED STUDIES OF AMBIENT AIR EXPOSURES  	6-42
            6.12.1   Mobile Laboratory Studies	6-42
            6.12.2   Aircraft Cabin Studies 	6-43
     6.13   SUMMARY  	6-43
     REFERENCES  	6-46

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	7-2
            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	7-5
                    7.1.3.2    Ozone Exposure Indices Used	7-8
                    7.1.3.3    Lag Time: Period between Ozone Exposure  and
                              Observed Health Effect	7-9
                    7.1.3.4    Model Specification to Adjust for Temporal Trends and
                              Meteorologic Effects	7-11
                    7.1.3.5    Confounding Effects of Copollutants 	7-14
                    7.1.3.6    Model Uncertainty from Multiple Hypothesis Testing	7-15


                                          I-xi

-------
                              Table of Contents
                                    (cont'd)
                                                                                 Page
                7.1.3.7    Impact of GAM Convergence Issue on Ozone Risk
                          Estimates	7-18
       7.1.4     Approach to Presenting Ozone Epidemiologic Evidence	7-20
7.2    FIELD STUDIES ADDRESSING ACUTE EFFECTS OF OZONE	7-21
       7.2.1     Summary of Key Findings on Field Studies of Acute Ozone Effects
                From the 1996 O3 AQCD	7-21
       7.2.2     Introduction to Recent Field Studies of Acute Ozone Effects	7-22
       7.2.3     Acute Ozone Exposure and Lung Function	7-23
                7.2.3.1    Acute Ozone Studies with Spirometry (FEVj) 	7-24
                7.2.3.2    Acute Ozone Studies of PEF	7-34
       7.2.4     Respiratory Symptoms	7-41
       7.2.5     Acute Airway Inflammation	7-49
       7.2.6     Acute Ozone Exposure and School Absences	7-50
       7.2.7     Cardiovascular Endpoints	7-52
                7.2.7.1    Cardiac Autonomic Control  	7-52
                7.2.7.2    Acute Myocardial Infarction	7-55
                7.2.7.3    Cardiovascular Endpoints in Human Clinical Studies 	7-56
                12.1 A    Summary of Field Studies with Cardiovascular Outcomes .... 7-56
       7.2.8     Summary of Field Studies Assessing Acute Ozone Effects	7-56
7.3    ACUTE EFFECTS OF OZONE ON DAILY EMERGENCY DEPARTMENT
       VISITS AND HOSPITAL ADMISSIONS	7-57
       7.3.1     Summary of Key Findings on Studies of Emergency Department
                Visits and Hospital Admissions from the 1996 O3 AQCD	7-57
       7.3.2     Review of Recent Studies of Emergency Department Visits for
                Respiratory Diseases  	7-58
       7.3.3     Studies of Hospital Admissions for Respiratory Diseases	7-62
                7.3.3.1    Potential Confounding of the Ozone Effect on Respiratory
                          Hospitalizations by Copollutants  	7-69
       7.3.4     Association of Ozone with Hospital Admissions for Cardiovascular
                Disease  	 7-71
       7.3.5     Summary of Acute Ozone Effects on Daily Emergency Department
                Visits and Hospital Admissions	7-73
7.4    ACUTE EFFECTS OF OZONE ON MORTALITY 	7-74
       7.4.1     Summary of Key Findings on Acute Effects of Ozone on Mortality
                From the 1996 O3 AQCD	7-74
       7.4.2     Introduction to Assessment of Current Ozone-Mortality Studies 	7-74
       7.4.3     Single-Pollutant Model Ozone-Mortality Risk Estimates  	7-75
       7.4.4     Meta-analyses of O3-Mortality Risk Estimates 	7-81
       7.4.5     Seasonal Variation in Ozone-Mortality Risk Estimates	7-84
       7.4.6     Ozone-Mortality Risk Estimates Adjusting for PM Exposure	7-87
       7.4.7     Ozone Risk Estimates for Specific Causes of Mortality  	7-89
       7.4.8     Ozone-Mortality Risk Estimates for Specific Subpopulations	7-94
       7.4.9     Summary of Acute Ozone Effects on Mortality	7-96
                                      I-xii

-------
                              Table of Contents
                                    (cont'd)
                                                                                Page
7.5     EFFECTS OF CHRONIC OZONE EXPOSURE	7-97
       7.5.1     Summary of Key Findings on Studies of Health Effects and Chronic
                Ozone Exposure from the 1996 O3 AQCD	7-97
       7.5.2     Introduction to Morbidity Effects of Chronic Ozone Exposure	7-97
       7.5.3     Seasonal Ozone Effects on Lung Function  	7-98
       7.5.4     Chronic Ozone  Exposure Effects on Lung Function and Respiratory
                Symptoms	7-100
       7.5.5     Chronic Ozone  Exposure and Respiratory Inflammation	7-106
       7.5.6     Risk of Asthma Development	7-108
       7.5.7     Respiratory Effects of Chronic Ozone Exposure on Susceptible
                Populations	7-110
       7.5.8     Effects of Chronic Ozone Exposure on Mortality and
                Cancer Incidence  	7-111
       7.5.9     Effects of Ozone on Birth-Related Health Outcomes	7-114
       7.5.10    Summary of Chronic  Ozone Exposure Effects on Morbidity
                and Mortality  	7-117
7.6     INTERPRETIVE ASSESSMENT OF THE EVIDENCE IN EPIDEMIOLOGIC
       STUDIES OF OZONE HEALTH EFFECTS	7-118
       7.6.1     Introduction  	7-118
       7.6.2     Ozone Exposure Indices	7-118
       7.6.3     Confounding by Temporal Trends and Meteorologic Effects in
                Time-Series Studies	7-120
                7.6.3.1     Assessment of Ozone Effects after Adj usting for
                          Temporal Trends and Meteorologic Effects	7-121
                7.6.3.2     Importance of Season-Specific Estimates of Ozone
                          Health Effects  	7-123
       7.6.4     Assessment of Confounding by Copollutants  	7-130
                7.6.4.1     Relationship between Personal Exposure to Ozone
                          and Copollutants	7-130
                7.6.4.2     Assessment of Confounding Using Multipollutant
                          Regression Models  	7-131
       7.6.5     Concentration-Response Function and Threshold	7-134
       7.6.6     Heterogeneity of Ozone Health Effects	7-138
       7.6.7     Health Effects of Ozone in Susceptible Populations 	7-142
                7.6.7.1     Health Effects Associated with Ambient Ozone
                          Exposure in Asthmatics	7-142
                7.6.7.2     Age-Related Differences in Ozone Effects	7-146
       7.6.8     Summary of Key Findings and Conclusions Derived From
                Ozone Epidemiologic Studies  	7-148
REFERENCES  	7-152
                                     I-xiii

-------
                             Table of Contents
                                    (cont'd)
                                                                                Pa
INTEGRATIVE SYNTHESIS: EXPOSURE AND HEALTH EFFECTS  	8-1
8.1     INTRODUCTION	8-1
       8.1.1     Chapter Organization	8-2
8.2     AMBIENT OZONE AIR QUALITY IN 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	8-5
       8.2.4     Interrelationships Between Ozone and Other Ambient Pollutants	8-6
       8.2.5     Policy Relevant Background (PRB) Ozone Concentrations 	8-7
8.3     FACTORS AFFECTING HUMAN EXPOSURE TO AMBIENT OZONE  	8-8
       8.3.1     Personal Exposure 	8-9
       8.3.2     Indoor Concentrations 	8-9
8.4     SYNTHESIS OF AVAILABLE INFORMATION ON OZONE-RELATED
       HEALTH EFFECTS	8-11
       8.4.1     Key Health-Related Findings and Conclusions from the
                1996 Ozone Air Quality Criteria Document 	8-12
       8.4.2     Assessment and Integration of New Experimental Evidence	8-18
                8.4.2.1    Background on Cross-Cutting Issues  	8-18
                8.4.2.2    Approaches to Experimental Evaluation of Ozone
                          Health Effects 	8-19
                8.4.2.3    Interspecies Comparison of Experimental Results:
                          Dosimetric Considerations  	8-20
                8.4.2.4    Critical Analysis of Toxicological Effects of
                          O3 Exposure	8-23
                          8.4.2.4.1      Pulmonary Function	8-23
                          8.4.2.4.2      Airway Responsiveness	8-28
                          8.4.2.4.3      Morphological and Biochemical
                                       Abnormalities	8-30
       8.4.3     Assessment of Epidemiological Evidence	8-39
                8.4.3.1    Strength and Consistency of Epidemiological
                          Associations 	8-40
                          8.4.3.1.1      Acute Exposure Studies	8-41
                8.4.3.2    Robustness of Epidemiological Associations	8-51
                          8.4.3.2.1      Exposure Issues: Ambient versus
                                       Personal 	8-53
                          8.4.3.2.2      Confounding by Temporal Trends
                                       and Meteorologic Effects	8-53
                          8.4.3.2.3      Assessment of Confounding
                                       by Copollutants 	8-54
                8.4.3.3    Lag Period between Ozone Exposure and
                          Health Response  	8-55
                8.4.3.4    Concentration-Response Functions  and Threshold	8-55
                                     I-xiv

-------
                                 Table of Contents
                                      (cont'd)
                                                                                Page
                    8.4.3.5    Consistency of Findings Across Epidemiologic Studies	8-56
                    8.4.3.6    Summary and Conclusions for Epidemiology Findings	8-57
     8.5    BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE
           FOR OZONE-RELATED HEALTH EFFECTS  	8-58
           8.5.1     Acute Ozone Exposure-Induced Health Effects	8-63
           8.5.2     Chronic O3 Exposure-Induced Health Effects  	8-69
           8.5.3     Mortality-Related Health Endpoints  	8-70
     8.6    SUSCEPTIBILITY FACTORS  	8-71
           8.6.1     Preexisting Disease as a Potential Risk Factor	8-73
           8.6.2     Potential Public Health Impacts	8-74
                    8.6.2.1    General Concepts Related to Defining of Adverse
                             Health Effects  	8-75
                    8.6.2.2    Estimation of Potential Numbers of Persons in At-Risk
                             Susceptible Population Groups in the United States	8-76
     8.7    SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS  	8-82
     REFERENCES  	8-88

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/MODE-OF-ACTION	9-5
     9.4    MODIFICATION OF FUNCTIONAL AND GROWTH RESPONSES  	9-6
     9.5    EFFECTS-BASED AIR QUALITY EXPOSURE INDICES	9-11
     9.6    OZONE EXPOSURE-PLANT RESPONSE RELATIONSHIPS	9-13
     9.7    EFFECTS OF OZONE EXPOSURE ON NATURAL ECOSYSTEMS  	9-17
     9.8    ECONOMICS	9-20
     REFERENCES  	9-22

10.   TROPOSPHERIC OZONE EFFECTS ON UV-B FLUX, AND ITS ROLE 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	10-2
                    10.2.1.1   UV Radiation::  Wavelengths, Energies and Depth of
                             Atmospheric Penetration	10-2
                    10.2.1.2   Temporal Variations in Solar Flux	10-3
                    10.2.1.3   Atmospheric Radiative Interactions with Solar
                             Ultraviolet Radiation	10-5
                    10.2.1.4   Data Requirements for a Surface UV-B Climatology	10-12
           10.2.2    Factors Governing Human Exposure to Ultraviolet Radiation  	10-13
                    10.2.2.1   Outdoor Activities	10-14
                    10.2.2.2   Occupation  	10-15


                                        I-xv

-------
                              Table of Contents
                                     (cont'd)
                                                                                  Page
                10.2.2.3    Age  	10-16
                10.2.2.4    Gender	10-16
                10.2.2.5    Geography	10-17
                10.2.2.6    Protective Behavior	10-17
                10.2.2.7    Summary of Factors that Affect Human Exposures to
                           Ultraviolet Radiation	10-18
       10.2.3    Factors Governing Human Health Effects due to Ultraviolet
                Radiation  	10-18
                10.2.3.1    Erythema	10-19
                10.2.3.2    Skin Cancer	10-21
                10.2.3.3    Ultraviolet Radiation Exposure and the Incidence of
                           Nonmelanoma Skin Cancers	10-22
                10.2.3.4    Ocular Effects of Ultraviolet Radiation Exposure  	10-29
                10.2.3.5    Ultraviolet Radiation and Immune System Suppression	10-31
                10.2.3.6    Protective Effects of Ultraviolet Radiation - Production
                           of Vitamin D	10-33
       10.2.4    Summary and Conclusions for Ozone Effects on UV-B Flux	10-34
10.3   TROPOSPHERIC OZONE AND CLIMATE CHANGE	10-35
       10.3.1    The Projected Impacts of Global Climate Change	10-36
       10.3.2    Solar Energy Transformation and the Components of the Earth's
                Climate System	10-40
       10.3.3    The Composition of the Atmosphere and the Earth's Radiative
                Equilibrium	10-42
                10.3.3.1    Forcing of the Earth's Radiative Balance	10-43
       10.3.4    Factors Affecting the Magnitude of Climate Forcing by Ozone 	10-45
                10.3.4.1    The Global Burden of Tropospheric Ozone  	10-46
                10.3.4.2    Background Concentrations versus Regionally-Oriented
                           Ozone Enhancements  	10-47
                10.3.4.3    Ozone Trends:  Globally and in North America	10-48
                10.3.4.4    The Sensitivity of Ozone-Related Forcing Surface
                           to Albedo	10-51
                10.3.4.5    The Altitude Dependence of Forcing by Tropospheric
                           Ozone  	10-52
                10.3.4.6    Co-occurrence of Ozone with Particulate Matter	10-52
       10.3.5    Estimated Forcing by Tropospheric Ozone  	10-53
                10.3.5.1    Direct Climate Forcing Due to Ozone	10-53
                10.3.5.2    Indirect Forcing Due to Ozone	10-55
                10.3.5.3    Predictions for Future Climate Forcing by
                           Anthropogenic  Ozone	10-56
       10.3.6    The Impact of a Warming Climate on Atmospheric
                Ozone Concentrations  	10-57
       10.3.7    Conclusion	10-57
REFERENCES 	10-59
                                      I-xvi

-------
                                Table of Contents
                                     (cont'd)

                                                                              Page

11.
EFFEi
11.1
11.2
11.3
11.4
11.5
11.6
REFE
CT OF OZONE ON MAN-MADE MATERIALS 	
ELASTOMERS
TEXTILES AND FABRICS 	
DYES, PIGMENTS, AND INKS 	
ARTISTS' PIGMENTS
SURFACE COATINGS 	
CONCLUSIONS
RENCES 	
	 11-1
11-1
	 11-3
	 11-4
11-5
	 11-12
11-13
	 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	3-61

3-4        Indoor/Outdoor Ozone Concentrations in Various Microenvironments 	3-63

7-la       Field Studies that Investigated the Association Between Acute Ambient O3
           Exposure and Changes in FEV] in Adults	7-25

7-lb       Percent Changes in FEVj (95% CI) Associated with Acute Ambient O3
           Exposures in Adults, Ordered by Size of the Estimate  	7-26

7-lc       Cross-day Percent Changes in FEVj (95% CI) Associated with Acute Ambient
           O3 Exposures in Adults, Ordered by Size of the Estimate	7-27

7-2a       Field Studies that Investigated the Association Between Acute Ambient O3
           Exposure and Changes in FEVj in Children 	7-28

7-2b       Percent Changes in FEV] (95% CI) Associated with Acute Ambient O3 Exposures
           in Children, Ordered by Size of the Estimate	7-29

7-2c       Cross-day Percent Changes in FEVj (95% CI) Associated with Acute Ambient
           O3 Exposures in Children, Ordered by Size of the Estimate 	7-30

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	7-103

8-1        Acute O3-induced Physiological and Biochemical Changes in Human
           and Animals  	  8-60

8-2        Gradation of Individual Responses to Short-Term Ozone Exposure in
           Healthy Persons  	8-76

8-3        Gradation of Individual Responses to Short-Term Ozone Exposure in
           Persons with Impaired Respiratory Systems 	8-77
                                           I-xviii

-------
                                      List of Tables
                                          (cont'd)

Number                                                                                Page

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)	8-79

8-5        Acute Respiratory Conditions per 100 Persons/Year by Age Group in the
           United States (1996 National Health Interview Survey) 	8-80

10-1       Examples of Impacts Resulting From Proj ected Changes in Extreme
           Climate Events	10-38

10-2       CTM Studies Assessed by the IPCC for its Estimate of the Change in Global
           and Total Column O3 Since the Preindustrial Era  	10-50

10-3       Tropospheric O3 Change (O3) in Dobson Units (DU) Since Preindustrial
           Times, and the Accompanying Net (SW plus LW) Radiative Forcings (WnT2),
           After Accounting for Stratospheric Temperature Adjustment (using the Fixed
           Dynamical Heating Method) 	10-54

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-xix

-------
                                      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	2-8

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)  	2-8

2-3        The diurnal evolution of the planetary boundary layer while high pressure
           prevails over land	2-10

2-4        Locations of low level jet occurrences in decreasing order of prevalence
           (most frequent, common, observed)	2-10

2-5        Conceptual two-reservoir model showing conditions in the PEL and in the
           lower free troposphere during a multiday O3 episode  	2-12

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	2-14

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) 	2-17

2-9        Main components of a comprehensive atmospheric chemistry modeling system,
           such as Models-3 	2-18

3-1        County wide mean daily maximum 8-h O3 concentrations, May to September
           2000 to 2004	3-4

3-2        County wide 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

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
                                            I-xx

-------
                                       List of Figures
                                           (cont'd)

Number                                                                                   Page

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 hour 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-lla-f    Diurnal variability in 8 hour averaged O3 in selected urban areas	3-22

3-llg-l    Diurnal variability in 8 hour averaged O3 in selected urban areas	3-23

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 hour averaged 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  	3-32

3-16       Composite diurnal variability  in 8-h O3 concentrations observed at
           CASTNET sites  	3-32

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
                                             I-xxi

-------
                                      List of Figures
                                           (cont'd)

Number                                                                                  Page

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	3-37

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 	3-38

3-21       Binned mean PM25 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	3-43

3-24       The co-occurrence pattern for O3 and PM2 5 using 2001 data from AQS	3-44

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  	3-49
3-27       Time-series of hourly average O3 concentrations observed at five national
           parks: Denali (AK), Voyageur (MM), Olympic (WA), Glacier (MT),
           and Yellowstone (WY)	3-51

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 - tj = T)  	3-54

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  	3-57

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


                                            I-xxii

-------
                                      List of Figures
                                           (cont'd)

Number                                                                                  Page

4-3        Ozone uptake efficiency as a function of breathing frequency at a minute
           ventilation of 30 L/min	4-9

5-1        Major secondary products of ozone interaction with epithelial lining fluid and
           lung cells 	5-3

5-2        (Reprinted from Molecular Aspects of Medicine, IS. Mudway  and F. J. Kelly,
           Ozone and the Lung:  a sensitive issue, page 36, (2000), with permission
           from Elsevier)	5-6

5-3        Mechanisms of  Ozone Toxicity	5-9

5-4        Mouse chromosomes on which genes or gene loci have been identified that
           modulateresponses to O3  	5-33

6-1        Triangular exposure profile - O3-induced FEV] decrements (top panel) and O3
           concentrations (bottom panel) as a function of exposure duration	6-9

6-2        Recovery of FEVj responses following a 2 h exposure to 0.4 ppm O3 with IE	6-11

6-3        Predicted O3-induced decrements in FEVj as a function of exposure duration
           and level of IE (line labels areVE levels) in young healthy adults (20 yrs of age)
           exposed to 0.3 ppm O3	6-25

6-4        Time course of acute responses seen in humans exposed to O3	6-34

7-1        Percent change (95% CI) in morning PEF in children per standardized increment
           (see Section 7.1.3.2)	7-35

7-2        Percent change (95% CI) in afternoon PEF in children per standardized increment
           (see Section 7.1.3.2)	7-36

7-3        Comparison of single-day lags (1-, 2-, 3-, 4-, 5-, and 6-day) to a cumulative
           multiday lag (1- to 5-day) for percent changes in PEF per 30 ppb increase in
           8-h avg O3 in urban children 	7-38

7-4        Density curves of the percent change in PEF per 30 ppb increase in 8-h avg
           O3 with a cumulative lag of 1 to 5 days for the individual eight NCICAS cities
           and the pooled average of all cities  	7-39

7-5        Odds ratios for the incidence of cough among asthmatic children per standardized
           increment (see Section 7.1.3.2)  	7-43
                                            I-xxiii

-------
                                       List of Figures
                                            (cont'd)

Number                                                                                    Page

7-6        Odds ratios for extra medication use among asthmatic children per standardized
           increment (see Section 7.1.3.2)  	7-44

7-7        Density curves of the odds ratios for the incidence of symptoms per 30 ppb
           increase in 8-h avg O3 with a cumulative lag of 1 to 4 days for the individual
           eight cities and the pooled average of all cities  	7-46

7-8        Ozone-associated percent change (95% CI) in emergency department visits for
           asthma per standardized increment (see Section 7.1.3.2)	7-59

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) 	7-64

7-10       Ozone-associated percent change (95% CI) in total respiratory hospitalizations
           by season per standardized increment (see Section 7.1.3.2)  	7-65

7-11       Comparison of single-day lags (0-, 1-, 2-, 3-,  4-, and 5-day) to a cumulative
           multiday lag (0- to 4-day) for percent changes in total respiratory hospitalizations
           per 40 ppb increase in 1-h max O3 in children less than two years of age	7-69

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)	7-70

7-13       Ozone-associated percent change (95% CI) in total cardiovascular hospitalizations
           per standardized increment (see Section 7.1.3.2)	7-72

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)	7-76

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)	7-77

7-16       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	7-79

7-17       Comparison of single-day lags (0-, 1-, 2-, and 3-day) to a cumulative multiday
           lag (0- to 6-day) for percent changes in all cause mortality per 20 ppb increase
           in 24-h avg O3 in all  ages	7-80

7-18       Combined all  cause (nonaccidental) O3 excess mortality risk estimates (95% CI)
           from recent meta-analyses per standardized increment (see Section 7.1.3.2)  	7-82
                                             I-xxiv

-------
                                       List of Figures
                                            (cont'd)

Number                                                                                   Page

7-19       All cause (nonaccidental) O3 excess mortality risk estimates (95% CI) by season
           per standardized increment (see Section 7.1.3.2)	7-86

7-20       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)	7-88

7-21       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)	7-90

7-22       Ozone-associated cardiovascular mortality risk estimates (95% CI) per
           standardized increment (see Section 7.1.3.2)	7-92

7-23       Adjusted average annual increases in FEV] and maximal midexpiratory flow
           (MMEF) versus the mean 8-h avg O3 (10 a.m. to 6 p.m.) concentration over
           a 4-year period in the 12 southern California communities of the Children's
           Health Study	7-102

7-24       Adjusted Cymortality 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	7-112

7-25       The relationship between PM and O3 in the summer (June through August)
           and the winter (December through February) as sorted and averaged by
           quintiles of PM	7-125

7-26       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) 	7-128

7-27       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)	7-129

7-28       Posterior means and 95% Pis of 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	7-132

7-29       Maximum likelihood estimates of Cymortality for 95 U.S. communities,
           determined using a constrained distributed lag model for lags 0 through 6 days 	7-133

8-1        Frequency distributions of FEVj decrements following 6.6-h exposures to
           O3  or filtered air  	  8-25
                                             I-xxv

-------
                                       List of Figures
                                            (cont'd)

Number                                                                                   Page

8-2        Proportion of moderately exercising healthy adults (24 yrs old) predicted
           to have 5, 10, or 15% decrements in FEVj as a function of concentration
           (0 to 0.12 ppm O3) times exposure duration (1 to 6.6 h) 	8-26
8-3        Effect of age on FEVj responses to O3 exposure (0.42 ppm for 1.5 h with
           intermittent exercise)  	8-27

8-4        Neutrophilia response in the distal airways postexposure (PE) to O3 or filtered air .... 8-34

8-5        Ozone-associated percent change (95% CI) in emergency department visits
           for asthma (A), total respiratory hospitalization by season (B), respiratory
           hospitalization with adjustment for PM indices (C) and (D) total cardiovascular
           hospitalization per 40 ppb increase in 1-h max O3 or equivalent	8-47

8-6        A. All cause (nonaccidental) O3 excess mortality risk estimates (95% CI) for
           all year analyses per standardized increment (see Section 7.1.3.2)	8-49

8-7        All causes (nonaccidental) O3 excess mortality risk estimates (95% CI) per
           standardized increment (see Section 7.1.3.2)	8-51

8-8        Ozone-associated percent change (95% CI) in cardiovascular risk estimates
           per standardized increment (see Section 7.1.3.2)	8-52

8-9        Resolution time-line for the physiological and biochemical parameters are
           derived from studies reported in Chapter 6 and Chapter 6 Annex	8-61

8-10       Os-induced cellular and molecular changes and their evolution depicted here
           is derived from the data reported in Leikauf et al. (1995) and Mudway and
           Kelly (2000)	8-62

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  	10-4

10-3       Ozone column abundances from the years 1990 to 1992 for 0, 40, and 80° N
           as well as 80° S	10-7

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)	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)	10-12


                                            I-xxvi

-------
                                      List of Figures
                                           (cont'd)

Number                                                                                 Page

10-6       Estimated global mean radiative forcing exerted by gas and various particle
           phase species for the year 2000, relative to 1750	10-45

10-7       Mid-tropospheric O3 abundance (ppb) in northern midlatitudes (36 °N-59 °N)
           for the years 1970 to 1996  	10-49

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

11-3       Observed color changes for natural colorant-on-paper systems during exposure
           to 0.40 ppm ozone 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
           ozone 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 CHEMISTRY OF OZONE IN THE ATMOSPHERE


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—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—Harvard University, Cambridge, MA

Dr. William Keene—University of Virginia, Charlottesville, VA

Dr. Tadeusz Kleindienst—National Exposure Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC

Dr. Jennie Moody—University of Virginia, Charlottesville, VA

Mr. Charles Piety—University of Maryland, College Park, MD

Dr. Sandy Sillman—University of Michigan, Ann Arbor, MI

Dr. Jeffrey Stehr—University of Maryland, College Park, MD

Dr. Bret Taubman—Pennsylvania State University, State College, PA
                                        I-xxviii

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)
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, NOAA/GFDL, Princeton, NJ

Mr. Chris Geron, NRML, U.S. EPA, Research Triangle Park, NC

Dr. David Golden, Stanford University, Palo Alto, CA

Dr. John Merrill, University of Rhode Island, Kingston, RI

Dr. Sam Oltmans, NOAA, CMDL, Boulder, CO

Dr. David Parrish, NOAA/AL, Boulder, CO

Dr. Perry Samson, Depart. Atmos. Ocean, and Space Sciences, University of Michigan,
Ann Arbor, MI

Dr. Sandy Sillman, University of Michigan, Ann Arbor, MI

Dr. Melvin Shapiro, National Center for Atmospheric Research, Boulder, CO
          CHAPTER 3 - ENVIRONMENTAL CONCENTRATIONS, PATTERNS,
                            AND EXPOSURE ESTIMATES
Principal Authors

Ms. Beverly Comfort—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. Arlene Fiore—NOAA/GFDL, Princeton, NJ

Dr. Daniel Jacob—Harvard University, Cambridge, MA
                                        I-xxix

-------
                       Authors, Contributors, and Reviewers
                                        (cont'd)
Principal Authors
(cont'd)

Dr. Alan S. Lefohn—ASL &Associates, Helena, MT

Dr. Clifford Weisel—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—U.S. EPA, NERL U.S. EPA, Research Triangle Park, NC

Contributors and Reviewers

Dr. Christoph Bruehl—Max Planck Institute for Atmospheric Chemistry, Mainz, Germany

Dr. Russell Dickerson—University of Maryland, College Park, MD

Dr. Judith  Graham—American Chemistry Council, Washington, D.C.

Dr. Laszlo Horvath—Hungarian Meteorological Service, Budapest, Hungary

Dr. Ted Johnson—TRJ Associates, Durham, NC

Dr. John Merrill—University of Rhode Island, Kingston, RI

Dr. Jennie Moody—University of Virginia, Charlottesville, VA

Dr. Sam Oltmans—NOAA CMDL, Boulder, CO

Dr. Michiel G.M. Roemer, TNO, The Netherlands

Dr. Sandy  Sillman—University of Michigan, Ann Arbor, MI

Dr. Tamas Weidinger—University of Budapest, Budapest, Hungary

                                         I-xxx

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)
           CHAPTER 4 - DOSIMETRY, SPECIES HOMOLOGY, SENSITIVITY,
                               AND EXTRAPOLATION
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

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—Wayne State University, Detroit, MI

Dr. Carroll Cross—University of California, Davis, CA

Dr. Mitch Cohen—NYU School of Medicine, New York University, New York, NY
                                        I-xxxi

-------
                       Authors, Contributors, and Reviewers
                                        (cont'd)
Contributors and Reviewers

Dr. Steven Kleeberger—National Institute of Environmental Health Sciences, Research
Triangle Park, NC 27711

Dr. George Liekauf—University of Cincinnati, Cincinnati, OH

Dr. David Basset—Wayne State University, Detroit, MI

Dr. E.M. Postlethwait—University of Texas Medical Branch, Galveston, TX

Dr. Kent Pinkerton—University of California, Davis, CA

Dr. Jack Harkema—Michigan State University, East Lansing, MI

Dr. Edward Schelegle—University of California, Davis, CA

Dr Judith Graham—American Chemical Council, Arlington, VA


              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—University of California, Davis, CA (retired)

Dr. Milian J. Hazucha—University of North Carolina, Chapel Hill, NC

Dr. E. William Spannhake—Johns Hopkins University, Baltimore, MD

Contributors and Reviewers

Dr. Edward Avol—University of Southern  California, Los Angeles, CA

Dr. Henry Gong—Ranches Los Amigos Medical Center, Los Angeles, CA
                                        I-xxxii

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)
Contributors and Reviewers
(cont'd)

Dr. Jane Q. Koenig—University of Washington, Seattle, WA

Dr. Michael Madden—National Health and Environmental Effects Research Laboratory,
U.S. Environmental Protection Agency, Chapel Hill, NC

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 Kaz Ito—New York University, New York, NY

Dr. Pat Kinney—School of Public Health, Columbia University, 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-xxxiii

-------
                      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

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


   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

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—Michigan Technological University, Houghton,  MI

Dr. Michael Nannini—Ilinois State Water  Survey, IL

Dr. Nancy Grulke—USDA Forest Service, Riverside, CA

                                       I-xxxiv

-------
                       Authors, Contributors, and Reviewers
                                        (cont'd)
Principal Authors
(cont'd)

Dr. Richard Adams—Oregon State University., Corvallis, OR

Dr. Robert Heath—University of California, Riverside, CA,

Dr. Victor Runeckle—Vancouver, B.C., CN

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)

Contributors and Reviewers

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

                                         I-xxxv

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)
Contributors and Reviewers
(cont'd)

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. Maria-Jose Sanz—Fundacion CEAM, c/Charles Darein, 14-Parque Te Valencia, Spain

Dr. James Short!e—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
                                        I-xxxvi

-------
                      Authors, Contributors, and Reviewers
                                      (cont'd)
           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)
[Note: Any inadvertently omitted names of authors/reviewers will be inserted in the final draft of
this O3 AQCD, as will more complete addresses for all authors/reviewers.]
                                      I-xxxvii

-------
               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-xxxviii

-------
               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-xxxix

-------
               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. Faye Silliman—Word Processor, InfoPro, Inc., 8200 Greensboro Drive, Suite  1450, McLean,
VA 22102

Mr. John A. Bennett—Technical Information Specialist, Library Associates of Maryland,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852

Ms. Sandra L. Hughey—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
                                          I-xl

-------
      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, (jbalmes@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 @nj c. org)

Dr. William (Jim) Gauderman, Associate Professor, Preventive Medicine, University of
Southerm 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. JackHarkema, 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-xli

-------
      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, Cary, 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-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. 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, (jsu@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-xliii

-------
                          Abbreviations and Acronyms
a
AA
ACh
ADSS
AER
AEROCE
AHR
AHSMOG
AIRS
AM
ANF
AOP2
APHEA
AQCD
AQS
ARIC
ATLAS
A/V
P
BAL
BALF
BHR
BS
BSA
BMZ
BP
C
C x T
CAA
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
Aerometric Information Retrieval System
alveolar macrophage
atrial natriuretic factor
antioxidant protein 2
Air Pollution on Health:  European Approach (study)
Air Quality Criteria Document
Air Quality System
Atherosclerosis Risk in Communities (study)
atmospheric model by Kurucz
surface-to-volume ratio
beta-coefficient; slope of an equation
bronchioalveolar lavage
bronchioalveolar lavage fluid
bronchial hyperresponsiveness
black smoke
body surface area
basement membrane zone
blood pressure
concentration
concentration  x time; concentration times duration of exposure
Clean Air Act
                                        I-xliv

-------
CADS
CAPs
CAR
CASAC
CASTNet
CC16
CCSP
CD96

c
^dyn
CE
CFD
C2H5-H
C5H8
C6H16
CHAD
CH3-CHO
CH3-CC13
CH3-CO
CH4
CI
CIE

CINC
CLM
CMAQ
CO
CO2
COD
COP
COPD
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
Air Quality Criteria Document for Ozone and Related Photochemical
Oxidants; O3 AQCD
dynamic lung compliance
continuous exercise
computational fluid dynamics
ethane
isoprene
terpene
Consolidated Human Activities Database
acetaldehyde
methyl chloroform
acetyl
methane
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
                                        I-xlv

-------
CRP
CTM
DHBA
DNA
DOAS
DPPC
DU
e
EEC
ECG
EDU
EEG
ELF
ENA-78
ENSO
EPA
ETS
F
FA
FACE
fB
FEF
FEF25.75
FEVj
FVC
GAM
GCM
GEE
GHGs
GLM
C-reactive protein
chemistry transport model
2,3-dehydroxybenzoic acid
deoxyribonucleic acid
differential optical absorption spectroscopy/spectrometry
dipalmitoylglycero-3-phosphocholine
Dob son units
epsilon, convergence precision
exhaled breath condensate (fluid)
el ectrocardi ographi c
ethylenediurea
el ectroencephal ographi c
epithelial lining fluid
epithelial cell-derived neutrophil-activating peptide 78
El Nino-Southern Oscillation
U.S. Environmental Protection Agency
environmental tobacco smoke
female
filtered air
free-air carbon dioxide exposure
breathing frequency
forced expiratory flow
forced expiratory flow between 25 and 75% of vital capacity
forced expiratory volume in 1 second
forced vital capacity
Generalized Additive Model
general circulation model
Generalized Estimating Equation
greenhouse gases
Generalized Linear Model
                                         I-xlvi

-------
GM-CSF
G6PD
GR
GSH
GSHPx
GSTM1
H+
H2CO, HCHO
HDMA
HFCs
HNE
HNO2, HONO
HNO3
HO
HO2
H202
HR
HRV
H2SO4
1C
ICAM
ICNIRP
IE
Ig
IL
iNOS
ip
IPCC
IR
K,
granulocyte-macrophage colony stimulating factor
glucose-6-phosphate dehydrogenase
glutathione reductase
glutathione; reduced glutathione
glutathione peroxidase
glutathione S-transferase |i-l (genotype)
hydrogen ion
formaldehyde
house dust mite allergen
hydrofluorocarbons
4-hydroxynonenal
nitrous acid
nitric acid
hydroxyl
hydroperoxyl; hydroperoxy
hydrogen peroxide
heart rate
heart rate variability
sulfuric acid
inspiratory  capacity
intracellular adhesion molecule
International Commission on Non-Ionizing Radiation Protection
intermittent exercise
immunoglobulin (e.g., IgA, IgE, IgG, IgM)
interleukin  (e.g., IL-1, IL-6, IL-8)
inducible nitric oxide synthase; NOS-2
intraperitoneal
Intergovernmental Panel on Climate Change
infrared
intrinsic mass transfer coefficient/parameter
                                         I-xlvii

-------
Kr
KTB
LDH
LIDAR
LIS
LLJ
LOESS
LOP
LPS
LRT
LT
LT
M
M
M
MAP
MCP
MENTOR
MI
MIP
MMEF
MONICA
MPAN
MPO
mRNA
MSA
MSL
MT
mass transfer coefficient for gas phase
mass transfer coefficient for liquid phase
reaction rate constant
terminal bronchiole region mass transfer coefficient
lactic acid dehydrogenase
Light Detection And Ranging
lateral intercellular space
low-level jet
locally estimated smoothing splines
lipid ozonization products
lipopolysaccharide
lower respiratory tract; lower airways
leukotriene (e.g., LTB4, LTC4, LTD4, LTE4)
local time
gas molecule
male
maximum number of iterations
mean arterial pressure
monocyte chemotactic protein
Modeling Environment for Total Risk Studies
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
                                        I-xlviii

-------
n,N
NAAQS
NADPH
NAS
NCEA-RTP

NCICAS
ND
NEM
NF
NF-KB
NH4HSO4
NHAPS
NIST
NK
NL
NM
NMHCs
NMMAPS
NO
NO2
NO3
NOS
NOS-1
NOS-2
NOS-3
NOX
N0y

NOZ
NP
number
National Ambient Air Quality Standards
reduced nicotinamide adenine dinucleotide phosphate
Normative Aging Study
National Center for Environmental Assessment Division in Research
Triangle Park, NC
National Cooperative Inner-City Asthma Study
not detectable; not detected
National Ambient Air Quality Standards Exposure Model
national forest
nuclear factor kappa B
ammonium bisulfate
National Human Activity Pattern Survey
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
nitrate
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  and NOZ; odd nitrogen
species
difference between NOy and NOx
national park
                                         I-xlix

-------
NPP
NQOlwt
NRC
NTP
NTS
NWR
OfD)
02
03
O3*
0(3P)
OAQPS
8-OHdG
OH
OTC
OVA
Ox
6PGD
P
"90
PAF
PAN
PAR
PEL
PBPK
PCI
PE
PEF
PEM
Penh
net primary productivity
NAD(P)H-quinone oxidoreductase wild type (genotype)
National Research Council
National Toxicology Program
nucleus tractus solitarius
national wildlife refuge
electronically excited oxygen atom
ground-state oxygen
ozone
electronically excited ozone
ground-state oxygen atom
Office of Air Quality Planning and Standards
8-hy droxy-2' -deoxy guanosine
hydroxyl; hydroxy
open-top chamber
ovalbumin
odd oxygen species
6-phosphogluconate dehydrogenase
probability value
values of the 90th percentile absolute difference in concentrations
platelet-activating factor
peroxyacetyl nitrate; peroxyacetic nitric anhydride
proximal alveolar region
planetary boundary layer
physiologically based pharmacokinetic (approach)
picryl chloride
postexposure
peak expiratory flow
personal exposure monitor
enhanced pause
                                          1-1

-------
PG
PI
PM
PM25
PM
   10
PM10.2.5

PMNs
pNEM
ppb
ppbv
ppm
PPN
PRB
PSA
PUFA
PWM
QCE
r
R
R2
RCO
R'C(O)-O2
RH
RL
RO2
ROOH
ROS
RR
RRMS
RT
prostaglandin (e.g., PGD2, PGE, PGEls PGE2 PGFla, PGF2a)
probability interval
particulate matter
fine particulate matter (mass median aerodynamic diameter <2.5 jim)
combination of coarse and fine particulate matter
coarse particulate matter (mass median aerodynamic diameter between 10
and 2.5 jim)
polymorphonuclear neutrolphil leukocytes; neutrophils
Probabilistic National Ambient Air Quality Standard Exposure Model
parts per billion
parts per billion by volume
parts per million
peroxypropionyl nitrate; peroxypropionic nitric anhydride
policy relevant background
picryl sulfonic acid
polyunsaturated fatty acid
pokeweed mitogen
quasi  continuous exercise
correlation coefficient
intraclass correlation coefficient
multiple correlation coefficient
acyl
acyl peroxy
relative humidity
total pulmonary resistance
organic peroxyl; organic peroxy
organic peroxides
reactive oxygen species
ribonucleotide reductase
relatively remote monitoring sites
respiratory tract
                                          I-li

-------
SAB
SAC
SAMD
SBUV
SC
SD, S-D
SD
SES
sGAW
SHEDS
SNPs
SO2
S042
SOD
SOS
SP
SP
sRAW
STE
STRF
SUM06
SUM08
SZA
t
T3
T4
TAR
TB
TEARS
Tc-DTPA
Science Advisory Board
Staphylococcus aureus Cowan 1 strain
S-adenosyl methionine decarboxylase
solar backscattered ultraviolet radiation
stratum corneum
Spraque-Dawley (rat)
standard deviation
socioeconomic status
specific airways conductance
Simulation of Human Exposure and Dose System
single nucleotide polymorphisms
sulfur dioxide
sulfate
superoxide dismutase
Southern Oxidant Study
substance P
surfactant protein (e.g., SP-A, SP-D)
specific airways resistance
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
Mest statistical value; t statistic
triiodothyronine
thyroxine
Third Assessment Report
terminal bronchioles
thiobarbituric acid reactive substances
radiolabeled diethylenetriaminepentaacetic acid; 99mTc-DTPA
                                         I-lii

-------
T
T
XCTL
TLC
TNF
TNFR
TOMS
TRIM
TRIM Expo
TSP
TWA
UA
UNFCCC
URT
USGCRP
UV
UV-A
UV-B
UV-C
VC
VD
VE
V02max
voc
VT
Vra
W126
WMO/UNEP
WT
core temperature
cytotoxic T-lymphocytes
total lung capacity
tumor necrosis factor
tumor necrosis factor receptor
Total Ozone Mapping 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 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
vital capacity
anatomic dead space
minute ventilation; expired volume per minute
maximal oxygen uptake (maximal aerobic capacity)
volatile organic compound
volumetric penetration
volume at which 50% of an inhaled bolus is absorbed
tidal volume
terminal bronchiole region volume
cumulative integrated exposure index with a sigmoidal weighting function
World Meteorological Organization/United Nations Environmental Program
wild type
                                         I-liii

-------
 i                              EXECUTIVE  SUMMARY
 2
 3
 4      E.I   INTRODUCTION
 5            Tropospheric or "surface-level" ozone is one of six major air pollutants regulated by
 6      National Ambient Air Quality Standards (NAAQS) under the U.S. Clean Air Act.  As mandated
 7      by the Clean Air Act, the U.S. Environmental Protection Agency (EPA) must periodically
 8      review the scientific bases (or "criteria") for the various NAAQS by assessing newly available
 9      scientific information on a given criteria air pollutant.  This draft document, Air Quality Criteria
10     for Ozone and other Photochemical Oxidants, is an updated revision of the 1996 Ozone Air
11      Quality Criteria Document (O3 AQCD) that provided scientific bases for the current O3 NAAQS
12      set in 1997.
13
14      E.I.I   Clean Air Act Legal Requirements
15            Clean Air Act (CAA)  Sections 108 and 109 govern establishment, review, and revision
16      of U.S. National Ambient Air Quality Standards (NAAQS).
17      »D Section 108 directs the U.S. Environmental Protection Agency (EPA) Administrator to list
18         ubiquitous (widespread) air pollutants that may reasonably be anticipated to endanger public
19         health or welfare and to issue air quality criteria for them.  The air quality criteria are to
20         reflect the latest scientific information useful in indicating the kind  and extent of all
21         exposure-related effects on public health and welfare expected from the presence of the
22         pollutant in the ambient air.
23
24      »D Section 109 directs the EPA Administrator to set and periodically revise, as appropriate, two
25         types of NAAQS: (a) primary NAAQS to protect against adverse health effects of listed
26         criteria pollutants among sensitive population groups, with an adequate margin of safety, and
27         (b) secondary NAAQS to protect against welfare effects (e.g., impacts on vegetation, crops,
28         ecosystems, visibility, climate, man-made materials, etc.).  Section  109 also requires peer
29         review of the NAAQS and their underlying scientific bases by the Clean Air Scientific
30         Advisory Committee (CASAC), a committee of independent non-EPA experts.


        August 2005                              E-l         DRAFT-DO NOT QUOTE OR CITE

-------
 1     E.1.2 Chronology of Ozone NAAQS Revisions
 2          In 1971, the U.S. EPA set primary and secondary standards for total photochemical
 3     oxidants. However, based on the criteria review completed in 1978, the original primary and
 4     secondary NAAQS set in 1971 were revised in 1979 to focus on O3 as the indicator for new
 5     primary and secondary standards that would be attained when the expected number of days per
 6     calender year with maximum  1-h average O3 concentrations >0.12 ppm did not exceed one.  The
 7     NAAQS for ambient O3 were revised in 1997 by replacing the 1-h standards with an 8-h primary
 8     standard that is met when the  3-year average of the annual fourth highest daily maximum 8-h
 9     average concentration is <0.08 ppm. The 1997 primary NAAQS was based on scientific data
10     from controlled human exposure, laboratory animal, and epidemiological studies and associated
11     analyses presented in the 1996 O3 AQCD and in the 1996 O3 Staff Paper (U.S. Environmental
12     Protection Agency, 1996b).
13
14     •DThis revised O3 AQCD is now being prepared by ORD's National Center for Environmental
15        Assessment (NCEA) to support EPA's ongoing Congress!onally-mandated periodic review
16        of O3 NAAQS under a consent decree (court-ordered) schedule that calls for issuance of the
17        revised AQCD in final form by February 28, 2006.  This document assesses the latest
18        available scientific information (published mainly through December 2004) judged to be
19        useful in deriving criteria as scientific bases for decisions on possible revision of the
20        current O3 NAAQS.
21
22     »DA separate EPA O3 Staff Paper will  draw upon key findings/conclusions from this document,
23        together with other analyses, to develop and present options for consideration by the EPA
24        Administrator regarding review and possible revision of the O3 NAAQS.
25
26     E.1.3  Document Organization and Structure
27        Volume I of this document consists  of the present Executive Summary and eleven main
28     chapters of this revised O3 AQCD.  Those main chapters focus primarily on interpretative
29     evaluation of key information, whereas more detailed  descriptive summarization of pertinent
30     studies and/or supporting analyses are provided in accompanying annexes. Volume II contains
31     the annexes for Chapters 4 through 7, whereas Volume III contains the annex for Chapter 9.

       August 2005                             E-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1          The topics covered in the main chapters are as follows:
 2      •DThis Executive Summary summarizes key findings and conclusions from Chapters 1 through
 3        1 1 of this revised O3 AQCD.
 4
 5      »D Chapter 1 provides a general introduction, including an overview of legal requirements,
 6        the chronology of past revisions of O3-related NAAQS, and orientation to the structure of
 7        the document.
 8
 9      •DChapters 2 and 3 provide background information on atmospheric chemistry/physics of O3
10        formation, air quality, and exposure aspects to help to place ensuing discussions of O3 health
1 1        and welfare effects into perspective.
12
13      »D Chapters 4 through 7 then assess dosimetry aspects, experimental (controlled human exposure
14        and laboratory animal) studies, and epidemiologic (field/panel; other observational) studies.
15
16      •DChapter 8 provides an integrative synthesis  of key findings and conclusions derived from the
17        preceding chapters with regard to ambient O3 concentrations, human exposures, dosimetry,
18        and health effects.
19
20      •DChapter 9 deals with effects of O3 on vegetation, crops, and natural ecosystems, whereas
21
22
                                       3            ,      ,                     ,
          Chapter 10 evaluates tropospheric O3 relationships to alterations in surface-level UVB flux
          and climate change and Chapter 1 1 assesses materials damage (these all being key types of
23        welfare effects of relevance to decisions regarding secondary O3 NAAQS review).
24
25
26     E.2   ATMOSPHERIC CHEMISTRY AND PHYSICS OF TROPOSPHERIC
27            OZONE FORMATION
28          Key findings/conclusions from Chapter 2 regarding the chemistry and physics of surface-
29     level O3 formation include the following:
30
       August 2005                               E-3         DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Ozone (O3) is a secondary pollutant formed by atmospheric reactions involving two classes
 2         of precursor compounds, volatile organic compounds (VOCs) and nitrogen oxides (NOX).
 3         Carbon monoxide also contributes to O3 formation.
 4
 5      •  The formation of O3 and associated compounds is a complex, nonlinear function of many
 6         factors, including the intensity and spectral distribution of sunlight; atmospheric mixing and
 7         other atmospheric processes; and the concentrations of the precursors in ambient air.
 8
 9      •  The photochemical oxidation of almost all anthropogenic and biogenic VOCs is initiated by
10         reaction with hydroxyl (OH) radicals.  At night, when they are most abundant, NO3 radicals
11         oxidize alkenes. In coastal and other select environments, Cl and Br radicals can also initiate
12         the oxidation of VOCs.
13
14      •  In urban areas, basically all classes of VOCs (alkanes, alkenes, aromatic hydrocarbons,
15         carbonyl compounds, etc.) and CO are important for ozone formation. Although knowledge
16         of the oxidative mechanisms of VOCs has improved over the past several years, gaps in
17         knowledge involving key classes, such as aromatic hydrocarbons, still remain. For example,
18         only about half of the carbon initially present in aromatic hydrocarbons in smog chamber
19         studies form compounds that can be identified.
20
21      •  In addition to gas phase reactions, reactions also occur on the surfaces of or within cloud
22         droplets and airborne particles.  Most of the well-established multiphase reactions tend to
23         reduce the rate of O3 formation in polluted environments. Direct reactions of O3 and
24         atmospheric particles appear to be too  slow to reduce O3 formation significantly at typical
25         ambient PM levels.
26
27      •  Oxidants other than O3 are found in the gas phase and in particles. The chemistry  occurring
28         in particle bound-water and, hence, the mechanisms leading to the formation of reactive
29         oxygen species in particles are largely unknown.
30
        August 2005                              E-4         DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Our basic understanding of meteorological processes associated with summertime O3
 2         episodes has not changed over the past several years. However, the realization is growing
 3         that long-range transport processes are important for determining O3 concentrations at the
 4         surface . In addition to synoptic scale flow fields, nocturnal low-level jets are capable
 5         of transporting pollutants hundreds of km from their sources in either the upper boundary
 6         layer or the lower free troposphere. Turbulence then brings O3 and other pollutants to
 7         the surface.
 8
 9      •  Even in the absence of photochemical reactions in the troposphere, some O3 would be found
10         near the earth's surface due to its downward transport from the stratosphere.  Intrusions of
11         stratospheric O3 that reach the surface are rare. Much more common are intrusions that
12         penetrate to the middle and upper troposphere. However, O3 transported to the middle and
13         upper troposphere can still affect surface concentrations through various mechanisms that
14         mix air between the planetary boundary layer and the free troposphere above.
15
16      •  Chemistry transport models are used to improve understanding of atmospheric chemical and
17         physical processes, as well as to develop air pollution control strategies. The performance of
18         these models must be evaluated by comparison with field data as part of an iterative cycle of
19         model improvement and subsequent evaluation. Discrepancies between model predictions
20         and observations  can be used to point out gaps in current understanding and thus to improve
21         parameterizations of atmospheric chemical and physical processes.
22
23      •  Model evaluation does not merely involve a straightforward comparison between model
24         predictions and observed concentration fields of a pollutant of interest (e.g., O3).  Such
25         comparisons may not be meaningful because it is difficult to determine if agreement between
26         measurements and model predictions  truly represents an accurate treatment of physical and
27         chemical processes in the model or the effects of compensating errors in model routines.
28
29      •  The main methods currently in use for routine monitoring of ambient ozone are based on
30         chemiluminescence or UV absorption. Measurements at most ambient monitoring sites are
31         based on UV absorption. Both of these methods are subject to interference by other

        August 2005                               E-5         DRAFT-DO NOT QUOTE OR CITE

-------
 1         atmospheric components.  Studies conducted in Mexico City and in a smog chamber have
 2         found positive interference, but a few studies conducted in urban plumes have not found
 3         evidence for significant positive interference in the UV absorption technique.
 4
 5
 6     E.3   ENVIRONMENTAL DISPERSAL, AMBIENT CONCENTRATIONS,
 7            AND HUMAN EXPOSURE TO OZONE
 8          Key findings/conclusions derived from Chapter 3 are as follows:
 9     •   Ozone is monitored in populated areas in the United States during "ozone seasons", which
10         vary in length depending on location. All monitors should be operational from May to
11         September. However, in many areas, O3 is monitored throughout the year.
12
13     •   The median of the mean daily maximum 8-h average O3 concentration from May to
14         September 2000 to 2004 across the U.S. was 0.049 ppm on a countywide average basis.
15         Ninety five per cent of countywide mean daily maximum 8-h average O3 concentrations were
16         less than 0.057 ppm for the same period. Because most monitors are located in the East,
17         these values should not be taken to represent conditions across the country.
18
19     •   The daily maximum  1-h O3 concentrations tend to be much higher in large urban areas or in
20         areas downwind of large urban areas. For example, daily maximum 1-h O3 concentrations in
21         Houston, TX approached 0.20 ppm during the same period.
22
23     •   Daily maximum 8-h  average O3 concentrations are lower than, but are highly correlated
24         with, 1-h daily maximum O3 concentrations. For example, in the Baltimore, MD area, the
25         correlation coefficient between the two quantities was 0.98 for data obtained from May to
26         September 1994 to 2004.
27
28     •   Within individual MSAs, O3 tends to be well correlated across monitoring sites. However,
29         there can be substantial spatial variations in concentrations.  Ozone in city centers tends to be
30         lower than in regions either upwind or downwind of the center, because of titration by NO
31         emitted by motor vehicles.

       August 2005                             E-6        DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Ozone concentrations tend to peak in early- to mid-afternoon in areas where there is strong
 2         photochemical production and later in the day in areas where transport is more important in
 3         determining O3 abundance.
 4
 5      •  Summertime maxima in O3 concentrations occur in areas in the United States where there is
 6         substantial photochemical activity involving O3 precursors emitted from human activities.
 7         Maxima can occur anytime from June through August.
 8
 9      •  Springtime maxima are observed in relatively remote sites in the western United States and
10         at various other relatively unpolluted sites throughout the Northern Hemisphere. Relatively
11         high O3 concentrations can also be found during winter in several cities throughout the
12         southern United States.
13
14      •  Long-term trends in O3 concentrations reflect notable decreases over time throughout the
15         United States, with decreases nationwide of approximately 29% in 2nd highest 1-h  O3
16         concentrations from 1980 to 2003 and of about 21% in 4th highest 8-h O3 concentrations
17         during the same time period.
18
19      •  These trends include dramatic decreases from peak 1-h O3 levels of 0.4 to 0.6 ppm seen in
20         the Los Angeles area at times in the late 1950's to 1970's to current peak levels of 0.17 ppm
21         and 0.15 ppm (1-h and 8-h avg, respectively) seen in the Los Angeles basin during
22         2000-2003.
23
24      •  Downward trends in the upper tail of the O3 concentration distribution do not reflect trends
25         for O3 values towards the center of the O3 concentration distribution nationwide. These latter
26         concentrations have remained more or less constant, and O3 values in the lower tail of the
27         distribution show some evidence of slight increases.
28
29      •  Policy relevant background (PRB) O3 concentrations are used for assessing risks to human
30         health associated with O3 produced from anthropogenic  sources in the United States, Canada
        August 2005                              E-7        DRAFT-DO NOT QUOTE OR CITE

-------
 1         and Mexico. Because of the nature of the definition of PRB concentrations, they cannot be
 2         derived from observations directly, instead they must be derived from model estimates.
 3
 4      •  Current model estimates indicate that PRB O3 concentrations in the United States surface air
 5         are generally 0.015 ppm to 0.035 ppm.  Such concentrations decline from spring to summer
 6         and are generally <0.025 ppm under conditions conducive to high O3 episodes.  PRB Ozone
 7         concentrations may be higher, especially at elevated sites during the spring, due to enhanced
 8         contributions from (a) pollution sources inside and outside North America and
 9         (b) stratospheric O3 exchange.
10
11      •  Sufficient data for other oxidants (e.g., H2O2, PAN) and oxidation products (e.g., HNO3,
12         H2SO4) in the atmosphere are not available for use in epidemiologic time series studies.
13         Limited data for oxidants besides O3 in the gas and particle phases suggest that their
14         combined concentrations are probably <10 % that of O3.
15
16      •  Relationships between O3 and PM2 5 are complex, in part because PM is not a distinct
17         chemical species, but is a mix of primary and secondary species. For example,  PM25
18         concentrations were positively correlated with O3 during summer, but negatively correlated
19         with O3 during the winter at Ft. Meade, MD. Similar relationships were found for PM10 and
20         O3 in data collected in a number of urban areas during the 1980s.
21
22      •  Humans are exposed to O3 either outdoors or in various microenvironments. Ozone in
23         indoor environments results mainly from infiltration from outdoors.  Once indoors, O3 is
24         removed by deposition on and reaction with surfaces and reactions with other pollutants.
25         Hence, O3 levels indoors tend to be notably lower than outdoor O3 concentrations measured
26         at nearby monitoring sites, although the indoor and ambient O3 concentrations tend to vary
27         together (i.e., the higher the ambient, the higher the indoor O3 levels).
28
29      •  Personal exposure to O3 tends to be positively associated with time spent outdoors.
30         Although O3 concentrations obtained at stationary monitoring sites may not explain the
        August 2005                               E-8        DRAFT-DO NOT QUOTE OR CITE

-------
 1         variance in individual personal exposures, they appear to serve reasonably well as surrogate
 2         measures for aggregate personal exposures.
 3
 4      •  Atmospheric reactions between O3 and certain other ambient airborne contaminants, e.g.,
 5         terpenes emitted by vegetation or wood products, contribute to generation of ultrafine
 6         particles, with formation of such particles being observed in both urban and rural areas.
 7         These reactions also occur in indoor environments and involve O3 infiltrating from outdoors
 8         and terpenes emitted by household products (e.g., air fresheners).  Gaseous products
 9         resulting form such reactions may also be toxic.
10
11
12      E.4    DOSIMETRIC STUDIES
13           Chapter 4 discusses dosimetric issues, including factors that are important to consider in
14      attempting animal-to-human extrapolations of experimentally-induced O3 effects.
15
16      •  Dosimetric studies seek to quantify dose and factors affecting the dose of O3 and/or its active
17         metabolites at specific lung regions, target tissues, or cells.
18
19      •  In both humans and animals, the efficiency of O3 uptake is greater in the nasal passages than
20         the  oral pathway.  In the lower respiratory tract, increasing tidal volume increases O3 uptake,
21         whereas increasing flow or breathing frequency decreases O3 uptake. As flow is increased,
22         O3 uptake shifts to the smaller peripheral airways.
23
24      •  In adult human females relative to males, the smaller airways and associated larger surface-
25         to-volume ratio enhance local O3 uptake and cause somewhat reduced penetration of O3 into
26         the  distal lung.  However, it is not clear from these findings if the actual anatomical location
27         of O3 uptake differs between males and females.
28
29      •  Similarly exposed individuals vary in the amount of actual dose received, but O3 uptake is
30         not  predictive of intersubject variability in FEVj.
31

        August 2005                              E-9        DRAFT-DO NOT QUOTE OR CITE

-------
1      •  The efficiency of O3 uptake is chemical-reaction rate dependent and the reaction products
2         (hydrogen peroxide, aldehydes, and hydroxyhydroperoxides) created by ozonolysis of lipids
3         in ELF and cell membranes appear to mediate O3 toxicity.
4
5
6
7
9
        •  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
           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.

10      •  New experimental work in rats suggests that the primary site of acute O3-induced cell injury
11         is the conducting airways, whereas prior modeling studies suggested that the proximal
12         alveolar and centriacinar regions may be principal O3 target sites.
13
14      •  In most clinical studies, humans are exposed to O3 during exercise. Under these conditions,
15         the switch from nasal to oral breathing, coupled with increases in respiratory flow (as occurs
16         during exercise), causes a shift in the O3 dose distribution, thusly allowing O3 to penetrate
17         deeper into the lung and thereby increasing the potential for damage to bronchiolar and
18         alveolar tissues.
19
20      •  Comparisons of acute exposures in rats and humans suggest that, though both species have
21         similar qualitative responses to O3 exposure, there are interspecies mechanistic disparities
22         that necessitate careful comparisons of dose-response relationships.  Currently available data
23         suggest that lowest observable effect levels in resting rats are approximately 4- to 5-fold
24         higher than for exercising humans for toxicological endpoints, including BAL protein and
25         BAL PMNs.
26
27
28      E.5   ANIMAL TOXICOLOGY ASPECTS
29         Key toxicology findings/conclusions from laboratory animal studies discussed in Chapter 5
30      include:
31

        August 2005                              E-10         DRAFT-DO NOT QUOTE OR CITE

-------
 1      E.5.1  Respiratory Tract Effects of Short-Term Exposures to Ozone
 2           In general, O3 concentration and duration of exposure (C and T), respectively, determine
 3      the dose and resultant health effects of O3. Concentration usually dominates the response, with
 4      the impact of T being C-dependent (at higher Cs, the impact of T tends to be greater).
 5
 6      Effects on Pulmonary Function
 1      •  Rapid shallow breathing, which is not protective, but does cause a more evenly distributed
 8         injury pattern, is the most common change in pulmonary function induced by acute (1-8 h)
 9         O3 exposure of-0.2 ppm.  Decreased lung volumes are observed in rats with acute exposures
10         at levels of 0.5  ppm. Breathing mechanics (compliance and resistance) are affected at
11         exposures of-1.0 ppm.
12
13      •  Attenuation of pulmonary function decrements occurs with 5 days of repeated acute O3
14         exposures, which are not accompanied by concurrent attenuation of lung injury and
15         morphological changes, indicating that the attenuation does not result in protection against
16         all the effects of O3.
17
18      •  Ozone-induced airway hyperresponsiveness (AHR) occurs in laboratory animals with acute
19         exposures (< 1 h) in the range of 0.5 to 1.0 ppm. Animal studies have shown that O3
20         exposure can augment OVA-induced AHR. A temporal relationship exists between
21         inflammatory cell influx and O3-induced AHR, but inflammation is not a prerequisite of
22         AHR. Repeated O3 exposures enhance AHR, possibly by modulating rapidly adapting
23         airway receptors or by altering the structure of conducting airways. In human asthmatics,
24         AHR appears to be due, in part, to chronic inflammation and airway remodeling.
25
26      •  Studies using repeated O3 exposure (<0.3 ppm) of nonsensitized laboratory animals have
27         shown equivocal results. A few studies in sensitized laboratory animals are consistent with
28         the O3-induced exacerbation of AHR reported in atopic humans with asthma.  However,
29         extrapolation of these data is difficult due to interindividual and interspecies differences in
30         responsiveness to bronchoprovocation and possible adaptation of airway responsiveness with
31         long-term, repeated O3 exposures.

        August 2005                              E-11       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Other Respiratory Tract Effects of Ozone
 2      •  Due to its high reactivity, O3 penetrates only about 0.1 to 0.2 jim into the extracellular lining
 3         fluid (ELF) of the respiratory tract. Ozone interacts with a wide range of components in ELF
 4         that include polyunsaturated fatty acids, cholesterol, amino acid residues, reduced
 5         glutathione, uric acid, vitamins C and E, and free amino acids.  Ozone's toxicity is dependent
 6         upon a cascade of reaction products, including ozonide, aldehyde, and hydroperoxide.
 7         Saturated phospholipids are thought to reduce the local dose and limit site-specific cell injury
 8         from O3 exposure.
 9
10      •  Antioxidants present in ELF act to protect lung tissue from O3-induced injury, but even with
11         environmentally relevant exposures, the reactivity of O3 is not quantitatively fully quenched.
12         Thus, cell injury occurs in both the upper and lower respiratory tract.  Short-term exposures
13         to <1 ppm O3 increase antioxidant metabolism.  Previous O3 exposure does not appear to be
14         protective upon re-exposures.
15
16      •  Both short- and long-term exposures to O3 have been shown to enhance lung xenobiotic
17         metabolism, possibly due to changes in the number and function of bronchiolar epithelial
18         Clara cells and alveolar epithelial Type 2 cells.  Elevations in enzyme activity appear to
19         increase as a function of age, suggesting that O3 exposure can cause greater lung injury in the
20         older animal.  Some studies found an effect on liver xenobiotic enzymes with exposure to O3
21         concentrations as low as 0.1 ppm, whereas others did not detect alterations in metabolic
22         enzymes even at 1 ppm, the effects appearing to be highly species-specific.
23
24      •  Acute exposures of 0.1 ppm O3 disrupt the barrier created by airway mucosa in the normal
25         lung, resulting in an increase in serum proteins, bioactive mediators, and neurophils in the
26         interstitium and air spaces of the lung. In rats, a single 3 h exposure to 0.5 ppm O3 produces
27         a significant increase in both lung permeability and inflammation.  Ozone-induced
28         permeability changes appear to occur predominantly in the trachea and bronchioalveolar
29         regions compared to nasal passages. Species differences exist in responses, with guinea pigs
30         being the most responsive; rabbits the least; and rats, hamsters, and mice intermediate.  With
31         continuing exposure, the increases in BALF protein and PMNs typically peak after a few

        August 2005                               E-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1         days and return toward control levels even with continuing exposure.  Though inflammation
 2         and increased permeability occur somewhat concurrently, they are distinct events controlled
 3         by independent mechanisms.
 4
 5      •  Important mechanisms of O3-induced inflammation and injury involve inflammatory
 6         cytokines and chemokines, which are released as a result of stimulation or injury of
 7         macrophages, epithelial cells and PMNs.  In vitro exposures to O3 induce release of the
 8         cytokines IL-6, TNF-a, IL-lp, and IL-8. In vivo exposures of O3 induce release of MTP-2,
 9         IL-6, MTP-la, CINC, eotaxin and fibronectin. Studies utilizing antibodies to selected pro- or
10         anti-inflammatory cytokines suggest a role of TNF-a, interleukin-10 (IL-10) and IL-lp in
11         O3-induced changes in permeability, inflammation and cytokine release.
12
13      •  Cell adhesion molecules (e.g., ICAM-1) and extracellular matrix proteins (e.g., fibronectin)
14         modulate O3-induced lung inflammation and injury. Ozone exposure also affects
15         macrophage functions by increasing their production of nitric oxide, superoxide anion
16         andPGE2.
17
18      •  Mucociliary clearance is affected in most test species at just under 1 ppm, with lower levels
19         (-0.1 ppm) increasing clearance and somewhat higher levels decreasing clearance.  At O3
20         exposures of 0.1 to 1.2 ppm, alveolar macrophage (AM) function is disrupted and the
21         number of AM are increased. Ozone exposures are linked to decreased resistance to
22         microbial pathogens.
23
24      •  Ozone exposures can enhance or suppress immune responsiveness, depending on the species
25         studied, the concentration of O3, the route of exposure of allergen, and exposure timing.
26         Continuous exposure to O3 impairs immune responses for the first several days of exposure,
27         followed by an adaptation to O3 that allows a return of normal immune responses.  Most
28         species show little effect of O3 exposures prior to immunization, but exhibit suppression of
29         responses to antigen with O3 exposures post-immunization.  Ozone exposures are linked to a
30         possible interaction between the innate and acquired immune system and a shift in the
31         immune response towards a Th-2-like pattern. Surfactant proteins A and D, which have an

        August 2005                              E-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1         immunomodulatory function in protecting against oxidative stress, are affected by O3
 2         exposures. Ozone exposures at levels of 0.1 to 1 ppm O3 for 1 week have been shown to
 3         cause, in general, increased mortality and morbidity, decreased clearance, increased bacterial
 4         growth, and increased severity of infection at exposure.
 5
 6      •  Age, gender, nutritional status, genetic variability, exercise and exposure to co-pollutants are
 7         all factors which can impact the effects of O3.  Control of the ventilatory response to O3 is
 8         determined, at least in part, by genetic factors.  Genetic loci that modulate pulmonary
 9         responses to O3 differ from each other and from loci controlling inflammatory responses.
10         The effects of age and gender on lung inflammation are not well characterized, but exercise
11         during O3 exposure clearly generally increases susceptibility.
12
13      •  Collagen increases with O3 exposure and can persist after exposure stops. Rats exposed
14         acutely or subchronically to 0.4 ppm O3 showed centriacinar thickening of septa. Collagen
15         content decreased with postexposure recovery time but not the structural fibrotic changes in
16         ductular septa and respiratory bronchioles, suggesting that subchronic O3 exposures in rats
17         creates a progression of structural lung injury that can evolve to a more chronic form that
18         likely includes fibrosis.
19
20      •  Ozone-induced alterations in lung structure have been shown across a variety of species
21         repeatedly exposed to O3 concentrations as low as 0.15 ppm. Cells in the centriacinar region
22         (CAR) are the primary targets of O3, but ciliated epithelial cells in the nasal cavity and
23         airways and Type 1 epithelial cells in the gas exchange region  are also targeted.  Ozone-
24         induced fibrotic changes in the CAR are maximal at 3 d of exposure and recover 3 d
25         postexposure with exposures to 0.2 ppm in rodents. Rats with induced allergic rhinitis are
26         more susceptible to 0.5 ppm than are controls.  The proximal respiratory bronchiole receives
27         the most acute epithelial injury from exposures < 1 ppm, while metabolic effects are greatest
28         in the distal bronchioles and minor daughter airways.
29
        August 2005                               E-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1     E.5.2 Respiratory Tract Effects of Chronic (Long-Term) Exposures to Ozone
 2          A variety of respiratory tract effects have been shown to occur as the result of more
 3     chronic, longer-term exposures of laboratory animals to O3. Some of the more notable types of
 4     effects are as follow.
 5     »D  Chronic O3 exposures in a range of 0.5 to 1.0 ppm induce a pattern of epithelial hyperplasia
 6          which is similar to the pattern of inflammation, with a peak over the first few day, a drop,
 7          and then disappearance.  In contrast, fibrotic changes in lung tissue increase very slowly
 8          over months of exposure, and, after exposure ceases, the changes sometimes persist or
 9          increase.  Compared to continuous exposure regimens, seasonal episodic exposures
10          demonstrated remodeling in the distal airways, abnormalities in tracheal basement
11          membrane, eosinophil accumulation in conducting airways, and decrements in airway
12          innervation. Also, long-term O3 exposures have demonstrated that repeated daily exposure
13          of rats to an episodic profile of O3 caused small, but significant decrements in lung
14          function that were consistent with early indicators of focal fibrogenesis in the proximal
15          alveolar region, without overt fibrosis.
16
17     E5.3   Other Types of Ozone Exposure Effects Observed in Laboratory
18             Animal Models
19     Systemic Effects of Ozone
20     •   Decreased heart rate, core temperature, and blood pressure, all collectively termed the
21          hypothermic response, are other types of effects observed at concentrations of 0.3 to
22          0.5 ppm.  Concentrations of O3  >0.5 ppm cause tissue edema (possibly mediated by atrial
23          natriuretic factor). Additionally, O3-induced production of platelet-activating factor and
24          oxysterols suggest mechanisms of cardiovascular injury.
25
26     •   Neurobehavioral effects attributed to O3 exposure (0.2 to 1.0 ppm) include decreased motor
27          activity, short- and long-term memory deficits, increased freezing behavior,  and decreased
28          exploratory behaviors. Near-ambient exposures to O3 elicit neuroendocrine  effects,
29          including morphological and hormonal changes in the pituitary-thyroid-adrenal axis and
30          alterations of visual and olfactory  neural  pathways.
       August 2005                             E-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1      •    No noticeable neurobehavioral or somatic effects have been observed with prenatal
 2           exposures of < 1.0 ppm. Effects on neonatal mortality are observed with exposures of 1.0
 3           to 1.5 ppm. Effects on spleen and thymus appear to only occur at high O3 concentrations
 4           (>1.0 ppm), whereas relevant urban ambient exposures have no effect on systemic immune
 5           function in rats.
 6
 7      Genotoxicity Potential of Ozone
 8      •    The weight of evidence from new experimental studies, utilizing non-lifetime exposures,
 9           does not appear to support ambient O3 as a pulmonary  carcinogen in laboratory animal
10           models.  New data are in agreement with the 1994 National Toxicology Program
11           evaluation of O3 carcinogenicity.  However, O3 could possibly act as a co-carcinogen
12           functioning to stimulate hyperplasia.
13
14      Interactions of Ozone with Other Co-occurring Pollutants
15      •    Bases for toxic interactions of O3 with co-occurring pollutants may include: adsorption
16           of O3 onto a co-pollutant with transport to another site;  production of lexicologically
17           active secondary products; biological or chemical alterations at target sites that affect
18           response to O3 or the co-pollutant; O3- or co-pollutant-induced physiological change, such
19           as alteration in ventilation pattern, resulting in changes in the penetration or deposition of
20           one pollutant when another is present.
21
22      •    Generalizations regarding interactions of O3 and co-pollutants include:  interactions of
23           O3-containing mixtures are generally synergistic; O3 may produce more significant
24           biological responses as a component of a mixture than  when inhaled alone; and, although
25           most studies  have shown that interaction occurs only at higher than ambient concentrations
26           with acute exposure, some have demonstrated interactions at more environmentally
27           relevant levels (e.g., 0.05 to 0.1 ppm O3 with NO2) and with repeated exposures.
28
29      Effects of Other Photochemical Oxidants
30      •    Ambient concentrations of the most abundant non-O3 oxidants (peroxyacetyl nitrate,
31           peroxypropionyl  nitrate, and H2O2) have not been  shown as being likely to cause adverse

        August 2005                               E-16       DRAFT-DO NOT QUOTE OR CITE

-------
 1           health effects. However, as constituents of ambient air mixes, other ambient oxidants may
 2           contribute to some effects attributed to O3.
 3
 4
 5      E.6   CONTROLLED HUMAN EXPOSURE STUDIES
 6           Key findings/conclusions derived from Chapter 6 assessment of experimental human
 7      studies include:
 8      •  Responses in humans exposed to ambient O3 concentrations include decreased inspiratory
 9         capacity; mild bronchoconstriction; rapid, shallow breathing pattern during exercise; and
10         symptoms of cough and pain on deep inspiration.  Ozone exposure also results in airway
11         hyperresponsiveness, inflammation, immune system activation, and epithelial injury.
12
13      •  Young healthy adults exposed to O3 concentrations of 0.08 ppm develop significant
14         reversible, transient decrements in pulmonary function if minute ventilation or exposure
15         duration are increased sufficiently.  Healthy children experience similar spirometric
16         responses but lesser symptoms from O3 exposure relative to young adults.  On average,
17         spirometric and symptom responses to O3 exposure appear to decline with increasing age
18         beyond approximately 18 years of age.
19
20      •  There is a tendency for slightly increased spirometric responses in mild asthmatics and
21         allergic rhinitics relative to healthy young adults.  Spirometric responses in asthmatics appear
22         to be affected by baseline lung function.
23
24      •  There is a large degree of intersubject variability in physiologic and symptomatic responses
25         of adults exposed to O3.  However, responses tend to be reproducible within a given
26         individual  over a period of several months. With increasing O3 concentration, the
27         distribution of FEVj decrements becomes asymmetrical with a few individuals experiencing
28         large decrements.  An individual's innate susceptibility to ozone may be linked to the genetic
29         background of an individual.  Additional studies, however, are needed to ascertain the link
30         between susceptibility and polymorphisms.
31

        August 2005                              E-17       DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Soluble mediators of inflammation such as the cytokines (IL-6, IL-8) and arachidonic acid
 2         metabolites (e.g., PGE2, PGF2", thromboxane, and leukotrienes [LTs] such as LTB4) have
 3         been measured in the BAL fluid of humans exposed to O3. There appears to be no strong
 4         correlation between any of the measured cellular and biochemical changes and changes in
 5         pulmonary function. A limited number of studies suggest that inflammatory responses may
 6         be detected following O3 exposures that are insufficient to cause decrements in pulmonary
 7         function.
 8
 9      •  With repeated O3 exposures over several days, spirometric and symptom responses become
10         attenuated, but this tolerance is lost after about a week without exposure.  Some markers of
11         airway inflammation and small airways dysfunction may not be attenuated by repeated O3
12         exposures.
13
14      •  An initial phase of recovery from O3 exposure in healthy individuals proceeds relatively
15         rapidly, with acute spirometric and symptom responses resolving within about 2 to 4 h.
16         Effects on the small airways, assessed by decrements in FEF25.75 and altered ventilation
17         distribution at and possibly beyond 24 h, may be partly due to inflammation.  Some
18         inflammatory and cellular changes may persist for up to 48 h, but the time course for these
19         parameters in humans has not been explored fully.
20
21
22      E.7   EPIDEMIOLOGIC STUDIES
23          Many epidemiologic studies, as discussed in Chapter 7, have shown associations of acute
24      exposure to ambient O3 with a variety of human health endpoints, including pulmonary function,
25      respiratory symptoms, hospital admissions, and mortality. Key findings and conclusions
26      regarding O3 health effects drawn from the epidemiologic evidence and the issues that may
27      affect the interpretation of the effect estimates can be briefly summarized as follows.
28
29      E.7.1  Health Effects Associated with Acute Ozone Exposures
30      •  Field/panel studies of acute O3 effects.  Recent field/panel studies continue to confirm that
31         short-term O3 exposure is associated with acute decrements in lung function and increased

        August 2005                              E-18       DRAFT-DO NOT QUOTE OR CITE

-------
 1         respiratory symptoms, particularly in children and asthmatics. There is also suggestive
 2         evidence that O3 is related to increased asthma medication use.  Taken together with the
 3         evidence from controlled human exposure studies, O3 is likely causally related to the various
 4         respiratory health outcomes. The current evidence is limited but supportive of a potential
 5         effect of O3 on heart rate variability, ventricular arrhythmias, and the incidence of
 6         myocardial infarctions.
 7
 8      •  Acute O? effects on emergency department visits and hospitalizations. Large multicity
 9         studies, as well as many studies from individual cities have reported an association of short-
10         term O3 concentrations with respiratory and cardiovascular hospital admissions. Studies
11         using year-round data noted some inconsistencies in the O3 effect on daily hospitalizations.
12         However, studies with data restricted to the summer or warm season, in general, indicated
13         positive and robust associations between short-term (e.g., 1 h or 8 h) ambient O3
14         concentrations and cardiopulmonary hospital admissions. Results for emergency department
15         visits are less consistent.
16
17      •  Acute (X effects on mortality.  The majority of the studies suggest that an elevated risk of
18         all-cause mortality is associated with acute exposure to O3, especially in the summer or warm
19         season when O3 levels are typically high.  Slightly greater O3 effects were observed for
20         cardiovascular mortality. Results from a recent, large U.S. multicity time-series study
21         provide the strongest evidence to-date for acute O3 exposure effects on mortality.  Recent
22         meta-analyses also showed consistent risk estimates that are unlikely to be confounded by
23         PM; however, future work is needed to better understand the influence of model
24         specifications on the risk coefficient.
25
26      •  Age-related differences in O? health effects.  Supporting evidence exists for heterogeneity in
27         the effects of O3 by age. The elderly population (>65 years of age) appear to be at greater
28         risk of O3-related hospitalizations and mortality compared to all age or younger populations.
29         In addition, potentially adverse respiratory health outcomes were associated with O3
30         exposure in children (<18 years of age).
31

        August 2005                               E-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Ozone health effects in asthmatics. The effects of O3 on asthmatics have been examined
 2         widely in both time-series studies and panel studies. Associations of O3 with various
 3         respiratory health outcomes (including lung function declines, increased respiratory
 4         symptoms, and emergency department visits) were observed. These findings, along with the
 5         pathophysiologic understanding of asthma as a chronic inflammatory disease, indicate that
 6         asthmatics may be a notably susceptible population affected by O3 exposures.
 7
 8      E.7.2   Issues Potentially Affecting Interpretation of Acute Exposure Studies
 9      •  Exposure assessment. Exposure misclassification may result from the use of stationary
10         ambient monitors to determine exposure in population studies.  Although central ambient
11         monitors do not explain the variance of individual personal exposures, significant
12         correlations are found between aggregate personal O3 measurements and O3 concentrations
13         from ambient monitors. A simulation study indicated that the use of ambient monitor data
14         will tend to underestimate the O3 effect. Better understanding of the factors that affect the
15         relationship between ambient concentrations and personal exposures should help to improve
16         interpretation of the O3 effect estimates.
17
18      •  Ozone exposure indices. The three most commonly used daily O3 exposure indices,  1-h max
19         O3, 8-max O3, and 24-h avg O3, were found to be highly correlated in studies conducted in
20         various regions. In addition, effect-size estimates and significance of associations across all
21         health outcomes were comparable when using standardized distributional increments of
22         40 ppb, 30 ppb, and 20 ppb for 1-h max O3,  8-h max O3, and 24-h avg O3, respectively.
23
24      •  Lag structures for (X exposure and effect. The lag time between O3 exposure and effect may
25         differ depending on various factors such as the specific health outcome of interest, the
26         mechanism of effect, and preexisting health conditions.  The majority of the studies found an
27         immediate O3 effect, with the strongest associations observed between health outcomes and
28         O3 exposure on the same day and/or previous day.  Some studies found large cumulative
29         effects of O3 over longer lag periods,  indicating that multiday lags also may be relevant for
30         some health outcomes, including mortality.
31

        August 2005                              E-20        DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Sensitivity to model specifications for temporal trends.  Ozone effect estimates that were
 2         reported in studies whose main focus was PM often were calculated using the same model
 3         specifications as PM.  While the sensitivity of the O3 risk estimates to alternative model
 4         specifications has not been throughly investigated, the limited available evidence indicates
 5         that O3 effects appear to be robust to various model specifications for temporal trend
 6         adjustment.
 7
 8      •  Influence of seasonal  factors. An evaluation of the confounding effects of meteorologic
 9         factors and copollutants on O3 risk estimates is complicated by their changing relationships
10         with O3 across seasons.  In addition, seasonal or seasonally-modified factors (e.g., air
11         conditioning use, time spent outdoors) complicate interpretation of all-year effect estimates,
12         as they affect the relationship between ambient concentrations and personal exposures.
13         Given the potentially  significant influence of season, season-specific analyses  are more
14         informative in assessing O3 health risks.
15
16      •  Confounding by copollutants.  Multipollutant regression models often are used to adjust for
17         confounding by copollutants. Although there is some concern regarding the use of
18         multipollutant models given the varying concurvity across pollutants, currently available
19         results generally suggest that the inclusion of copollutants into the models do not
20         substantially affect O3 risk estimates. These findings indicate that effects of O3 on  various
21         health outcomes are robust and independent of the effects of other copollutants.
22
23      •  Concentration-response function. In the limited mortality and morbidity studies that have
24         specifically examined the O3 concentration-response relationship, the evidence is
25         inconclusive regarding the detection of any clear effect threshold. Factors such as  exposure
26         measurement error may reduce the ability to detect a threshold in population studies.
27
28      •  Heterogeneity of O3 health effects. Consistent O3 effect  estimates have generally been
29         observed for mortality, hospitalizations, and other respiratory health outcomes in multicity
30         studies. Some other reported geographic heterogeneity in effect sizes may be attributable to
31         differences in relative personal exposure to O3, which is  affected by variations in factors such

        August 2005                               E-21       DRAFT-DO NOT QUOTE OR CITE

-------
 1         as air conditioning prevalence and human activity patterns as well as the varying
 2         concentrations and compositions of copollutants present by region.
 3
 4      E.7.3   Health Effects Associated with Chronic Ozone Exposure
 5           Many fewer studies have investigated the effects of chronic O3 exposure on morbidity and
 6      mortality.  The strongest evidence is for negative seasonal effects of chronic O3 exposure on lung
 7      function in adults and children.  Less conclusive are longer-term studies investigating the
 8      association of chronic O3 exposure on yearly lung function, asthma incidence, and respiratory
 9      symptoms. Studies of potential chronic O3 exposure-mortality relationships have generally
10      observed inconsistencies across exposure periods, cause-specific mortality outcomes,
11      and gender.
12
13
14      E.8   INTEGRATIVE SYNTHESIS
15            This section summarizes key conclusions derived from the  Chapter 8 integrated synthesis
16      of information regarding health effects associated with ambient O3 exposures. The conclusions
17      were derived based on an integrated analysis of available laboratory animal, human clinical, and
18      epidemiological studies  that have evaluated health effects associated with short-term, repeated,
19      and long-term exposures to O3 alone or in combination with other ambient pollutants.  The
20      Chapter  8 synthesis utilized experimental evidence from dosimetric and human and animal
21      toxicological studies presented in Chapters 4, 5, and 6 both to evaluate the biological plausibility
22      of health effects observed in epidemiologic studies discussed in Chapter 7 and to inform
23      delineation of O3 exposure-dose-response relations and likely underlying mechanisms of action.
24      These evaluations are also aimed at identifying susceptible populations that are at potentially
25      greater risk for effects of O3 exposure.
26
27      1. Health effects of acute (short-term) exposures to Ozone
28            Numerous field panel and time-series epidemiologic studies (using better weather models
29      and adjustments to confounding copollutants than those assessed in the 1996 O3 AQCD ) have
30      evaluated the effects of short-term exposure to O3 on a wide range  of health endpoints, from lung
        August 2005                              E-22        DRAFT-DO NOT QUOTE OR CITE

-------
 1      function decrements to mortality. Results from the majority of studies continue to support the
 2      conclusions reported in the 1996 O3 AQCD.
 3      •  Panel studies typically have evaluated the effects of short-term O3 exposure on both healthy
 4         individuals and people with cardiopulmonary diseases. These evaluations included
 5         measurement of lung function changes, respiratory symptoms and use of asthma medication.
 6
 7      •  Clinical controlled exposure studies in humans indicate changes in lung function and
 8         respiratory symptoms that vary as a function of exposure concentration, duration and level
 9         of exercise.
10
11      •  Newer meta-analyses confirmed the interindividual differences in lung function decrements
12         reported in the 1996 O3 AQCD. Age-specific differences in the lung function responses were
13         also observed.  Spirometric responses (due to decrements in lung function) in healthy adults
14         exposed to near ambient O3 levels typically resolve to near baseline values within 4-6 h.
15
16      •  Meta-analyses of four controlled human exposure studies (two new and two reported in the
17         1996 O3 AQCD) reporting the effects of prolonged (6.6 h) exposures to 0.08 ppm O3 during
18         moderate exercise on pulmonary function in young healthy adults (M = 90, F = 30; mean
19         age, 23 yrs) indicate an absolute FEVj decrease of approximately 6%, whereas FEVj
20         increased by about 1% following free air (FA) exposures.
21
22      •  Recent meta-analyses on numerous clinical studies indicate interindividual differences in
23         lung inflammatory response to short-term O3 exposures.
24
25      •  Inflammatory and permeability responses also generally resolve (in some instances complete
26         recovery) within a week or two after cessation of O3 exposure, but exhibit differential
27         attenuation profiles between normal healthy subjects and people with preexisting respiratory
28         diseases. However, some lung inflammation markers may not completely resolve readily,
29         and mild persistent inflammation has been reported.
30
        August 2005                              E-23        DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Field/panel studies of healthy individuals and asthmatics have found positive associations
 2         between short-term exposure to O3 and decrements in lung function analogous to those
 3         shown by studies of controlled short-term (1 to 8 h) human exposures to O3.
 4
 5      •  Associations between short-term O3 exposures and school absenteeism (due to respiratory
 6         illness) have also been suggested.
 7
 8      •  With regard to cardiac impacts, a limited number of field studies that examined the
 9         relationship between short-term O3 exposures and cardiovascular effects (heart rate
10         variability,  myocardial infarction) suggest an association.
11
12      •  A large multicity and several single-city studies have indicated a positive association
13         between increased  O3 levels (especially during the warm season) and increased risk for
14         hospital admissions.  On the other hand epidemiologic data on emergency department visits
15         do not suggest such an association with increase in ambient O3 levels.
16
17      •  The results of two large multicity studies from the U.S. and several single-city studies
18         suggest a positive association between increases in O3  levels and all-cause (non-accidental)
19         daily mortality. Meta-analyses on the influence of season suggest  a causal association.
20         Additional  meta-analyses on cause-specific mortality are suggestive  of a likely positive
21         association between increases in ambient O3 levels and cardiovascular mortality.
22
23      •  Short-term  O3-induced lung function decrements, respiratory symptoms, inflammation and
24         permeability changes observed in animal toxicology studies are consistent with human
25         studies.
26
27      2. Health effects of repeated short-term exposures to Ozone
28           The results of new controlled human exposure studies of repeated short-term O3 exposures
29      continue to support the health effects findings/conclusions reported in  the 1996 O3 AQCD.
30      •  Repeated exposure studies at higher concentrations typically show that FEVj response to O3
31         is enhanced on the  second of several days of exposure.  Such an enhanced response was not

        August 2005                              E-24        DRAFT-DO NOT QUOTE OR CITE

-------
 1         observed at lower O3 concentrations. With repeated O3 exposures over several days,
 2         spirometric and symptom responses become attenuated, but this tolerance is lost after about a
 3         week without exposure.
 4
 5      •  In humans repeatedly exposed to 0.4 ppm O3 for 5 consecutive days, several indicators of
 6         inflammation (e.g., PMN influx, IL-6, PGE2, BAL protein, fibronectin) were attenuated after
 7         5 days of exposure.  However, lung injury and permeability markers (LDH, IL-8, total
 8         protein, epithelial cells) did not show attenuation, indicating that tissue damage probably
 9         continues to occur during repeated exposure. The recovery of the inflammatory response
10         occurred for some markers after 10 days, but some responses were not normalized even after
11         20 days.  The continued presence of cellular injury markers indicates a persistent effect that
12         may not necessarily be recognized due to the attenuation of spirometric and symptom
13         responses.
14
15      •  Repeated daily exposure to lower concentrations of O3 (0.125 ppm for 4 days) causes an
16         increased response to bronchial allergen challenge in subjects with preexisting allergic
17         airway disease, with or without asthma. In these subjects, changes in airway responsiveness
18         after O3 exposure appear to be resolved more slowly than changes in FEVj or respiratory
19         symptoms.
20
21      3. Health effects of long-term exposures to Ozone
22           Assessment of human health effects associated with long-term O3 exposures is hampered
23      by the lack of pertinent data from human clinical and epidemiologic studies.  Chronic animal
24      toxicology studies continue to support structural  alterations in several regions of the respiratory
25      tract and identify  the centriacinar region of the lung as the most affected region.
26      »D Animal toxicology studies that utilized exposure regimens to simulate seasonal  exposure
27         pattern also report increased lung injury compared to conventional chronic stable exposures.
28         One long-term study of infant rhesus monkeys exposed to simulated seasonal O3 patterns
29         (0.5 ppm 8h/day for 5 days, every 14 days for 11 episodes) demonstrated: (1) remodeling in
30         the distal airways; (2) abnormalities in tracheal basement membrane; (3) eosinophil
31         accumulation in conducting airways; (4) decrements in airway innervation. These findings

        August 2005                              E-25        DRAFT-DO NOT QUOTE OR CITE

-------
 1         advance earlier information regarding possible injury-repair processes occurring with
 2         seasonal O3 exposures.
 3
 4      •  Effects of O3 on the upper respiratory tract of F344 rats exposed to O3 (0.12, 0.5, or 1.0 ppm
 5         for 20 months) included marked mucous cell metaplasia in the rats exposed to 0.5 and
 6         1.0 pm O3, but not at 0.12 ppm O3. The persistent nature of the O3-induced mucous cell
 7         metaplasia suggests that O3 exposure may have the potential to induce similar long-lasting
 8         alterations in the airways of humans.  Hyperplasia in the nasal epithelium of rats exposed to
 9         0.25 and 0.5 ppm, 8h/day, 7 days/week, for 13 weeks has been reported.
10
11      •  Pathophysiological changes associated with chronic O3 exposures observed in animal studies
12         suggest possible similar alterations in humans.  The pulmonary function changes observed in
13         children in polluted metropolitan areas and lung structural alterations reported in an autopsy
14         study in Los Angeles suggest a role for long-term ambient O3 exposure, but such possible
15         effects need to be further evaluated with improved study design(s).
16
17      4. Susceptibility factors associated with exposure to ozone
18           Various factors such as age, gender, nutrition, socioeconomic, activity patterns, and disease
19      status have been shown to influence the response to environmental air pollutants.  Controlled
20      human exposure studies clearly established differential biological response to O3 based on
21      physical activity (exertion) and age.  These studies also demonstrated a large variation in
22      sensitivity and responsiveness to O3. The specific factors that contribute to this intersubject
23      variability are yet to be identified.
24      »D Increased  hospital admissions for asthma and COPD in summer (with increased levels of
25         ambient O3) suggest that people with these respiratory diseases as potential sub-population
26         for O3-induced health effects.
27
28      •  Similarly, based on O3-induced differential responses in lung inflammation and in airway
29         hyperresponsiveness, asthmatics (including children) appear to have potentially increased
30         susceptibility to O3. However, there is no supportive data from controlled human studies
31         suggesting individuals with COPD are more sensitive to O3-induced health effects.

        August 2005                               E-26        DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Animal toxicology studies provided supportive evidence to the observations of varied
 2         susceptibility. Various strains of mice and rats have demonstrated the importance of genetic
 3         background in O3 susceptibility. Moreover, genetic and molecular characterization studies in
 4         laboratory animals identified genetic loci responsible for both sensitivity and resistance.
 5
 6      •  Consistent with the 1996 O3 AQCD,  the scarcity of data prevents determination of the role of
 7         ethnic or racial background and nutrition status on O3-induced health effects. However,
 8         as presented in this document, exercising (moderate to high physical exertion) healthy
 9         adolescents and asthmatics appear to demonstrate increased responsiveness to ambient
10         concentrations of O3 and may be susceptible for O3-induced health effects.
11
12      5. Health effects of binary pollutant mixtures containing ozone
13           A limited number of controlled human exposure studies and a few animal toxicology
14      studies of binary mixtures containing O3 suggest potential  interactions, depending on specific
15      exposure regimens and copollutant constituents.
16      »D Continuous exposure to SO2 and NO2 increased inhaled bolus O3 absorption, while
17         continuous exposure to O3 decreased O3 bolus absorption. Asthmatics exhibited enhanced
18         airway reactivity to house dust mite following exposures to O3, NO2, and the combination of
19         the two gases.  Spirometric response, however, was impaired only by O3 and O3+NO2 at
20         higher concentrations.
21
22      •  Animal toxicology studies with O3 in mixture with NO2, formaldehyde, and PM
23         demonstrated additive, synergistic or antagonistic effects, depending on the exposure
24         regimen and the endpoints evaluated.
25
26      •  One  controlled exposure study of children, designed to approximate exposure conditions of
27         an epidemiologic study by matching  the population and exposure atmosphere (0.1 ppm O3,
28         0.1 ppm SO2 and 101 |ig/m2 H2SO4), failed to support the  findings of the epidemiologic study.
29         This study points out difficulties in trying to link the outcomes of epidemiologic and
30         controlled exposure studies by use of binary pollutant mixtures.
31

        August 2005                              E-27        DRAFT-DO NOT QUOTE OR CITE

-------
 1      E.9   VEGETATION AND ECOLOGICAL EFFECTS
 2      General
 3      •  Data published since 1996 continue to support the conclusions of previous O3 AQCDs that
 4         there is strong evidence that ambient O3 concentrations cause foliar injury along with growth
 5         and yield damage to numerous common and economically valuable plant and tree species.
 6
 7      •  Research to date has continued its focus at the species level, with very few new studies at the
 8         ecosystem level. The lack of quantification of biotic and abiotic factors impinging on the
 9         individual to population organizational levels results in a limited ability to scale O3 responses
10         to the ecosystem level.  Therefore, a high degree of uncertainty remains in our ability to
11         assess ozone risk to ecological resources and the services they provide.
12
13      Methodologies
14      •  Since the 1996 AQCD free-air exposure (FACE) systems have come into more frequent use.
15         FACE systems eliminate many of the concerns raised about closed or open-top chamber
16         (OTC) experiments including small plot size, altered microclimate within OTCs, and the
17         effect of charcoal filtering on overall air quality within OTCs. One of the advantages of the
18         application of plume systems to O3 research is the ability to compare response of plants in
19         open-field systems with results from OTCs. In particular,  studies with quaking aspen
20         (Populus tremuloides L.) performed in OTCs, FACE, and  also at sites along an ambient O3
21         gradient showed that O3 symptom expression was generally similar, supporting the
22         previously observed level of variation among aspen clones in OTC studies.
23
24      •  The lack of rural monitors continues to be a major problem in the characterization of O3
25         exposures in remote areas, as well as in linking effects to exposure in natural ecosystems.
26         Since the 1996 O3 AQCD, the use of passive samplers has expanded monitoring efforts to
27         include remote areas that were previously uncharacterized.
28
29      •  Advancements in biomonitoring have been made since the 1996 O3 AQCD, primarily in the
30         area of identification and symptom verification of sensitive species . The U.S. Department
31         of Agriculture (USD A) Forest Service continues its program to monitor O3 effects in forested

        August 2005                             E-28        DRAFT-DO NOT QUOTE OR CITE

-------
 1         ecosystems throughout the United States.  Although results are not useful for developing
 2         exposure-response relationships or for quantifying responses to O3, they can provide an
 3         annual assessment and correlative information regarding the extent of O3 injury occurring
 4         across many regions of the United States.
 5
 6      Mode of Action
 7      •  The new information available on the mode of action of O3 is, in part, a result of improved
 8         molecular tools for following rapid changes that occur within the leaf.  Many changes occur
 9         within hours or possibly days following O3 exposure. Other O3 effects take longer to occur
10         and tend to be most obvious only under exposure to low O3 concentrations for long periods.
11         These low-exposure chronic effects have been linked to the senescence process or some
12         physiological response very closely linked to senescence (e.g., translocation, reabsorption,
13         allocation of nutrients and carbon).
14
15      Modification of Growth Response
16      •  It has been known for decades that several factors, both biotic and abiotic, alter plant
17         response to O3. However, only a few studies reported since the 1996 O3 AQCD have
18         improved our understanding of the role of these interactions in modifying plant O3 response.
19
20      •  Recent studies have supported the earlier conclusion that O3 often increases the likelihood
21         and success of insect attacks, but only with respect to chewing insects. Although it seems
22         likely that some insect problems could increase as a result  of greater O3 levels, we are still far
23         from being able to predict the nature of any particular O3-plant-insect interaction, its
24         likelihood, or its severity.
25
26      •  O3 exposure generally increases plant diseases associated with facultative necrotrophic plant
27         pathogens. Generally, pathogens that benefit from damage to cells are enhanced by O3 stress
28         of their hosts, whereas pathogens and pests that require healthy hosts are depressed by O3
29         stress.
30
31

        August 2005                               E-29        DRAFT-DO NOT QUOTE  OR CITE

-------
 1      Exposure Indices
 2      •  Exposure indices are metrics that relate measured plant damage (i.e., reduced growth) to
 3         monitored ambient O3 concentrations over time to provide a consistent metric for reviewing
 4         and comparing exposure-response effects obtained from various studies. Since the 1996 O3
 5         AQCD, there has been no direct experimental testing of the adequacy of exposure indices
 6         proposed in 1996; therefore, there is no new information to alter the basic conclusions put
 7         forth in the 1996 O3 AQCD.
 8
 9      •  The proposed indices in the 1996 O3 AQCD (i.e., SUM06, W126, AOT40) included various
10         functional and statistical summaries of monitored hourly O3 concentrations over designated
11         time periods.  The few studies that have been published since the 1996 O3 AQCD continue to
12         support the earlier conclusions, including the importance of peak concentrations, and the
13         duration and occurrence of O3 exposures in altering plant growth and yield.
14
15      •  A large body of new research, mostly out of Europe, addresses the need for an index related
16         to the actual flux of O3 into the plant. Despite additional research linking estimates of flux
17         with plant response since 1996, information is still insufficient to identify a flux-based model
18         that incorporates the necessary complexity across space and time to be non-site or non-
19         species specific. Based on the current state of knowledge, exposure indices that cumulate
20         and differentially weight the higher hourly average concentrations,  but include the mid-level
21         values (e.g., SUM06, W126, AOT40), still represent the best approach for relating vegetation
22         effects to O3 exposure in the United States.
23
24      Ozone Exposure-Plant Response Relationships
25      •  Data published since 1996 continue to support the conclusions of previous O3 AQCDs that
26         there is strong evidence that ambient O3 concentrations cause foliar injury and growth and
27         yield damage to numerous common and economically valuable plant and tree species.
28
29      •  In addition to reductions in crop yield, O3 may also reduce the quality or nutritive value of
30         annual species.  Many recent studies have found O3 effects on various measures of plant
31         organs that affect quality, with most of those studies focusing on characteristics important for

        August 2005                              E-30        DRAFT-DO NOT QUOTE OR CITE

-------
 1         food or fodder.  These studies indicate that ambient O3 may have economically important
 2         effects on the quality of crop and forage species.
 3
 4      •  Results since 1996 support the conclusion of the 1996 O3 AQCD that deciduous trees are
 5         generally less O3 sensitive than are most annual plants, with the exception of a few very
 6         sensitive genera such as Populus and sensitive species  such as black cherry. Various
 7         evergreen tree species and genotypes have widely varying O3 sensitivities.  Based on OTC
 8         studies with seedlings, major evergreen species in the United States are generally less
 9         sensitive than are most deciduous trees, and slower-growing evergreen species are less
10         sensitive than are faster-growing species. For all types of perennial vegetation, cumulative
11         effects over more than one growing season may be important; studies for only a single
12         season may underestimate effects.
13
14      Ecosystem Effects
15      •  There is evidence that tropospheric O3 is an important stressor of ecosystems, with
16         documented impacts on the biotic condition, ecological processes, and chemical/physical
17         nature of natural ecosystems. Effects on individual keystone species and their associated
18         microflora and fauna, which have been shown experimentally, may cascade through the
19         ecosystem to the landscape level, although this has not yet been demonstrated.
20
21      •  Systematic injury surveys (e.g., USDA Forest Service's ozone bioindicator plot network and
22         Europe's TCP Forests) demonstrate that foliar injury occurs on O3 sensitive species in many
23         regions of the United States and Europe.  Frequent lack of correspondence between foliar
24         symptoms and growth effects means that other methods must be used to estimate the regional
25         effects of O3 on tree growth.  Investigations of the radial growth of mature trees,  combined
26         with data from many controlled studies with seedlings and a few studies with mature trees
27         suggest that ambient O3 is reducing the growth of mature trees in some U.S. locations.
28
29      •  The study of genetic aspects of O3 impacts on natural ecosystems has been largely based on
30         correlations, and it remains to be shown more conclusively whether O3 affects biodiversity or
31         genetic diversity.

        August 2005                              E-31        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Economics
 2      •  The physical and economic effects on agriculture are well documented and provide useful
 3         information for the consideration of establishing air quality standards for crops. Effects on
 4         forests and natural ecosystems remain problematic, due to limitations in biological response
 5         data and economic methods. The problem is even more acute for valuing natural ecosystem
 6         goods and services.
 7
 8
 9      E.10   TROPOSPHERIC OZONE EFFECTS ON UV-B FLUX AND  ITS
10             ROLE IN CLIMATE CHANGE
11          The molecular properties specific to O3 include a capacity for absorbing incoming
12      ultraviolet (UV) and infrared (IR) radiation,  and both incoming solar and outgoing terrestrial IR
13      radiation.  Consequently, O3 plays an essential role in shielding the earth's surface from harmful
14      levels of UV-B radiation, by way of the stratospheric O3 layer.  Its effectiveness as a screen for
15      the residual UV-B flux that penetrates the stratosphere and passes into the troposphere and its
16      role in reducing UV-induced human health effects are addressed in Chapter 10.  The radiation-
17      absorbing properties of O3 also make it a greenhouse gas (GHG) having global and, more
18      importantly, regional consequence for climate, as also addressed in Chapter 10.  Important
19      conclusions from Chapter 10 are summarized below.
20
21      •  The distribution of (X within the atmosphere. Ozone is distributed very unevenly  within the
22         atmosphere, with -90% of the total atmospheric burden present in the stratosphere.  The
23         remaining -10% is distributed within the troposphere, with higher relative concentrations
24         near the source of its precursors at the surface.  Concentrations of O3 at the mid- and upper-
25         troposphere vary, depending upon meteorological conditions.
26
27      •  Multiple factors govern the flux of UV-B radiation at the Earth's surface. Latitude and
28         altitude are the two most important factors that  define the residual UV-B flux at the  surface.
29         Natural variation in the total column density of stratospheric O3 is also an important factor.
30         All of these factors are followed in importance by tropospheric clouds, particulate matter
        August 2005                              E-32        DRAFT-DO NOT QUOTE OR CITE

-------
 1         (PM) and O3. The effect of natural stratospheric variation, clouds, PM and tropospheric O3
 2         on UV fluxes within the troposphere and at the surface are each very difficult to predict.
 3
 4      •  A UV-B "climatology" is needed to predict human exposure levels. A UV-B climatology,
 5         representing patterns and trends in UV-B flux at the Earth's surface, must be based on
 6         extended in situ observations in order to adequately capture natural variability and the effects
 7         of human activities on atmospheric UV-B absorbers. At present, the body of UV-B
 8         measurements cannot support the development of a climatology.
 9
10      •  Human exposure to UV-B radiation. Quantitative evaluation of human exposure to UV-B
11         radiation is necessary to perform health risk assessment of UV-B-related health effects.
12         Individuals who participate in outdoor sports and activities, work outdoors, live in
13         geographic areas with higher solar flux, and/or engage in high-risk behavior (e.g., extended
14         sun bathing) can reasonably be projected to be at increased risk for higher UV radiation
15         exposures.  However, little is known about the impact of variability in these factors on
16         individual exposure to UV radiation.
17
18      •  Human health effects of UV-B radiation. Exposure to UV-B radiation is associated with
19         increased risk of erythema, nonmelanoma and melanoma skin cancers, ocular damage, and
20         immune system suppression. Some studies have attempted to estimate the potential effects
21         of changes in surface-level UV flux resulting from stratospheric O3 depletion on these health
22         outcomes; however, the numerous simplifying assumptions made in the  assessments limit the
23         usefulness of the risk estimates. The effect of changes in surface-level O3 concentrations on
24         UV-induced health outcomes cannot yet be critically assessed within reasonable uncertainty.
25
26      •  Vitamin D-related health benefits of UV-B radiation. A potential health benefit of increased
27         UV-B exposure relates to the production  of vitamin D in humans.  Several studies have
28         found that UV-B radiation, by increasing vitamin D production, is associated with reduced
29         risks of various cancers. However, as with other impacts of UV-B on human health, this
30         beneficial effect of UV-B has not been studied in sufficient detail to allow for a credible
31         health benefits assessment.

        August 2005                              E-33         DRAFT-DO NOT QUOTE OR CITE

-------
 1      •  Ozone is a potent GHG. Ozone traps incoming solar radiation at both ends of the spectrum,
 2         as well as shortwave radiation that is scattered from high-albedo portions of the Earth's
 3         surface.  Outgoing terrestrial IR is absorbed by O3 within the range where water vapor does
 4         not absorb, so that natural variability in humidity does not alter its radiative impact. These
 5         effects directly force climate. By participating in the oxidative chemistry of the atmosphere,
 6         O3 can indirectly and negatively force climate by the removal of other greenhouse gases.
 7
 8      •  Multiple factors influence the forcing effect of tropospheric O?.  Estimates of present-day
 9         forcing by O3 depend upon the available information on pre-industrial and current
10         concentrations.  Both are limited and, therefore, very uncertain.  Other factors, including the
11         albedo of underlying surface, altitude and co-occurrence of PM  can also complicate the
12         calculation of globally-averaged forcing.
13
14      •  Globally-averaged direct forcing by O3. On the basis of the best available information, a
15         2001 Intergovernmental Panel on Climate Change (TPCC) report offered an estimated value
16         of 0.35 ± 0.15 WnT2 for the annual, globally-averaged direct forcing by tropospheric O3.
17         Another recent estimate places this value at 0.5 ± 0.2 WnT2.
18
19      •  Projections of forcing by O? into the future. A CTM-climate modeling intercomparison
20         study carried out as  part of the third assessment by the IPCC yielded an estimated 0.4 to 0.78
21         WnT2 forcing by O3 by the year 2100. The authors of this study concluded that O3 can be
22         expected to be an important contributor to climate forcing into the future.
23
24      •  Climate forcing by O3 at the regional scale may be its most import impact on climate.
25         Satellites have detected high O3 concentrations localized at the regional scale that are
26         associated with large urban centers and extensive biomass burning. Climate forcing by these
27         high, regional-scale O3 concentrations have been estimated to be on the order of 1 WnT2 (a
28         substantial fraction of the direct, globally-averaged forcing due to well-mixed GHGs,
29         including CO2). The impact of climate forcing at this level depends upon the particular
30         characteristics of the region in which it occurs.  At present, regional-scale modeling studies

        August 2005                              E-34        DRAFT-DO NOT QUOTE OR CITE

-------
 1         are not available that provide estimates of these effects.  Research efforts to do so are
 2         underway.
 3
 4
 5      E.ll  MATERIALS DAMAGE
 6           The Chapter 11 discussion of O3 effects on man-made materials mainly summarizes key
 7      information from the 1996 O3 AQCD, given that little new pertinent research information on O3-
 8      related materials damage has been published since then. Key points include:
 9
10      •  Ozone and other photochemical oxidants react with many economically important man-made
11         materials, decreasing their useful life and aesthetic appearance.  Materials damaged by O3
12         include elastomers; textiles and fibers; dyes, pigments, and inks; and paints and other surface
13         coatings.
14
15      •  Elastomeric compounds (natural rubber and synthetic polymers and copolymers of
16         butadiene, isoprene, and styrene) are highly susceptible, even to low O3 concentrations.
17         Ozone damages these compounds by breaking the molecular chain at the carbon-carbon
18         double bond and by adding a chain of three oxygen atoms directly across the double bond.
19         This structure change promotes characteristic cracking of stressed/stretched rubber called
20         "weathering."  Tensile strain produces cracks on the surface of the rubber that increase in
21         size and number with increased stress/stretching.  The rate of crack growth is dependent on
22         degree of stress, type of rubber compound, O3 concentration, duration of exposure, O3
23         velocity, and temperature. After initial cracking,  further O3 penetration results in additional
24         cracking and, eventually, mechanical weakening.
25
26      •  Ozone can damage textiles and fabrics by mechanisms similar to those associated with
27         elastomers. Generally, synthetic fibers are less affected by O3 than natural  fibers. Overall,
28         O3 contribution to degradation of textiles and fabrics is not considered significant.
29


        August 2005                              E-3 5       DRAFT-DO NOT  QUOTE OR CITE

-------
1      •  Ozone fading of textile dyes is a diffusion-controlled process, with the rate of fading being
2         controlled by diffusion of the dye to the fiber surface. Many textile dyes react with O3. The
3         rate and severity of the O3 attack is influenced by the chemical nature of the textile fiber and
4         the manner in which the dye has been applied.
5
6      •  Paints applied to exterior surfaces of buildings and other structures (e.g., bridges), as well as
7         several artists' pigments, are also sensitive to fading and oxidation by O3 at concentrations
8         found in urban areas.
9
       August 2005                               E-36        DRAFT-DO NOT QUOTE OR CITE

-------
 i                                 1.  INTRODUCTION
 2
 3
 4           This is an update revision of the document, "Air Quality Criteria for Ozone and Related
 5     Photochemical Oxidants" published by the U.S. Environmental Protection Agency (EPA) in
 6      1996 (U.S. Environmental Protection Agency, 1996). That 1996 Ozone Air Quality Criteria
 7     Document (O3 AQCD) provided scientific bases for Congressionally-mandated periodic review
 8     by the EPA of the National Ambient Air Quality Standards for Ozone (O3 NAAQS), which
 9     culminated in promulgation of new O3 NAAQS by EPA in 1997.
10           The present document critically assesses the latest scientific information relative to
11     determining the health and welfare effects associated with the presence of various concentrations
12     of O3 and related oxidants in ambient air. It builds upon the previous 1996 EPA O3 AQCD,
13     by focusing on evaluation and integration of information relevant to O3 NAAQS criteria
14     development that has become available since that covered by the 1996 criteria review; and it will
15     provide scientific bases for the current periodic review of the O3 NAAQS.
16           This introductory chapter of the revised O3 AQCD presents: (a) background information
17     on legislative requirements, the criteria and NAAQS review process, and the history of O3
18     NAAQS reviews (including a chronology of changes in key elements of the O3 standards);
19     (b) an overview of the current O3 criteria review process and projected schedule (including
20     approaches and procedures used to prepare this document, as well as projected key milestones);
21     and (c) an orientation to the general  organizational structure and content of the document.
22
23
24      1.1   LEGAL AND HISTORICAL BACKGROUND
25      1.1.1  Legislative Requirements
26           Two sections of the Clean Air  Act  (CAA) govern the establishment, review, and revision
27     of National Ambient Air Quality Standards (NAAQS).  Section 108 (42 U.S.C. 7408) directs the
28     Administrator of the U.S. Environmental Protection Agency (EPA) to identify ambient air
29     pollutants that may be reasonably anticipated to endanger public health or welfare and to issue
30     air quality criteria for them.  These air quality criteria are to reflect the latest scientific


       August 2005                               1-1         DRAFT-DO NOT QUOTE OR CITE

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

22
23
24
25
26
27
28
29
30
31
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.
     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.
      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
   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 Section 109(b).  Alternatively,
the Administrator may find that revision of the standards is inappropriate and conclude the
review by leaving the existing standards unchanged.  Section 109(d)(2) of the 1977 CAA
Amendments 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 (CAS AC) of EPA's Science Advisory Board (SAB).
August 2005
                                                    1-2
     DRAFT-DO NOT QUOTE OR CITE

-------
 1      1.1.2   Criteria and NAAQS Review Process
 2          Periodic reviews by EPA of criteria and NAAQS for a given criteria air pollutant progress
 3      through a number of steps, beginning with preparation by EPA's National Center for
 4      Environmental Assessment Division in Research Triangle Park,  NC (NCEA-RTP) of an air
 5      quality criteria document  (AQCD).  The AQCD provides a critical assessment of the latest
 6      available scientific information upon which the NAAQS are to be based.  Drawing upon the
 7      AQCD, staff of EPA's Office of Air Quality Planning and Standards (OAQPS) prepare a Staff
 8      Paper that evaluates policy implications of the key studies and scientific information contained
 9      in the AQCD and presents EPA staff conclusions and recommendations for standard-setting
10      options for the EPA Administrator to consider. The Staff Paper is intended to help "bridge the
11      gap" between the scientific assessment contained in the AQCD and the judgments required of
12      the Administrator in determining whether it is appropriate to retain or to revise the NAAQS.
13      Iterative drafts of both the AQCD and the Staff Paper (as well as other analyses, such as
14      exposure and/or risk assessments, supporting the Staff Paper)  are made available for public
15      comment and CAS AC review. The final versions of the AQCD and Staff Paper incorporate
16      changes made in response to CAS AC and public review.  Based on the information in these
17      documents, the Administrator proposes decisions on whether to  retain or revise the NAAQS,
18      taking into account public comments and CASAC advice and  recommendations.  The
19      Administrator's proposed decisions are published in the Federal Register, with a preamble that
20      presents the rationale for the decisions and solicits public comment. The Administrator
21      makes a final decision after considering comments received on the proposed decisions. The
22      Administrator's final decisions are promulgated in a Federal Register notice that addresses
23      significant comments received on the proposal.
24          NAAQS decisions involve consideration of the four basic elements of a standard:
25      indicator, averaging time, form, and level. The indicator defines the pollutant to be measured in
26      the ambient air for the purpose of determining compliance with the standard. The averaging
27      time defines the time period over which air quality measurements are to be obtained and
28      averaged, considering evidence of effects associated with various time periods of exposure.
29      The form of a standard defines the air quality statistic that is to be compared to the level of the
30      standard (i.e., an ambient  concentration of the indicator pollutant) in determining whether an
31      area attains the standard.  The form  of the standard specifies the air quality measurements that

        August 2005                               1 -3        DRAFT-DO NOT QUOTE OR CITE

-------
 1      are to be used for compliance purposes (e.g., the 98th percentile of an annual distribution of
 2      daily concentrations; the annual arithmetic average), the monitors from which the measurements
 3      are to be obtained (e.g., one or more population-oriented monitors in an area), and whether the
 4      statistic is to be averaged across multiple years. These basic elements of a standard are the
 5      primary focus of the staff conclusions and recommendations in the Staff Paper and in the
 6      subsequent rulemaking, building upon the policy-relevant scientific information assessed in the
 7      AQCD and on the policy analyses contained in the Staff Paper.  These four elements taken
 8      together determine the degree of public health and welfare protection afforded by the NAAQS.
 9
10      1.1.3   Regulatory Chronology1
11          On April 30, 1971, the EPA promulgated primary and secondary NAAQS for
12      photochemical oxidants under Section 109 of the CAA (36 FR 8186).  These were set at an
13      hourly average of 0.08 ppm total photochemical oxidants, not to be exceeded more than 1 h per
14      year. On April 20, 1977, the EPA announced (42 FR 20493) the first review and updating of the
15      1970 Air Quality  Criteria Document for Photochemical Oxidants in accordance with Section
16      109(d) of the CAA. In preparing that AQCD, the EPA made two external review drafts of the
17      document available for public comment, and these drafts were peer reviewed by the
18      Subcommittee on Scientific Criteria for Photochemical Oxidants of EPA's Science Advisory
19      Board (SAB). A final revised AQCD for ozone (O3) and other photochemical oxidants was
20      published on June 22, 1978.
21          Based on the 1978 revised AQCD and taking into account the advice and recommendations
22      of the SAB Subcommittee and public comments, the EPA announced (44 FR 8202) a final
23      decision to revise the NAAQS for photochemical oxidants on February 8, 1979.  That final
24      rulemaking revised the primary standard from 0.08 ppm to 0.12 ppm, set the secondary standard
25      to be the same as the primary standard, changed the chemical designation of the standards from
26      photochemical oxidants to O3, and revised the definition of the point at which the standard is
27      attained as indicated in Table 1-1.
28
              '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).

        August 2005                                1 -4        DRAFT-DO NOT QUOTE OR CITE

-------
       	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 10'4 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).
 1          On March 17, 1982, in response to requirements of Section 109(d) of the CAA, the EPA
 2     announced (47 FR 11561) that it planned to revise the existing 1978 AQCD for O3 and Other
 3     Photochemical Oxidants; and, on August 22,  1983, it announced (48 FR 38009) that review of
 4     the primary and secondary NAAQS for O3 had been initiated.  The EPA provided a number of
 5     opportunities for expert review and public comment on revised chapters of the AQCD, including
 6     two public peer-review workshops in December 1982 and November 1983.  Comments made at
 7     both workshops were considered by EPA  in preparing the First External Review Draft that was
 8     made available (49 FR 29845) on July 24, 1984, for public review. On February 13,  1985
 9     (50 FR 6049) and then on April 2, 1986 (51 FR 11339), the EPA announced two public CASAC
10     meetings, which were held on March 4-6,  1985 and April 21-22, 1986, respectively.  At these
11     meetings, the CASAC reviewed external review drafts of the revised AQCD for O3 and Other
12     Photochemical Oxidants.  After these two reviews, the Chair summarized CASAC's consensus
13     view in an October 1986 letter to the EPA Administrator, which stated that the document
14     "represents a scientifically balanced and defensible summary of the extensive scientific
15     literature." Taking into account public and CASAC comments on the two external review  drafts,
16     revisions were made by EPA and the  final document was released by EPA in August 1986.
17          The first draft of the Staff Paper "Review of the National Ambient Air Quality Standards
18     for Ozone: Assessment of Scientific and Technical Information" drew upon key findings and
19     conclusions from the  AQCD and was reviewed by CASAC at the April 21-22, 1986 public
20     meeting. At that meeting, the CASAC recommended that new information on prolonged O3

       August 2005                               1-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposure effects be considered in a second draft of the Staff Paper. The CAS AC reviewed the
 2      resulting second draft and also heard a presentation of new and emerging information on the
 3      health and welfare effects of O3 at a December 14-15, 1987 public review meeting. The CASAC
 4      concluded that sufficient new information existed to recommend incorporation of relevant
 5      new data into a supplement to the 1986 AQCD (O3 Supplement) and in a third draft of the
 6      Staff Paper.
 7          A draft O3 Supplement, "Summary of Selected New Information on Effects of Ozone on
 8      Health and Vegetation: Draft Supplement to Air Quality Criteria for Ozone and Other
 9      Photochemical Oxidants," and the revised Staff Paper were made available to CASAC and to the
10      public in November 1988. The O3 Supplement assessed selected literature concerning exposure-
11      and concentration-response relationships observed for health effects in humans and experimental
12      animals and for vegetation effects that appeared in papers published or in-press from  1986
13      through early 1989.  On December 14-15, 1988, CASAC held a public meeting to review these
14      documents and then sent the EPA Administrator a letter (dated May 1, 1989), which stated that
15      the draft O3 Supplement, the 1986 AQCD, and the draft Staff Paper "provide an adequate
16      scientific basis for the EPA to retain or revise the  primary and secondary standards of ozone."
17      The CASAC concluded (a) that it would be some  time before sufficient new information on the
18      health effects of multihour and chronic exposure to O3 would be published in scientific journals
19      to receive full peer review and, thus, be suitable for inclusion in a criteria document and (b) that
20      such information could be considered in the next review of the O3 NAAQS. A final version of
21      the O3 Supplement was published in 1992 (U.S. Environmental Protection Agency, 1992).
22          On October 22, 1991, the American Lung Association and other plaintiffs filed  suit to
23      compel the Agency to complete the  review of the  criteria and standards for O3 in accordance
24      with the CAA.  The  U.S. District Court for the Eastern District of New York subsequently issued
25      an order requiring the EPA to announce its proposed decision on whether to revise the standards
26      for O3 by August  1,  1992 and to announce its final decision by March  1, 1993.
27          The proposed decision on O3, which appeared in the Federal Register on August 10, 1992
28      (57 FR 35542), indicated that revision of the existing 1-h NAAQS was not appropriate at that
29      time. A public hearing on this decision was held in Washington, DC on September 1, 1992; and
30      public comments were received through October 9, 1992.  The final decision not to revise the
31      1-h NAAQS was published in the Federal Register on March 9, 1993 (58 FR 13008). However,

        August 2005                              1 -6        DRAFT-DO NOT QUOTE OR CITE

-------
 1     that decision did not take into consideration a number of more recent studies on the health
 2     and welfare effects of O3 that had been published since the last of the literature assessed in
 3     the O3 Supplement (i.e., studies available through 1985 and into early 1986).
 4           The Agency initiated consideration of such studies as part of the next congress!onally-
 5     mandated periodic review of criteria and NAAQS for Ozone. The new studies were assessed in
 6     revised draft O3 AQCD chapters that were peer reviewed in July and September 1993
 7     workshops, followed by public release of the First External Review Draft in February 1994 and
 8     CAS AC review on July 20-21, 1994. Further drafts of the O3 AQCD, revised in response to
 9     public comments and CAS AC review, were reviewed by CAS AC on March 21-25, 1995, and at
10     a final CAS AC review meeting on September 19-20, 1995. The scientific soundness of the
11     revised O3 AQCD was recognized by CAS AC in a November 28, 1995 letter to the EPA
12     Administrator; and the final AQCD for O3 was published in July 1996.
13           The first draft of the associated Staff Paper, "Review of the National Ambient Air Quality
14     Standards for Ozone: Assessment of Scientific and Technical Information," was also reviewed
15     by CAS AC at the March 21-22, 1995 public meeting. CAS AC also  reviewed subsequent drafts
16     of the Staff Paper at public meetings on September 19-20, 1995 and  March 21, 1996, with
17     completion of CAS AC review of the primary and secondary standard portions of the draft Staff
18     Paper being communicated in letters to the EPA Administrator dated November 30, 1995 and
19     April 4, 1996, respectively.  The final O3 Staff Paper was published in June 1996.
20           On December 13, 1996 EPA published its proposed decision to revise the O3 NAAQS
21     (61 FR 65716). EPA provided extensive opportunities for public comment on the proposed
22     decision, including several public hearings and two national satellite telecasts. EPA's final
23     decision to promulgate a new 8-h O3 NAAQS (see Table 1-1) was published on July 18,  1997
24     (62 FR 38856).
25           Following promulgation of the new standards, numerous petitions for review of the
26     standards were filed in the U.S. Court of Appeals for the District of Columbia Circuit (D.C.
27     Circuit)2. On May 14, 1999, the Court remanded the O3 NAAQS to  EPA, finding that section
28     109 of the CAA, as interpreted by EPA, effected an unconstitutional delegation of legislative
              :''American Trucking Associations v'. EPA, No. 97-1441

       August 2005                               1 -7        DRAFT-DO NOT QUOTE OR CITE

-------
 1      authority3. In addition, the Court directed that, in responding to the remand, EPA should
 2      consider the potential beneficial health effects of O3 pollution in shielding the public from the
 3      effects of solar ultraviolet (UV) radiation. On January 27, 2000, EPA petitioned the U.S.
 4      Supreme Court for certiorari on the constitutional issue (and two other issues), but did not
 5      request review of the D.C. Circuit ruling regarding the potential beneficial health effects of O3.
 6      On February 27, 2001  the U.S. Supreme Court unanimously reversed the judgment of the D.C.
 7      Circuit on the constitutional issue, holding that section 109 of the CAA does not delegate
 8      legislative power to the EPA in contravention of the Constitution, and remanded the case to the
 9      D.C. Circuit to consider challenges to the O3 NAAQS that had not been addressed by that Court's
10      earlier decisions4. On March 26, 2002, the D.C. Circuit issued its final decision, finding the
11      1997 O3 NAAQS to be "neither arbitrary  nor capricious," and denied the remaining petitions
12      for review5.
13          On November 14, 2001 EPA proposed to respond to the Court's remand to consider the
14      potential beneficial health effects of O3 pollution in shielding the public from the effects of solar
15      UV radiation by leaving the 1997 8-h NAAQS unchanged. Following  a review of information in
16      the record and the substantive comments received on the proposed response, EPA issued a final
17      response to the remand, reaffirming the 8 h O3 NAAQS (68 FR 614, January 6, 2003).
18
19
20      1.2   CURRENT OZONE CRITERIA AND NAAQS REVIEW
21      1.2.1   Key Milestones and Procedures for Document Preparation
22          It is important to note at the outset that development of the present document has and will
23      continue to include substantial external expert review and opportunities for public input through
24      (a) public workshops involving the general scientific community, (b) iterative reviews of
25      successive drafts by CASAC,  and (c) comments from the public on successive drafts. Extensive
26      external inputs received through such reviews will help to ensure that the review of the  O3
              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)

        August 2005                              1-8         DRAFT-DO NOT QUOTE OR CITE

-------
 1      standards will be based on critical assessment in this document of the latest available pertinent
 2      science.
 3           The procedures for developing this revised O3 AQCD build on experience derived from the
 4      other recent criteria document preparation efforts, with key milestones for development of
 5      this O3 AQCD being listed in Table 1-2. Briefly, the respective responsibilities for production
 6      of the document and key milestones are as follows. An NCEA-RTP Ozone Team is responsible
 7      for the creation and implementation of a project plan for developing the O3 AQCD, taking into
 8      account input from individuals in other EPA program and policy offices identified as part of the
 9      EPA Ozone Work Group.  The resulting plan, i.e., the Project Work Plan for Revised Air
10      Criteria for Ozone and Related Photochemical Oxidants (November 2002), was discussed with
11      CASAC in January  2003. An ongoing literature search that was underway prior to initiation of
12      work on this document has continued throughout its preparation to identify pertinent O3
13      literature published  since early 1996. Under the processes established in Sections 108 and 109
14      of the CAA, the EPA officially initiated the current criteria and NAAQS review by announcing
15      the commencement of the review in the Federal Register (65 FR 57810, September, 2000) with a
16      call for information. That Federal Register notice included (1) a request asking for recently
17      available research information on O3 that may not yet have been published and (2) a request for
18      individuals with the appropriate type and level of expertise to contribute to the writing of O3
19      AQCD materials to  identify themselves. The specific authors of chapters  or sections of the
20      proposed document included both EPA and non-EPA scientific experts, who were selected on
21      the basis of their expertise on the subject areas and their familiarity with the relevant literature.
22      The project team defined critical issues and topics to be addressed by the authors and provided
23      direction in order to focus on evaluation of those studies most clearly identified as important for
24      standard setting.
25           As with other NAAQS reviews, critical assessment of relevant scientific information is
26      presented in this updated O3 AQCD. The main focus of this document is the evaluation and
27      interpretation of pertinent atmospheric science information, air quality data, human exposure
28      information,  and health and welfare effects information newly published since that assessed in
29      the 1996 O3 AQCD. Draft versions of AQCD chapter materials were evaluated via expert peer-
30      consultation workshop discussions (see Table 1-2) that focused on the selection of pertinent
31      studies to be included  in the chapters, the potential need for additional information to be added to

        August 2005                               1 -9        DRAFT-DO NOT QUOTE OR CITE

-------
               Table 1-2. Key Milestones for Development of Revised Ozone Air Quality
                                          Criteria Document"
        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 Review
         5.  CASAC Review of 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
        14.  Final O3 AQCD
                        Target Dates
                        Ongoing
                        September 2000
                        Dec 2001-March 2002
                        December 2002
                        January 2003
                        April 2003
                        July 2004

                        January 2005
                        Feb - April 2005
                        May 2005
                        August 2005
                        Sept - Nov 2005
                        December 2005
                        February 2006
        "Proposed schedule will be modified from time to time, as necessary, to reflect actual project requirements
         and progress.
1      the chapters, and the quality of the characterization and interpretation of the literature. The
2      authors of the draft chapters then revised them on the basis of the workshop and/or other expert
3      review comments6.  These and other integrative materials were then incorporated into the First
4      External Review Draft of this O3 AQCD (January 2005), which was made available for public
5      comment and CASAC review (see Table 1-2).
6           Following review of the First External Review Draft at a May 4-5, 2005 CASAC meeting,
7      EPA incorporated revisions into the draft O3 AQCD in response to comments from CASAC and
8      the public and has made this  Second External Review Draft (August, 2005) available for further
              6It should be noted that materials contributed by non-EPA authors have, at times, been 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.
       August 2005
1-10
DRAFT-DO NOT QUOTE OR CITE

-------
 1     public comment and CAS AC review according to the schedule projected in Table 1-2.  More
 2     specifically, this Second External Review Draft is available for public comment (90 days) during
 3     September-November, 2005, and will be reviewed by CASAC at a public meeting in December
 4     2005 (the site and specific dates to be announced in the Federal Register).  The final O3 AQCD is
 5     to be completed by February 28, 2006, and it is to be made publicly available electronically via
 6     an EPA website and then subsequently printed. Its availability will be announced in the Federal
 7     Register.
 8          The EPA's Office of Air Quality Planning and Standards (OAQPS) staff has also prepared
 9     a first draft O3 Staff Paper drawing upon key information contained in this Second External
10     Review Draft O3 AQCD, which presents recommendations regarding whether to retain or, if
11     appropriate, to revise the O3 NAAQS. After review of that draft Staff Paper (dated September,
12     2005) by the public and by CASAC, EPA will take public and CASAC comments into account
13     in producing a Second Draft Staff Paper.  That Second Draft Staff Paper (based on the final
14     version of this O3 AQCD) will also be made available for further public comment and CASAC
15     review before EPA produces a final ozone Staff Paper by September 30, 2006.
16
17
18     1.3   ORGANIZATIONAL STRUCTURE OF THE DOCUMENT
19     1.3.1   General Document Format
20          The general format used in preparing this draft document is to open each new section for
21     the updated document with concise summarization of key findings and conclusions from the
22     previous 1996 O3 AQCD (U.S.  Environmental Protection Agency, 1996).  After presentation of
23     such background information, the remainder of each section typically provides an updated
24     discussion of newer literature and resulting key conclusions.  In some cases where no new
25     information is available, the summary of key findings and conclusions from the previous criteria
26     document must suffice as the basis for current key conclusions. Increased emphasis is placed in
27     the main chapters of this revised O3 AQCD on interpretative evaluation and integration of
28     evidence pertaining to a given topic than has been typical of previous EPA air quality criteria
29     documents, with more detailed  descriptions of individual studies being provided in a series of
30     accompanying annexes.
       August 2005                             1-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1           A list of references published since completion of the 1996 criteria document was made
 2      available to the authors.  The references were selected from information data base searches
 3      conducted by EPA. Additional references have been added to the list (e.g., missed or recently
 4      published papers or "in press" publications) as work has proceeded in creating the draft
 5      document materials. As an aid in selecting pertinent new literature, the authors were also
 6      provided with a summary of issues that need to be addressed in the revised air quality criteria
 7      document for O3. These issues were identified by authors and reviewers of the previous
 8      documents and continue to be expanded, as appropriate, based on public discussions, workshops,
 9      or other  comments received by EPA.
10
11      1.3.2   Organization and Content of the Document
12           This  revised AQCD for O3  and Related Photochemical Oxidants critically assesses
13      scientific information on the health and welfare effects associated with exposure to the
14      concentrations of these pollutants in ambient air.  The document does not provide a detailed
15      literature review; but, rather, discusses cited references that reflect the current state of knowledge
16      on the most relevant issues pertinent to the NAAQS for O3. Although emphasis is placed on
17      discussion of health and welfare  effects information, other scientific data are presented and
18      evaluated in order to provide a better understanding of the nature, sources,  distribution,
19      measurement, and concentrations of O3 and related photochemical oxidants in ambient air,
20      as well as the measurement of population exposure to these pollutants.
21           The main focus of the scientific information discussed in the text comes from literature
22      published since completion of the 1996 O3 AQCD (U.S. Environmental Protection Agency,
23      1996). Emphasis is placed on studies conducted at or near O3 concentrations found in ambient
24      air.  Other studies are included if they contain unique data, such as the documentation of a
25      previously unreported effect or of a mechanism for an observed effect; or if they were multiple-
26      concentration studies designed to provide exposure-response relationships. Generally, this is not
27      an issue  for human clinical or epidemiology studies.  However, for animal toxicology studies,
28      consideration is given mainly to those studies conducted at less than 1 ppm O3. Key information
29      from studies  assessed in the previous O3 AQCD and whose data impacted the derivation of the
30      current NAAQS are briefly summarized in the text, along with specific citations to the previous
31      document. Prior studies are also discussed if they (1) are open to reinterpretation in light of

        August 2005                              1-12         DRAFT-DO NOT QUOTE OR CITE

-------
 1      newer data, or (2) are potentially useful in deriving revised standards for O3. Generally, only
 2      information that has undergone scientific peer review and has been published (or accepted for
 3      publication) through December 2004 is included in this draft document. A few particularly
 4      pertinent and important new studies published or accepted for publication beyond the end of
 5      2004 are also considered.
 6           This document consists of three volumes.  The first volume includes an Executive
 7      Summary and Conclusions, as well as Chapters 1 through 11 of the O3 AQCD. This introductory
 8      chapter (Chapter 1) presents background information on the purpose of the document, legislative
 9      requirements, and the history of past O3 NAAQS regulatory actions, as well as an overview of
10      the organization and content of the document. Chapter 2 provides information on the physics
11      and chemistry of O3 and related photochemical oxidants in the atmosphere. Chapter 3 covers
12      tropospheric O3 environmental concentrations, patterns, and exposure estimates.  The
13      accompanying annexes to each of these background chapters are found in Volume II.
14           Health information pertinent to derivation of the primary O3 NAAQS is then mainly
15      covered in the next several chapters (Chapters 4 through 8). Chapter 4 discusses O3 dosimetry
16      aspects, and Chapters 5, 6, and 7 discuss animal toxicological studies, human health effects from
17      controlled-exposure studies, and epidemiologic studies of ambient air exposure effects on human
18      populations, respectively.  Chapter 8 then provides an integrative and  interpretive evaluation of
19      key information relevant to O3  exposure and health risks, of most pertinence to the review of
20      primary O3 NAAQS. The annexes to these health-related chapters are found in Volume II.
21           The remaining three chapters of the document assess welfare effects information pertinent
22      to the review of secondary O3 NAAQS. Chapter 9 deals with ecological and other
23      environmental effects of O3 and related photochemical oxidants.  Chapter 10 assesses
24      tropospheric O3 involvement in climate change processes, including determination of solar UV
25      flux in Earth's lower atmosphere. Lastly, Chapter 11 discusses O3 effects on man-made
26      materials as a third type of welfare effect of potential concern.  Annex materials related to
27      welfare effects (especially vegetation/ecological effects) are contained in Volume III.
        August 2005                              1-13        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES
  2
  3       Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30)
  4             36:8186-8201.
  5       Federal Register. (1977) Review of the photochemical oxidant and hydrocarbon air quality standards. F. R.
  6             (April 20) 42: 20493-20494.
  7       Federal Register. (1979) National primary and secondary ambient air quality standards: revisions to the national
  8             ambient air quality standards for photochemical oxidants. F. R. (February 8) 44: 8202-8237.
  9       Federal Register. (1982) Air quality criteria document for ozone and other photochemical oxidants. F. R. (March 17)
10             47: 11561.
11       Federal Register. (1983) Review of the national ambient air quality standards for ozone. F. R. (August 22)
12             48:38009.
13       Federal Register. (1984) Draft air quality criteria document for ozone and other photochemical oxidants. F. R.
14             (July 24)  49: 29845.
15       Federal Register. (1985) Science Advisory Board; Clean Air Scientific Advisory Committee; open meeting. F. R.
16             (February 13) 50: 6049.
17       Federal Register. (1986) Science Advisory Board; Clean Air Scientific Advisory Committee; open meeting. F. R.
18             (April 2)  51: 11339.
19       Federal Register. (1992) National ambient air quality standards for ozone; proposed decision. F. R. (August 10)
20             57: 35542-35557.
21       Federal Register. (1993) National ambient air quality standards for ozone - final decision. F. R. (March 9)
22             58: 13008-13019.
23       Federal Register. (1996) National ambient air quality standards for ozone: proposed decision. F. R. (December 13)
24             61:65,716-65,750.
25       Federal Register. (1997) National ambient air quality standards for ozone; final rule. F. R. (July  18)
26             62: 38856-38896.
27       Federal Register. (2000) Air Quality  Criteria for Ozone and Related Photochemical Oxidants; notice; call for
28             information. F. R. (September 26) 65: 57810.
29       Federal Register. (2003) National ambient air quality standards for ozone: final response to remand; final rule. F. R.
30             (January 6) 68: 614-645.
31       U.S. Code. (2003a) Clean Air Act, § 108, air quality criteria and control techniques.. U. S. C. 42: §7408.
32       U.S. Code. (2003b) Clean Air Act, §109, national ambient air quality standards. U. S. C. 42: §7409.
33       U.S. Court of Appeals for the District of Columbia. (1999a) American Trucking Associations, Inc. v. U.S.
34             Environmental Protection Agency. 195 F.3d 4 (D.C. Cir. 1999).
35       U.S. Court of Appeals for the District of Columbia. (1999b) American Trucking Associations, Inc. v. U.S.
36             Environmental Protection Agency. 175 F.3d 1027 (D.C. Cir. 1999).
37       U.S. Court of Appeals for the District of Columbia. (2002) American Trucking Associations, Inc. v. U.S.
38             Environmental Protection Agency. 283 F.3d 355, 378-79 (D.C. Cir. 2002).
39       U.S. Environmental Protection Agency. (1992) Summary of selected new information on effects of ozone on health
40             and vegetation: supplement to 1986 air quality criteria for ozone and other photochemical oxidants.  Research
41             Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
42             Office; report no. EPA/600/8-88/105F. Available from: NTIS, Springfield, VA; PB92-235670.
43       U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
44             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
45             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
46             http://cfpub2.epa.gov/ncea/.
47       U.S. Supreme Court. (2001) Whitman v. American Trucking Association. 531 U.S. 457 (nos. 99-1257 and 99-1426).
48
         August 2005                                    1-14         DRAFT-DO NOT QUOTE OR CITE

-------
 i               2.  PHYSICS AND CHEMISTRY OF OZONE
 2                              IN THE ATMOSPHERE
 3
 4
 5     2.1   INTRODUCTION
 6           Ozone (O3) and other oxidants, such as peroxacyl nitrates and hydrogen peroxide (H2O2)
 7     form in polluted areas by atmospheric reactions involving two main classes of precursor
 8     pollutants, volatile organic compounds (VOCs) and nitrogen oxides (NOX). Carbon monoxide
 9     (CO) is also important for ozone formation in polluted areas. Ozone is thus a secondary
10     pollutant. The formation of O3 other oxidants and oxidation products from these precursors is a
11     complex, nonlinear function of many factors: the intensity and spectral distribution of sunlight;
12     atmospheric mixing and processing on cloud and aerosol particles; the concentrations of the
13     precursors in ambient air; and the rates of chemical reactions of the precursors. Information
14     contained in this chapter and in greater detail in Annex AX2 describes these processes,
15     numerical models that incorporate these processes to calculate O3 concentrations, and techniques
16     for measuring concentrations of ambient oxidants.
17           The atmosphere can be divided into several distinct vertical layers, based primarily on the
18     major mechanisms by which they are heated and cooled. The lowest major layer is the
19     troposphere, which extends from the earth's surface to about 8 km above polar regions and to
20     about 16 km above tropical regions. The planetary boundary layer (PEL) is the lower sublayer
21     of the troposphere, extending from the surface to about 1 or 2 km and is most strongly affected
22     by surface conditions. The stratosphere extends from the tropopause, or the top of the
23     troposphere, to about 50 km in altitude (Annex AX2.2.1). The emphasis in this chapter is placed
24     on chemical and physical processes occurring in the troposphere, in particular in the PEL.  The
25     processes responsible for producing summertime O3 episodes are fairly well understood, as
26     outlined in the previous Air Quality Criteria Document for Ozone and Related Photochemical
27     Oxidants (CD96). This chapter mainly considers topics for which there is substantial new
28     information and on topics that form the basis for discussions in later chapters.
       August 2005                              2-1         DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.2   CHEMICAL PROCESSES INVOLVED IN OZONE FORMATION
 2           AND DESTRUCTION
 3          Ozone occurs not only in polluted urban atmospheres but throughout the troposphere, even
 4     in remote areas of the globe. The same basic processes, involving sunlight-driven reactions of
 5     NOX and VOCs contribute to O3 formation throughout the troposphere. These processes also
 6     lead to the formation of other photochemical products, such as peroxyacetyl nitrate (PAN), nitric
 7     acid (HNO3), and sulfuric acid (H2SO4), and to other compounds, such as formaldehyde (HCHO)
 8     and other carbonyl compounds, such as aldehydes and ketones.
 9          The photochemical formation of O3 in the troposphere proceeds through the oxidation of
10     nitric oxide (NO) to nitrogen dioxide (NO2) by organic (RO2) or hydro-peroxy (HO2) radicals.
11     The photolysis of NO2 yields nitric oxide (NO) and a ground-state oxygen atom, O(3P), which
12     then reacts with molecular oxygen to form O3. Free radicals oxidizing NO to NO2 are formed
13     during the oxidation of VOCs (Annex AX2.2.2).
14          The term VOC refers to all carbon-containing gas-phase compounds in the atmosphere,
15     both biogenic and anthropogenic in origin, excluding carbon monoxide (CO)  and carbon
16     dioxide (CO2). Classes of organic compounds important for the photochemical formation of O3
17     include alkanes, alkenes, aromatic hydrocarbons, carbonyl compounds (e.g., aldehydes and
18     ketones), alcohols, organic peroxides, and halogenated organic compounds (e.g.,  alkyl  halides).
19     This array of compounds encompasses a wide range of chemical properties and lifetimes:
20     isoprene has an atmospheric lifetime of approximately an hour, whereas methane has an
21     atmospheric lifetime of about a decade.
22          In urban areas, compounds representing all classes of VOCs, and CO are important for O3
23     formation.  In nonurban vegetated areas, biogenic VOCs emitted from vegetation tend  to be the
24     most important.  In the remote troposphere, CH4 and CO are the main carbon-containing
25     precursors to O3 formation.  CO also can play an important role in O3 formation in urban areas.
26     The oxidation of VOCs is initiated mainly by reaction with hydroxyl (OH) radicals. The primary
27     source of OH radicals in the atmosphere is the reaction of electronically excited O atoms, O(JD),
28     with water vapor. O(JD) is produced by the photolysis of O3 in the Hartley bands.  In polluted
29     areas, the photolysis of aldehydes (e.g., HCHO),  nitrous acid (HONO) and hydrogen
30     peroxide (H2O2) can also be  significant sources of OH or HO2 radicals that can rapidly  be
31     converted to OH (Eisele et al., 1997). Ozone can oxidize alkenes, and, at night, when they are

       August 2005                              2-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1      most abundant, NO3 radicals also oxidize alkenes.  In coastal environments and other selected
 2      environments, atomic Cl and Br radicals can also initiate the oxidation of VOCs (Annex
 3      AX2.2.3).
 4           There are a large number of oxidized nitrogen containing compounds in the atmosphere
 5      including NO, NO2, NO3, HNO2, HNO3, N2O5, HNO4, PAN and its homologues, other organic
 6      nitrates and particulate nitrate. Collectively these species are referred to as NOy. Oxidized
 7      nitrogen compounds are emitted to the atmosphere mainly as NO which rapidly interconverts
 8      with NO2 and so NO and NO2 are often "lumped" together into their own group or family,
 9      or NOX.  NOX can be oxidized to reservoir and termination species (PAN and its homologues,
10      organic nitrates, HNO3, HNO4 and particulate nitrate).  These reservoir and termination species
11      are referred to as NOZ. The major reactions involving inter-conversions of oxidized nitrogen
12      species are discussed in Annex AX2.2.4.
13           The photochemical cycles by which the oxidation of hydrocarbons leads to O3 production
14      are best understood by considering the oxidation of methane, structurally the simplest VOC.
15      The CH4 oxidation cycle serves as a model for the chemistry of the relatively clean or unpolluted
16      troposphere (although this is a simplification because vegetation releases large quantities of
17      complex VOCs, such as isoprene, into the atmosphere). In the polluted atmosphere, the
18      underlying chemical principles are the same, as discussed in Annex AX2.2.5. The conversion of
19      NO to NO2 occurring with the oxidation of VOCs is accompanied by the production of O3 and
20      the efficient regeneration of the OH  radical, which in turn can react with other VOCs.
21      A schematic overview showing the major processes involved in O3 production and loss in the
22      troposphere and stratosphere is given in Figure 2-1.
23           The oxidation of alkanes and alkenes in the atmosphere has been treated in depth in CD96
24      and is updated in Annexes AX2.2.6  and AX2.2.7. In contrast to simple hydrocarbons containing
25      one or two carbon atoms, detailed kinetic information about the gas phase oxidation pathways of
26      many anthropogenic hydrocarbons (e.g., aromatic compounds, such as benzene  and toluene),
27      biogenic hydrocarbons (e.g., isoprene, the monoterpenes), and their intermediate oxidation
28      products (e.g., epoxides, nitrates, and carbonyl compounds) is lacking. Reaction with OH
29      radicals represents the major loss process for alkanes. Reaction with chlorine atoms is an
30      additional sink for alkanes. Stable products of alkane photooxidation are known to include
31      carbonyl compounds,  alkyl nitrates,  and J-hydroxycarbonyls.  Major uncertainties in the

        August 2005                               2-3         DRAFT-DO NOT QUOTE OR CITE

-------
                                           Stratosphere
                                             Non.Po.ar
      Figure 2-1. Schematic overview of O3 photochemistry in the stratosphere and troposphere.
1     atmospheric chemistry of the alkanes concern the chemistry of alkyl nitrate formation; these
2     uncertainties affect the amount of NO-to-NO2 conversion occurring and, hence, the amounts
3     of O3 formed during photochemical degradation of the alkanes.
4           The reaction of OH radicals with aldehydes produced during the oxidation of alkanes
5     forms acyl (R'CO) radicals, and acyl peroxy radicals (R'C(O)-O2) are formed by the further
6     addition of O2. As an example, the oxidation of ethane (C2H5-H) yields acetaldehyde
7     (CH3- CHO). The reaction of CH3- CHO with OH radicals yields acetyl radicals (CH3- CO).
8     The acetyl radicals will then participate with O2 in a termolecular recombination reaction to form
      August 2005
2-4
DRAFT-DO NOT QUOTE OR CITE

-------
 1      acetyl peroxy radicals, which can then react with NO to form CH3 + CO2 or they can react
 2      with NO2 to form PAN.  PAN acts as a temporary reservoir for NO2. Upon the thermal
 3      decomposition of PAN, either locally  or elsewhere, NO2 is released to participate in the O3
 4      formation process again.
 5           Alkenes react in ambient air with OH, NO3, and Cl radicals and with O3. All of these
 6      reactions are important atmospheric transformation processes, and all proceed by initial addition
 7      to the >C=C< bonds. Products of alkene photooxidation include carbonyl compounds,
 8      hydroxynitrates and nitratocarbonyls,  and decomposition products from the energy-rich
 9      biradicals formed in alkene-O3 reactions. Major uncertainties in the atmospheric chemistry of
10      the alkenes concern the products and mechanisms of their reactions with O3, especially the yields
11      of free radicals that participate in O3 formation.  Examples of oxidation mechanisms of complex
12      alkanes and alkenes can be found in comprehensive texts such as Seinfeld and Pandis (1998).
13           The oxidation of aromatic hydrocarbons constitutes an important component of the
14      chemistry of O3 formation in urban atmospheres (Annex AX2.2.8). Virtually all of the important
15      aromatic hydrocarbon precursors emitted in urban atmospheres are lost through reaction with the
16      hydroxyl radical.  Loss rates for these compounds vary from slow (i.e., benzene) to moderate
17      (e.g., toluene), to very rapid (e.g., xylene and trimethylbenzene isomers).  These loss rates are
18      very well understood at room temperature and atmospheric pressure and numerous experiments
19      have  been conducted that verify this.  However, the mechanism for the oxidation of aromatic
20      hydrocarbons following reaction with OH is poorly understood, as evident from the poor mass
21      balance of the reaction products.  The mechanism for the oxidation of toluene has been studied
22      most thoroughly and there is general agreement on  the initial steps in the mechanism.  However,
23      at present there is no promising approach for resolving the remaining issues concerning the later
24      steps. The oxidation of aromatic hydrocarbons also leads  to particle formation which could
25      remove gas-phase constituents that participate in O3 formation.  The chemistry of secondary
26      organic aerosol formation from gaseous precursors  was  summarized in the latest AQCD for
27      particulate matter.
28           The reactions of oxygenated VOCs are also important components of O3 formation (Annex
29      AX2.2.9).  They may be produced either by the  oxidation  of hydrocarbons or they may be
30      present in ambient air as the result of direct emissions. For example, motor vehicles and some
31      industrial processes emit formaldehyde and vegetation emits methanol.

        August 2005                              2-5         DRAFT-DO NOT QUOTE OR CITE

-------
 1           As much as 30% of the carbon in hydrocarbons in many urban areas is in the form of
 2      aromatic compounds. Yet, mass balance analyses performed on irradiated smog chamber
 3      mixtures of aromatic hydrocarbons indicate that only about one-half of the carbon is in the form
 4      of compounds that can be identified.  The situation is not much better for some smaller
 5      anthropogenic hydrocarbons. For example, only about 60% of the initial carbon can be
 6      accounted for in the OH initiated oxidation of 1,3-butadiene. About two-thirds of the initial
 7      carbon can be identified in product analyses of isoprene oxidation.  Adequate analytical
 8      techniques needed to identify and quantify key intermediate species are not available for many
 9      compounds.  In addition, methods to synthesize many of the suspected intermediate compounds
10      are not available so that laboratory studies of their reaction kinetics cannot be performed.
11      Similar considerations apply to the oxidation of biogenic hydrocarbons besides isoprene.
12           In addition to reactions occurring in the gas phase, reactions occurring on the surfaces of or
13      within cloud droplets and airborne particles also occur. Their collective surface area is huge
14      implying that collisions with gas phase species occur on very short time scales. In addition to
15      hydrometeors (e.g., cloud and fog droplets and snow and ice crystals) there are also potential
16      reactions involving atmospheric  particles of varying composition (e.g., wet [deliquesced]
17      inorganic particles, mineral dust, carbon chain agglomerates and organic carbon particles) to
18      consider. Most of the well-established multiphase reactions tend to reduce the rate of O3
19      formation in the polluted troposphere. Removal of HOX and NOX onto hydrated particles will
20      reduce the production of O3. However, the photolysis of HONO formed in reactions such as
21      these can increase the production of O3. The reactions of Br and Cl containing radicals
22      deplete O3 in selected environments such as the Arctic during spring, the tropical  marine
23      boundary layer and inland salt lakes.  Direct reactions of O3 and atmospheric particles appear to
24      be too slow to reduce O3  formation significantly at typical ambient PM levels. In addition, the
25      oxidation of hydrocarbons by Cl radicals could lead to the rapid formation of peroxy radicals and
26      higher rates of O3 production in selected coastal environments. It should be stressed that
27      knowledge of multiphase processes is still  evolving and there are still many questions that
28      remain to be answered as outlined in Annex AX2.2.10.
29           The oxidants, other than O3, that are formed from the chemistry described above could
30      exert effects on human health and perhaps  also on vegetation.  Gas phase oxidants include
31      PAN, H2O2 and CH3OOH and other organic hydroperoxides (Annex AX2.2). In addition to

        August 2005                               2-6        DRAFT-DO NOT QUOTE OR CITE

-------
 1     transfer from the gas phase, oxidants can be formed by photochemical reactions occurring in
 2     particles (Annex 2.2.10.6).  However, the pathways leading to the formation of oxidants in the
 3     particle phase are not as well understood as they are in the gas phase.
 4
 5
 6     2.3   METEOROLOGICAL PROCESSES AFFECTING OZONE
 7           Since CD96, substantial new information about transport processes has become available
 8     from numerical models, field experiments and satellite-based observations. Ozone is produced
 9     naturally by photochemical reactions in the stratosphere as shown in Figure 2-1.  Some of this O3
10     is transported downward into the troposphere throughout the year, with maximum contributions
11     during late winter and early spring mainly in a process known as tropopause folding.  Figure
12     2-2a shows a synoptic situation associated with a tropopause folding event. A vertical cross
13     section taken through the atmosphere from a to a' is shown in Figure 2-2b. In this figure the
14     tropopause fold is shown folding downward above and slightly behind the surface cold front,
15     bringing stratospheric air with it. Although the tropopause is drawn with a solid line, it should
16     not be taken to mean that it is a material surface, through which there is no exchange.  Rather
17     these folds should be thought of as regions in which mixing of tropospheric and stratospheric air
18     is occurring (Shapiro,  1980). This imported stratospheric air contributes to the natural
19     background of O3 in the troposphere, especially in the free troposphere.  It should be noted that
20     there is considerable uncertainty in the magnitude and distribution of this potentially important
21     source of tropospheric O3.  Stratospheric intrusions that reach the surface are rare. Much more
22     common are intrusions which penetrate only to the middle and upper troposphere. However, O3
23     transported to the upper and middle troposphere can still affect surface concentrations through
24     various exchange mechanisms that mix air from the free troposphere with air in the planetary
25     boundary layer.  Substantial photochemical production of O3 in the  troposphere also begins in
26     late winter and early spring; therefore, it cannot be assumed that O3 present at these times is only
27     stratospheric in origin.  The basic atmospheric dynamics and thermodynamics of stratospheric-
28     tropospheric exchange are  outlined in Annex AX2.3.1.
29
        August 2005                              2-7        DRAFT-DO NOT QUOTE OR CITE

-------
              12UTC, 23Feb94
103.6kPa
Figure 2-2a.   Surface weather chart showing sea level (MSL) pressure (kPa), and
              surface fronts.

Source:  Stall (2000).
                     D)
                     'o>
                     I
                           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).
August 2005
    2-8
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Our understanding of the meterological processes associated with summertime O3 episodes
 2      remains basically the same as outlined in CD96. Major episodes of high O3 concentrations in the
 3      eastern United States and in Europe are associated with slow moving, high pressure systems.
 4      High pressure systems during the warmer seasons are associated with the sinking of air, resulting
 5      in warm, generally cloudless skies, with light winds. The sinking of air results in the
 6      development of stable conditions near the surface which inhibit or reduce the vertical mixing
 7      of O3 precursors.  The combination of inhibited vertical mixing and light winds minimizes the
 8      dispersal of pollutants emitted in urban areas, allowing their concentrations to build up.
 9      Photochemical activity involving these precursors is enhanced because of higher temperatures
10      and the availability of sunlight.  In the eastern United States, high O3 concentrations during a
11      large scale episode can extend over hundreds of thousands of square kilometers for several days.
12      These conditions have been described in greater detail in CD96. The transport of pollutants
13      downwind of major urban centers is characterized by the development of urban plumes.
14      However, the presence of mountain barriers limits mixing as in Los Angeles and Mexico City
15      and will result in a higher frequency and duration of days with high O3 concentrations. Ozone
16      concentrations in southern urban areas, such as Houston, TX and Atlanta, GA tend to decrease
17      with increasing wind speed.  In northern cities such as Chicago, IL; New York, NY; Boston,
18      MA;  and Portland, ME, the average O3 concentrations over the metropolitan areas increase with
19      wind speed indicating that transport of O3 and its precursors from upwind areas is important
20      (Husar and Renard, 1998; Schichtel and Husar, 2001).
21           Ozone and other secondary pollutants are determined by meteorological and chemical
22      processes extending typically over spatial scales of several hundred kilometers (e.g., Civerolo
23      et al., 2003; Rao et al., 2003). An analysis of the output of regional model  studies conducted by
24      Kasibhatla and Chameides (2000) suggests that O3 can be transported over a few thousand
25      kilometers in the upper boundary layer of the eastern half of the United States during specific O3
26      episodes.  Convection is capable of transporting O3 and its precursors vertically through the
27      troposphere as shown in Annex AX2.3.2.  Nocturnal low level jets (LLJs) can also transport
28      pollutants hundreds of kilometers (Annex AX2.3.3). Schematic diagrams showing the
29      atmospheric conditions during the formation of low level jets and the regions in which they are
30      most prevalent are given in Figures 2-3 and 2-4. They have also been observed off the coast of
31

        August 2005                               2-9         DRAFT-DO NOT QUOTE OR CITE

-------
                    2-
                 •-  1-1
                 
-------
 1      California. Turbulence associated with LLJs can bring these pollutants to the surface and result
 2      in secondary O3 maxima in the early morning in many locations (Corsmeier et al., 1997).
 3           Aircraft observations indicate that there can be substantial differences in mixing ratios of
 4      key species between the surface and the atmosphere above (Fehsenfeld et al.,  1996; Berkowitz
 5      and Shaw, 1997). In particular, mixing ratios of O3 can be higher in the lower free troposphere
 6      (aloft) than in the planetary boundary layer during multiday O3 episodes (Taubmann et al.,
 7      2004). These conditions are illustrated schematically in Figure 2-5. Convective processes and
 8      small scale turbulence transport O3 and other pollutants both upward and downward throughout
 9      the planetary boundary layer and the free troposphere. Ozone and its precursors can be
10      transported vertically by convection into the upper part of the mixed layer on one day, then
11      transported overnight as a layer of elevated mixing ratios, and then entrained into a growing
12      convective boundary layer downwind and brought back down to the surface. High
13      concentrations of O3 showing large diurnal variations at the surface in southern New England
14      were associated with the presence of such layers (Berkowitz et al., 1998). Because of wind
15      shear, winds several hundred meters above the ground can bring pollutants from the west, even
16      though surface winds are from the southwest during periods of high O3 in the eastern United
17      States (Blumenthal  et al., 1997).  These considerations suggest that in many areas of the United
18      States, O3 formation involves processes occurring over hundreds if not thousands of square
19      kilometers.
20           Although the vast majority of measurements are made near the Earth's surface, there is
21      substantial photochemistry and transport of O3 occurring above the boundary layer in the free
22      troposphere.  In the free troposphere, pollutants are chemically more stable and can be
23      transported over much longer distances and O3 is produced more efficiently than in the planetary
24      boundary layer.  Results from the Atmosphere/Ocean Chemistry Experiment (AEROCE)
25      indicated that springtime maxima in surface O3 over the western North Atlantic Ocean result
26      from tropopause folding in close proximity to convective  clouds (Annex AX2.3.4). The
27      convection lifts O3 and its precursors to the free troposphere where they mix with O3 from the
28      stratosphere and the mixture is transported eastward.  Results from the North Atlantic Regional
29      Experiment (Annex AX2.3.4) indicated that summertime  air is transported along the East Coast
30      northeastward and upward ahead of cold fronts. New England and the Maritime Provinces of
31      Canada receive substantial amounts of O3  and other pollutants through this mechanism.

        August 2005                              2-11        DRAFT-DO NOT QUOTE OR CITE

-------
                                         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 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).
1     Pollutants transported in this way can then be entrained in stronger and more stable westerly

2     winds aloft and can travel across the North Atlantic Ocean.  The pollutants can then be brought

3     to the surface by subsidence in high pressure systems (typically behind the cold front in advance

4     of the one mentioned above).  Thus, pollutants from North America can be brought down either

5     over the North Atlantic Ocean or in Europe.  Pollutants can be transported across the North

6     Pacific Ocean from Asia to North America in a similar way. Behind an advancing cold front,

7     cold and dry stratospheric air is also being transported downward and southward. Stratospheric
      August 2005
               2-12
             DRAFT-DO NOT QUOTE OR CITE

-------
 1      constituents and tropospheric constituents can then mix by small-scale turbulent exchange
 2      processes.  The results of these studies suggest that the mechanisms involved in the long-range
 3      transport of O3 and its precursors are closely tied to the processes involved in stratospheric-
 4      tropospheric exchange.
 5
 6
 7      2.4   RELATIONS OF OZONE TO ITS PRECURSORS
 8           The local rate of O3 formation depends on atmospheric conditions such as the availability
 9      of solar ultraviolet radiation capable of initiating photolysis reactions, air temperatures and the
10      concentrations of chemical precursors (Annex AX2.3.5). The dependence of daily maximum
11      8-h O3 concentrations on daily maximum temperature is illustrated in Figure 2-6 for the
12      Baltimore,  MD area.  As can be seen, O3 concentrations tend to increase with temperature
13      (r = 0.74).  However, this trend is absent in data from Phoenix, AZ as can be seen in Figure 2-7
14      (r = 0.14).  These figures show that relations of O3 to precursor variables are location-specific
15      and relations observed in one area cannot be readily extrapolated to another. Factors that may be
16      responsible for the differences in O3 behavior in the two areas are discussed in Section
17      AX2.3.5.3.
18           Rather than varying directly with emissions of its precursors, O3 changes in a nonlinear
19      fashion with the concentrations of its precursors (Annex AX2.4). At the low NOX concentrations
20      found in most environments, ranging from remote continental areas to rural and suburban areas
21      downwind  of urban centers (low - NOX regime), the net production  of O3 increases with
22      increasing NOX.  At the high NOX concentrations found in downtown metropolitan areas,
23      especially near busy streets and roadways, and in power plant plumes there is scavenging
24      (titration) of O3 by reaction with NO (high - NOX regime).  In between these two regimes there is
25      a transition stage in which O3 shows only a weak dependence on NOX concentrations.  In the
26      high - NOX regime, NO2 scavenges OH radicals which would otherwise oxidize VOCs to
27      produce peroxy radicals, which in turn would oxidize NO to NO2.  In this regime, O3 production
28      is limited by the availability of free radicals. The production of free radicals is in turn limited by
29      the availability of solar UV radiation capable of photolyzing O3 (in the Hartley bands) or
30      aldehydes and/or by the abundance of VOCs whose oxidation produce more radicals than they
31      consume.  In the low-NOx regime, the overall effect of the oxidation of VOCs is to generate (or

        August 2005                              2-13       DRAFT-DO NOT QUOTE OR CITE

-------

Q.
S

•fi
2

O
O
o
£
00

E

S
>,

Q 2U









D
D D
°_D_ °n°l
o l°nPB PI
n " a """pSni" 01"'

•" -i^a^PJUo
n n a° o nn










"
iDu
|l
fil

Dn












n°i

||
IB
1
t a
DDu
a a
DO








B
POO B
B Bal
B 1

1 :
i ;
IDJ
n o=fT— ,

°°| n°n




n ° 0
°D B

DPD H^nnB Q
"
I llPll°Sn • °
1 nl DD
|||||5Bno
gnBBnB QQ n
DHBDB °
pHln
i H°
3nSn
•J DD




                            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).
               140'
            a
            a.
            o
            o
            00
            E
            n
            D
               120'
60
               40
               20
                      5     10    15     20    25    30     35    40    45
                                      Tmax at PHX °C
                                                                      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).
August 2005
                          2-14
DRAFT-DO NOT QUOTE OR CITE

-------
 1      at least not consume) free radicals, and O3 production varies directly with NOX. There are a
 2      number of ways to refer to the chemistry in these two chemical regimes.  Sometimes the terms
 3      VOC-limited and NOx-limited are used.  However, there are difficulties with this usage because
 4      (1) VOC measurements are not as abundant as they are for nitrogen oxides, (2) rate coefficients
 5      for reaction of individual VOCs with free radicals vary over an extremely wide range, and (3)
 6      consideration is not given to CO nor to reactions that can produce free radicals without involving
 7      VOCs. The terms NOx-limited and NOx-saturated (e.g., Jaegle et al., 2001) will be used
 8      wherever possible to more adequately  describe these two regimes. However, the terminology
 9      used in original articles will also be used here.
10           The chemistry of OH radicals, which are responsible for initiating the oxidation of
11      hydrocarbons, shows behavior similar to that for O3 with respect to NOX concentrations (Hameed
12      et al.,  1979; Pinto et al., 1993; Poppe et al., 1993; Zimmerman and Poppe, 1993).  These
13      considerations introduce a high degree of uncertainty into attempts to relate changes in O3
14      concentrations to emissions of precursors. There are no definitive rules governing the levels
15      of NOX at which the transition from NOx-limited to NOx-saturated conditions occurs. The
16      transition between these two regimes is highly spatially and temporally dependent and depends
17      also on the nature and abundance of the hydrocarbons that are present.
18           Trainer et al. (1993) and Olszyna et al. (1994) have shown that O3 and NOy are highly
19      correlated in rural areas in the eastern United States.  Trainer et al. (1993) also showed that O3
20      levels correlate even better with NOZ than with NOy,  as may be expected  because NOZ represents
21      the amount of NOX that has been oxidized, forming O3 in the process. NOZ is equal to the
22      difference between measured total reactive nitrogen (NOy) and NOX and represents the summed
23      products of the oxidation of NOX.  NOZ is composed mainly of HNO3, PAN and other organic
24      nitrates, particulate nitrate,  and HNO4.
25           Trainer et al. (1993) also suggested that the slope of the regression line between O3
26      and NOZ can be used to estimate the rate of O3 production per NOX oxidized (also known as
27      the O3 production efficiency,  or OPE). Ryerson et al. (1998, 2001) used measured correlations
28      between O3 and NOZ to identify different rates of O3 production in plumes from large point
29      sources.  A number of studies in the planetary boundary layer over the continental United States
30      have found that the OPE ranges typically from one to nearly ten.  However, it may be higher in
        August 2005                              2-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1      the upper troposhere and in certain areas, such as the Houston-Galveston area. Observations
 2      indicate that the OPE depends mainly on the abundance of NOX.
 3           Various techniques have been proposed to use ambient NOX and VOC measurements to
 4      derive information about the dependence of O3 production on their concentrations.  For example,
 5      it has been suggested that O3 formation in individual urban areas could be understood in terms of
 6      measurements of ambient NOX and VOC concentrations during the early morning (e.g., National
 7      Research Council, 1991). In this approach, the ratio of summed (unweighted) VOC to NOX is
 8      used to determine whether conditions were NOx-limited or VOC limited. This procedure is
 9      inadequate because it omits many factors that are important for O3 production such as the impact
10      of biogenic VOCs (which are typically not present in urban  centers during early morning);
11      important differences in the ability of individual VOCs to generate free radicals (rather than just
12      total VOC) and other differences in O3 forming potential for individual VOCs (Carter et al.,
13      1995); and changes in the VOC to NOX ratio due to photochemical reactions and deposition as
14      air moves downwind from urban areas (Milford et  al., 1994).
15           Photochemical production of O3 generally occurs simultaneously with the production of
16      various other species such as nitric acid (HNO3), organic nitrates, and other oxidants such as
17      hydrogen peroxide.  The relative rate of production of O3 and other species varies depending on
18      photochemical conditions, and can be used to provide information about O3-precursor
19      sensitivity.  Sillman (1995) and Sillman and He (2002) identified several secondary reaction
20      products that show different correlation patterns for NOx-limited and NOx-saturated conditions.
21      The most important correlations are for O3 versus NOy, O3 versus NOZ, O3 versus HNO3,
22      and H2O2 versus HNO3. The correlations between  O3 and NOy, and O3 and NOZ are especially
23      important because measurements of NOy and NOX are more  widely available than for VOCs.
24      Measured O3 versus NOZ (Figure 2-8) shows distinctly different patterns in different locations.
25      In rural areas and in urban areas such as Nashville, TN, O3 is highly correlated with NOZ. By
26      contrast, in Los Angeles, CA, O3 is not as highly correlated with NOZ, and the rate of increase
27      of O3 with NOZ is lower and the O3 concentrations for a given NOZ value are generally lower.
28      The different O3 versus NOZ relations in Nashville, TN and Los Angeles, CA reflects the
29      difference between NOx-limited conditions in Nashville versus an approach to NOX- saturated
30      conditions in Los Angeles.
        August 2005                              2-16        DRAFT-DO NOT QUOTE OR CITE

-------
                                                  X
                                                        X
                                                             X
                                                             X
                                                      X X
                                                     X
                                                                 X
                                        10
                                                  20
                                              NOZ (ppb)
    30
40
       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).
1
2
3
4
5
6
7
            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.
 9     2.5   THE ROLE OF CHEMISTRY-TRANSPORT MODELS IN
10           UNDERSTANDING ATMOSPHERIC OZONE
11          Chemistry-transport models (CTMs) are used to improve understanding of atmospheric
12     chemical processes and to develop control strategies (Annex AX2.5).  The main components of a
13     CTM are summarized in Figure 2-9. Models such as the CMAQ (Community Model for Air
14     Quality) system incorporate numerical algorithms describing the processes shown in Figure 2-9.
       August 2005
                                              2-17
DRAFT-DO NOT QUOTE OR CITE

-------
                     Initial/Boundary
                     Conditions and
                   Continuous Updates
                      of Met. Fields
                    from Observations
Meteorological
    Model
     Emissions
       Model
    Anthropogenic
    (point, area sources)
                                                                 Biogenic Emissions

X

' \

1
                                                                           Aerosol
                                                                           Chemistry
                                                                             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.
 1     Also shown in Figure 2-9 is the meteorological model used to provide the inputs for calculating
 2     the transport of species in the CTM. Meteorological models, such as the MM5 model, which
 3     supply these inputs to the CTMs mentioned above, also provide daily weather forecasts. The
 4     domains of these models extend typically over areas of millions of square kilometers.
 5          Because these models are computationally intensive, it is  often impractical to run them
 6     over larger domains without sacrificing some features. For these reasons, both the
 7     meteorological model and the CTM rely on boundary conditions that allow processes occurring
 8     outside the model domain to influence their predictions.  The entire system, consisting of
 9     meteorological model, emissions processor, and output processors shown in Figure 2-9
10     constitutes the framework of EPA's Models-3.
       August 2005
        2-18
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Because of the large number of chemical species and reactions that are involved in the
 2      oxidation of realistic mixtures of anthropogenic and biogenic hydrocarbons, condensed
 3      mechanisms must be used in atmospheric models. These mechanisms are tested by comparison
 4      with smog chamber data. However, the existing chemical mechanisms often neglect many
 5      important processes such as the formation and subsequent reactions of long-lived carbonyl
 6      compounds, the incorporation of the most recent information about intermediate compounds, and
 7      heterogeneous reactions involving cloud droplets and aerosol particles.
 8           Emissions inventories are compiled for O3 precursors ( NOX, VOCs, and CO). Recent
 9      estimates and more detailed discussions of the estimates are given in Annex AX2.5.2.
10      Anthropogenic NOX emissions are associated with combustion processes. Most emissions are in
11      the form of NO, which is formed at high combustion temperatures from atmospheric nitrogen
12      and oxygen and from fuel nitrogen.  The two largest sources of NOX are electric power
13      generation plants and motor vehicles. Emissions of NOX therefore are highest in areas having a
14      high density of power plants and in urban regions having high traffic density. Natural NOX
15      sources include stratospheric intrusions, lightning, soils, and wildfires. Lightning, fertilized
16      soils, and wildfires are the major natural sources of NOX in the United States. Both nitrifying
17      and denitrifying organisms in the soil can produce NOX, mainly in the form of NO.  Emission
18      rates depend mainly on fertilization levels and soil temperature and moisture. Spatial and
19      temporal variability in soil NOX emissions leads to considerable uncertainty in emissions
20      estimates.  Nationwide, about 60% of lightning generated NOX occurs in the southern United
21      States and about 60%  the total NOX emitted by soils occurs in the central corn belt of the United
22      States. The oxidation of NH3 emitted mainly by livestock and soils, leads to the formation of a
23      small amount of NO.  Uncertainties in natural NOX inventories are much larger than for
24      anthropogenic NOX emissions.
25           Hundreds of VOCs, containing mainly two to about twelve carbon atoms, are emitted by
26      evaporation and combustion processes from a large number of anthropogenic sources.  The two
27      largest source categories in the U.S.  EPA's emissions inventories are industrial processes and
28      transportation.  Emissions of VOCs from highway vehicles account for roughly two-thirds of the
29      transportation-related  emissions.
30           The accuracy of VOC emission estimates is difficult to determine, both for stationary and
31      mobile sources. Evaporative emissions, which depend  on temperature and other environmental

        August 2005                              2-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1      factors, compound the difficulties of assigning accurate emission factors. In assigning VOC
 2      emission estimates to the mobile source category, models are used that incorporate numerous
 3      input parameters (e.g., type of fuel used, type of emission controls, age of vehicle), each of
 4      which has some degree of uncertainty.  Data for the ratio of CO to NOX and NMHC to NOX in
 5      traffic tunnels (e.g., Pierson et al., 1990) indicated that emissions of NMHCs and CO from motor
 6      vehicles have been underestimated by as much as a factor of two (based on the assumption that
 7      emissions of NOX were reasonably well represented in the inventories). However, the results of
 8      more recent studies have been mixed, with many studies showing agreement to within ±50%
 9      (summarized in Air Quality Criteria for Carbon Monoxide [U.S. Environmental Protection
10      Agency, 2000]). Remote sensing data (Stedman et al., 1991) indicate that about 50% of NMHC
11      and CO emissions are produced by about 10% of the vehicles.  These "super-emitters" are
12      typically poorly maintained.  Vehicles of any age engaged in off-cycle operations (e.g., rapid
13      accelerations) emit much more than if operated in normal driving modes.
14           Vegetation emits  significant quantities of VOCs such as terpenoid compounds  (isoprene,
15      2-methyl-3-buten-2-ol, monoterpenes), compounds in the hexanal family, alkenes, aldehydes,
16      organic acids, alcohols, ketones, and alkanes. The major chemicals emitted by plants are
17      isoprene (35%), 19 other terpenoid compounds and 17 non-terpenoid compounds including
18      oxygenated compounds (40%) (Guenther et al., 2000). Coniferous forests represent  the largest
19      source on a nationwide basis, because of their extensive land coverage. Most biogenic emissions
20      occur during the summer, because of their dependence on temperature and incident sunlight.
21      Biogenic emissions are also higher in southern states than in northern states for these reasons and
22      because of species variations. The uncertainty in natural emissions is about  50% for isoprene
23      under midday summer conditions and  could be as much as a factor often higher for  some
24      compounds (Guenther et al.,  2000). Uncertainties in both biogenic and anthropogenic VOC
25      emission inventories prevent determination of the relative contributions of these two categories
26      at least in many urban areas.  On the regional and global scales, emissions of VOCs  from
27      vegetation are much larger than those from anthropogenic  sources.
28           The performance of CTMs must be evaluated by comparison with field data as part of a
29      cycle of model evaluations and subsequent improvements. Discrepancies between model
30      predictions and observations can be used to point out gaps in current understanding of
31      atmospheric chemistry  and to spur improvements in parameterizations of atmospheric chemical

        August 2005                              2-20        DRAFT-DO NOT QUOTE OR CITE

-------
 1     and physical processes. Model evaluation does not merely involve a straightforward comparison
 2     between model predictions and the concentration field of the pollutant of interest. Such
 3     comparisons may not be meaningful because it is difficult to determine if agreement between
 4     model predictions and observations truly represents an accurate treatment of physical and
 5     chemical processes in the CTM or the effects of compensating errors in complex model routines.
 6     Ideally, each of the model components (emissions inventories, chemical mechanism,
 7     meteorological driver) should be evaluated individually, however this is rarely done in practice.
 8     A comparison between free radical concentrations predicted by parameterized chemical
 9     mechanisms and observations suggests that radical concentrations were overestimated by current
10     chemical mechanisms for NOX concentrations <~5 ppb (Volz-Thomas et al., 2003).
11           In addition to comparisons  between concentrations of calculated and measured species,
12     comparisons of correlations between measured primary VOCs and NOX and modeled VOCs
13     and NOX are especially useful for evaluating results from chemistry-transport models. Likewise,
14     comparisons of correlations between measured species and modeled species can be used  to
15     provide information about the chemical state of the atmosphere and to evaluate model
16     representations (including O3 production per NOX, O3-NOX-VOC sensitivity, and the general
17     accuracy of photochemical representations). A CTM that demonstrates the accuracy of both its
18     computed VOC and NOX in comparison with ambient measurements and the spatial and temporal
19     relations among the critical secondary species  associated with O3 has a higher probability of
20     representing O3-precursor relations correctly than one that does not.
21
22
23     2.6   TECHNIQUES FOR MEASURING OZONE AND ITS PRECURSORS
24           Several techniques have been developed  for sampling and measurement of O3 in the
25     ambient atmosphere at ground level.  Although the chemiluminescence method (CLM) using
26     ethylene is designated as the Federal Reference Method for measuring O3, monitoring in  the
27     NAMS/SLAMS networks is conducted mainly with UV absorption spectrometry using
28     commercial short path instruments.  The primary reference standard instrument is a relatively
29     long-path UV absorption spectrometer maintained under carefully controlled  conditions at NIST
30     (e.g., Fried and Hodgeson, 1982). Episodic measurements are made with a variety of other
       August 2005                              2-21        DRAFT-DO NOT QUOTE OR CITE

-------
 1     techniques based on the principles of chemiluminescence, electrochemistry, differential optical
 2     absorption spectroscopy (DOAS), and LIDAR.
 3           In principle, each of these methods is subject to interference.  Kleindienst et al. (1993)
 4     found that water vapor could cause a positive interference in the CLM with an average positive
 5     deviation of 3% ozone/% water vapor at 25 °C.  The UV absorption spectrometers are subject to
 6     positive interference by atmospheric constituents, such as certain aromatic aldehydes that absorb
 7     at the 253.7 nm Hg resonance line and are at least partially removed by the MnO2 scrubber.
 8     Parrish and Fehsenfeld (2000) did not find any evidence for significant interference (>1%) in
 9     flights through the Nashville urban plume. The  same group tested the air of Houston, El Paso,
10     Nashville, Los Angeles, San Francisco and the East Coast. They observed only one instance of
11     substantive positive interference defined as the UV absorption technique showing more than a
12     few ppb more than the CLM.  This occurred in Laporte, TX under heavily polluted conditions
13     and a low inversion, at night (Jobson et al., 2004).  Leston et al. (2005) observed interference of
14     from  20 to 40 ppb in Mexico City and in a separate smog chamber study. However, the
15     concentrations of relevant compounds were many times higher than found in U. S. urban areas.
16     Thus, it is not likely that such interference could be more than a few ppb under typical ambient
17     conditions. However, Leston et al. (2005) suggested that the use of other materials in the
18     scrubber could have eliminated the interference  seen in their smog chamber study.
19           By far,  most measurements of NO are made using the CLM, based on its reaction with O3.
20     Commercial instruments for measuring NO and  NO2 are constructed with an internal converter
21     for reducing NO2 to NO and then measuring NO by the CLM.  In principle, this technique yields
22     a measurement of NOX with NO2 found by difference between NOX and NO. However, these
23     converters also reduce NOZ compounds thereby  introducing a positive interference in the
24     measurement of NO2.  Other methods for measuring NO2 are available, such as photolytic
25     reduction followed by CLM, laser-induced fluorescence and DOAS. However, they require
26     further development before they can be used for routine monitoring in the NAMS/SLAMS
27     networks.  More detailed  descriptions  of the issues and techniques discussed above and
28     techniques for measuring HNO3 and VOCs can be found in Annex AX2.6.
29
30
        August 2005                              2-22       DRAFT-DO NOT QUOTE OR CITE

-------
 1      2.7   SUMMARY
 2           Ozone (O3) is formed by atmospheric reactions involving two classes of precursor
 3      compounds, volatile organic compounds (VOCs) and nitrogen oxides (NOX). Ozone is thus a
 4      secondary pollutant.  Ozone is ubiquitous throughout the atmosphere; it is present even in remote
 5      areas of the globe. The photochemical oxidation of almost all anthropogenic and biogenic VOCs
 6      is initiated by reaction with hydroxyl (OH) radicals. At night, when they are most
 7      abundant, NO3 radicals also oxidize alkenes.  In coastal and other select environments, Cl and Br
 8      radicals can also initiate the oxidation of VOCs.
 9           In urban areas, basically all classes of VOCs (alkanes, alkenes,  aromatic hydrocarbons,
10      carbonyl compounds, etc.) and CO are important for O3 formation. Although knowledge of the
11      oxidative mechanisms of VOCs has improved over the past several years, gaps in knowledge
12      involving key classes, such as aromatic hydrocarbons, still remain. For example, only about half
13      of the carbon initially present in aromatic hydrocarbons in smog chamber studies form
14      compounds that can be identified.
15           In addition to gas phase reactions, reactions also occur on the surfaces of, or within cloud
16      droplets and airborne particles.  Most of the well-established multiphase reactions tend to reduce
17      the rate of O3 formation  in polluted environments. Reactions of Cl and Br containing radicals
18      deplete O3 in selected environments such as the Arctic during spring, the tropical marine
19      boundary layer and inland salt lakes. Direct reactions of O3 with atmospheric particles appear to
20      be too slow to reduce O3 formation significantly at typical ambient PM levels.
21           Our basic understanding of the meteorological processes associated with summertime O3
22      episodes has not changed over the past several years. However, the realization that long-range
23      transport processes are important for determining O3 concentrations at the surface is growing.  In
24      addition to synoptic scale flow fields, nocturnal low-level jets are capable of transporting
25      pollutants hundreds of km from their sources in either the upper boundary layer or the lower free
26      troposphere.  Turbulence then brings O3 and other pollutants to the surface. On larger scales,
27      important progress has been made in identifying the mechanisms of intercontinental transport
28      of O3 and other pollutants.
29           Some O3 would be found near the earth's surface as the result of its downward transport
30      from the stratosphere, even in the absence of photochemical reactions in the troposphere.
31      Intrusions of stratospheric O3 that reach the surface are rare. Much more common are intrusions

        August 2005                              2-23        DRAFT-DO NOT QUOTE OR CITE

-------
 1      that penetrate to the middle and upper troposphere. However, O3 transported to the middle and
 2      upper troposphere can still affect surface concentrations through various mechanisms that mix
 3      air between the planetary boundary layer and the free troposphere above.
 4           The formation of O3 and associated compounds is a complex, nonlinear function of many
 5      factors including the intensity and spectral distribution of sunlight; atmospheric mixing and other
 6      atmospheric processes; and the concentrations of the precursors in ambient air. At the
 7      lower NOX concentrations found in most environments, ranging from remote continental areas to
 8      rural  and suburban areas downwind of urban centers, the net production of O3 increases with
 9      increasing NOX. At the higher concentrations found in downtown metropolitan areas, especially
10      near busy streets and highways, and in power plant plumes there is net destruction of O3 by
11      reaction with NO.  In between these two regimes there is a transition stage, in which O3
12      production shows only a weak dependence on NOX concentrations. The efficiency of O3
13      production per NOX oxidized is generally highest in areas where NOX concentrations are lowest
14      and decrease with increasing NOX concentration.
15           Chemistry transport models are used to improve understanding of atmospheric chemical
16      and physical processes as well as to develop air pollution control strategies. The performance of
17      these models must be evaluated by comparison with field data as part of a cycle of model
18      evaluations and subsequent improvements.  Discrepancies between model predictions and
19      observations  can be used to point out gaps in current understanding and thus to improve
20      parameterizations of atmospheric chemical and physical processes. Model evaluation does not
21      merely involve a straightforward comparison between model predictions and the concentration
22      fields of a pollutant of interest (e.g., O3). Such comparisons may not be meaningful because it is
23      difficult to determine if agreement between measurements and model predictions truly represents
24      an accurate treatment of physical and chemical processes in the model or the effects of
25      compensating errors in model routines.
26           The main methods in use for routine monitoring of ambient O3 are based on
27      chemiluminescence or UV absorption.  Measurements at most ambient monitoring sites are
28      based on UV absorption.  Both of these  methods are subject to interference by other atmospheric
29      components.  One study found large positive interference in Mexico City and in a smog
30      chamber, but few studies conducted in urban plumes did not find significant positive interference
31      in the UV absorption technique.

        August 2005                               2-24        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES
  2
  3       Berkowitz, C. M; Shaw, W. J. (1997) Airborne measurements of boundary layer chemistry during the Southern
  4             Oxidant Study: a case study. J. Geophys. Res. [Atmos.] 102: 12,795-12,804.
  5       Berkowitz, C. M.; Fast, J. D.;  Sprinston, S. R.; Larsen, R. J.; Spicer, C. W.; Doskey, P. V.; Hubbe, J. M.; Plastridge,
  6             R. (1998) Formation mechanisms and chemical characteristics of elevated photochemical layers over the
  7             northeast United States. J. Geophys. Res. [Atmos.] 103: 10,631-10,647.
  8       Blumenthal, D. L.; Lurmann, F. W.; Kumar, N.; Dye, T. S.; Ray, S. E.; Korc, M. E.; Londergan, R.; Moore, G.
  9             (1997) Transport and mixing phenomena related to ozone exceedances in the northeast U.S. (analysis based
10             on NARSTO-northeast data). Available: http://capita.wustl.edu/otag/reports/otagrept/otagrept.html (30
11             October 2003).
12       Bonner, W. D. (1968) Climatology of the low level jet. Mon. Weather Rev. 96: 833-850.
13       Carter, W. P. L. (1995) Computer modeling of environmental chamber studies of maximum incremental reactivities
14             of volatile organic compounds. Atmos. Environ. 29: 2513.
15       Civerolo, K. L.; Mao, H. T.; Rao, S. T. (2003) The airshed for ozone and fine paniculate pollution in the eastern
16             United States. Pure Appl. Geophys. 160: 81-105.
17       Corsmeier, U.; Kalthhoff, N.; Kolle, O.; Motzian, M.; Fiedler, F. (1997) Ozone concentration jump in the stable
18             nocturnal boundary layer during a LLJ-event. Atmos. Environ. 31: 1977-1989.
19       Eisele, F. L.; Mount, G.  H.; Tanner, D.; Jefferson, A.; Shelter, R.; Harder, J. W.; Williams, E. J. (1997)
20             Understanding the production and interconversion of the hydroxyl radical during the tropospheric OH
21             photochemistry experiment. J. Geophys. Res. 102: 6457-6465.
22       Fehsenfeld, F. C.; Trainer, M.; Parrish, D. D.; Volz-Thomas, A.; Penkett, S. (1996) North Atlantic Regional
23             Experiment (NARE) 1993 summer intensive: foreword. J. Geophys. Res. [Atmos.] 101: 28,869-28,875.
24       Fried, A.; Hodgeson, J. (1982) Laser photoacoustic detection of nitrogen dioxide in the gas-phase titration of nitric
25             oxide with ozone. Anal. Chem. 54: 278-282.
26       Guenther, A.; Geron, C.; Pierce, T.; Lamb, B.; Harley, P.; Fall, R. (2000) Natural emissions of non-methane volatile
27             organic compounds, carbon monoxide, and oxides of nitrogen from North America. Atmos. Environ. 34:
28             2205-2230.
29       Hameed, S.; Pinto, J. P.; Stewart, R. W. (1979) Sensitivity of the predicted CO-OH-CH4 perturbation to tropospheric
30             NO-j concentrations. J. Geophys. Res. C: Oceans Atmos. 84: 763-768.
31       Husar, R. B.; Renard, W. P. (1998) Ozone as a function of local wind speed and direction: Evidence of local and
32             regional transport. Presented at: 91st annual meeting and exhibition of the Air & Waste Management
33             Association; June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association; online paper no.
34             98-A922. Available:  http://capita.wustl.edu/capita/CapitaReports/REPORTSl.HTML (13 November 2003).
35       Jacob, D. J.; Horowitz, L. W.; Munger, J. W.; Heikes, B. G.; Dickerson, R. R.; Artz, R. S.; Keene, W. C. (1995)
3 6             Seasonal transition from NOX- to hydrocarbon-limited conditions for ozone production over the eastern
37             United States in September. J. Geophys. Res. 100: 9315-9324.
38       Jaegle, L.; Jacob, D. J.; Brune, W.  H.; Wennberg, P. O. (2001) Chemistry of HOx radicals in the upper troposphere.
39             Atmos. Environ. 35: 469-489.
40       Jobson, B. T.; Berkowitz, C. M.; Kuster, W. C.; Goldan, P. D.; Williams, E. J.; Fehsenfeld, F. C.; Apel, E. C.; Karl,
41             T.; Lonneman, W. A.; Riemer, D. (2004) Hydrocarbon source signatures in Houston, Texas: influence of the
42             petrochemical industry. J. Geophys. Res. 109: D24305: 10.1029/2004JD004887.
43       Kasibhatla, P.; Chameides, W. L. (2000) Seasonal modeling of regional ozone pollution in the eastern United States.
44             Geophys. Res. Lett. 27: 1415-1418.
45       Kleindienst, T. E.; Hudgens, E. E.; Smith, D. F.; McElroy, F. F.; Bufalini, J. J. (1993)  Comparison of
46             chemiluminescence and ultraviolet ozone monitor responses in the presence of humidity  and photochemical
47             pollutants. Air Waste 43: 213-222.
48       Leston, A.; et al. (2005) J. Air Waste Manage. Assoc.: in press.
49       Milford, J. B.; Gao, D.; Sillman, S.; Blossey, P.; Russell, A. G. (1994) Total reactive nitrogen (NOy) as an indicator
50             of the  sensitivity of ozone to reductions in hydrocarbon and NOX emissions.  J. Geophys.  Res. 99:  3533-3542.
51       National Research Council. (1991) Rethinking the ozone problem in urban and regional air pollution. Washington,
52             DC: National Academy Press. Available: http://www.nap.edu/books/0309046319/html/ [26 March, 2004].
53       Olszyna, K. J.; Bailey, E. M.;  Simonaitis, R.; Meagher, J. F. (1994) O3 and NOy relationships at a rural site. J.
54             Geophys. Res.  [Atmos.] 99: 14,557-14,563.
55       Parrish, D. D.; Fehsenfeld, F. C. (2000) Methods for gas-phase measurements  of ozone, ozone precursors and
56             aerosol precursors. Atmos. Environ. 34: 1921-1957.


         August 2005                                    2-25          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Pierson, W. R.; Gertler, A. W.; Bradow, R. L. (1990) Comparison of the SCAQS tunnel study with historical data.
  2             Presented at: 83rd annual meeting & exhibition of the Air and Waste Management Association; June;
  3             Pittsburgh, PA. Pittsburgh, PA: Air and Waste Management Association; paper no. 90-175.3.
  4       Piety, C. A. (2005) The relation between daily maximum ozone and  daily maximum temperature [memorandum to
  5             Dr. Joseph Pinto]. Research Triangle Park, NC: U.S. Environmental Protection Agency; July 18.
  6       Pinto, J. P.; Bruhl, C.; Thompson, A. M. (1993) The current and future envirionmental role of atmospheric methane.
  7             In: Khalil, M. A. K., ed. Atmospheric methane sources, sinks, and role in global change, p. 514-531. (NATO
  8             ASI Series, v. 113).
  9       Poppe, D.; Wallasch, M.;  Zimmermann, J. (1993) The dependence of the concentration of OH on its precursors
10             under moderately polluted conditions:  a model study. J. Atmos. Chem. 16: 61-78.
11       Rao, S. T.; Ku, J.-Y.; Herman, S.; Zhang, K.; Mao, H. (2003) Summertime characteristics of the atmospheric
12             boundary layer and relationships to ozone levels over the eastern United States. Pure Appl. Geophys. 160:
13             21-55.
14       Ryerson, T. B.; Buhr, M. P.; Frost, G. J.; Goldan, P. D.; Holloway, J. S.; Htibler, G.; Jobson, B. T.; Kuster, W. C.;
15             McKeen, S. A.; Parrish, D. D.; Roberts, J. M.; Sueper, D. T.;  Trainer, M.; Williams, J.; Fehsenfeld, F. C.
16             (1998) Emissions lifetimes and ozone formation in power plant plumes. J. Geophys. Res. 103(D17):
17             22,569-22,583.
18       Ryerson, T. B.; Trainer, M.; Holloway, J. S.;  Parrish, D. D.; Huey, L. G.; Sueper, D. T.; Frost, G. J.; Donnelly, S.
19             G.; Schauffler, S.; Atlas, E. L.; Kuster, W. C.; Goldan, P. D.;  Htibler, G.; Meagher, J. F.; Fehsenfeld, F. C.
20             (2001) Observations of ozone formation in power plant plumes and implications for ozone control strategies.
21             Science (Washington, DC) 292: 719-723.
22       Sakugawa, H.; Kaplan, I. R. (1989) H2O2 and O3 in the atmosphere of Los Angeles and its vicinity: factors
23             controlling their formation and their role as oxidants of SO2. J. Geophys. Res. [Atmos.] 94:  12,957-12,973.
24       Schichtel, B. A.; Husar, R. B. (2001) Eastern North American transport climatology during high- and low-ozone
25             days. Atmos. Environ. 35: 1029-1038.
26       Seinfeld, J. H.; Pandis, S.  N. (1998)  Atmospheric chemistry and physics: from air pollution to climate change. New
27             York, NY: John Wiley & Sons, Inc.
28       Shapiro, M. A. (1980) Turbulent mixing within tropopause folds as a mechanism for the exchange of chemical
29             constituents between the stratosphere and troposphere. J. Atmos. Sci. 37: 994-1004.
30       Sillman, S. (1995) The use of NOy, H2O2 and HNO3 as indicators for ozone-NOx-hydrocarbon sensitivity in urban
31             locations. J. Geophys. Res. 100:  14,175-14,188.
32       Sillman, S.; He, D.-Y. (2002) Some  theoretical results concerning O3-NOX-VOC chemistry and NOX-VOC
33             indicators. J. Geophys. Res. (Atmos.) 107: 10.1029/2001JDOO1123.
34       Sillman, S.; He, D.; Cardelino, C.; Imhoff, R. E. (1997) The use of photochemical indicators to evaluate
3 5             ozone-NOx-hydrocarbon sensitivity: case studies from Atlanta, New York, and Los Angeles. J. Air Waste
36             Manage. Assoc. 47: 1030-1040.
37       Sillman, S.; He, D.; Pippin, M. R.; Daum, P. H.; Imre, D. G.;  Kleinman, L. I.; Lee, J. H.; Weinstein-Lloyd, J. (1998)
3 8             Model correlations for ozone, reactive nitrogen, and peroxides for Nashville in comparison with
39             measurements: implications for O3-NOx-hydrocarbon chemistry. J. Geophys. Res. [Atmos.] 103:
40             22,629-22,644.
41       Stedman, D. H.; Bishop, G.; Peterson, J. E.; Guenther, P. L. (1991) On-road CO remote sensing in the Los Angeles
42             Basin: final report. Sacramento, CA: California Air Resources Board, ARE Contract No. A932-189.
43       Stall, R. B. (1999) An introduction to boundary layer meteorology. Dordrecht, The Netherlands: Kluwer Academic
44             Publishers; pp. 9-15, 500-505.
45       Stall, R. B. (2000) Meteorology for  scientists and engineers: a technical companion book with Ahrens' Meteorology
46             Today. 2nd ed. Pacific Grove, CA: Brooks/Cole.
47       Taubman, B. F.; Marufu, L. T.; Piety, C. A.; Doddridge, E.G.; Stehr, J. W.; Dickerson, R. R. (2004) Airborne
48             characterization of the chemical, optical, and meteorological properties, and origins of a combined
49             ozone-haze episode over the eastern United States. J. Atmos.  Sci. 61: 1781-1793.
50       Trainer, M.; Parrish, D. D.; Buhr, M. P.; Norton, R. B.; Fehsenfeld, F. C.;  Anlauf, K. G.; Bottenheim, J. W.; Tang,
51             Y. Z.; Wiebe, H. A.; Roberts, J. M.; Tanner, R. L.; Newman, L.; Bowersox, V.  C.; Meagher, J. F.; Olszyna,
52             K. J.; Rodgers, M.  O.; Wang, T.; Berresheim, H.; Demerjian,  K. L.; Roychowdhury, U. K. (1993) Correlation
53             of ozone with NOy in photochemically aged air. J. Geophys. Res. [Atmos.] 98: 2917-2925.
54       U.S. Environmental Protection Agency. (2000) Air quality criteria for carbon monoxide. Research Triangle Park,
55             NC: National Center for Environmental Assessment; report no. EPA/600/P-99/00 IF. Available:
56             http://www.epa.gov/ncea/pdfs/coaqcd.pdf (7 May 2003).
         August 2005                                   2-26         DRAFT-DO NOT QUOTE OR CITE

-------
1      Volz-Thomas, A.; Geiss, H.; Hofzumahaus, A.; Becker, K.-H. (2003) Introduction to special section: photochemistry
2            experiment in BERLIOZ. J. Geophys. Res. [Atmos.] 108(D4): 10.1029/JD002029.
3      Zimmermann, J.; Poppe, D. (1993) Nonlinear chemical couplings in the tropospheric NOX-HOX gas phase chemistry.
4            J. Atmos. Chem. 17: 141-155.
5
       August 2005                                  2-27         DRAFT-DO NOT QUOTE OR CITE

-------
 i              3.  ENVIRONMENTAL CONCENTRATIONS,
 2               PATTERNS, AND EXPOSURE ESTIMATES
 3
 4
 5     3.1   INTRODUCTION
 6     Identification and Use of Existing Air Quality Data
 1          Topics discussed in this chapter include the characterization of ambient air quality data for
 8     ozone (O3), the uses of these data in assessing the exposure of vegetation to O3, concentrations
 9     of O3 in microenvironments, and a discussion of the currently available human exposure data and
10     exposure model development. The information contained in this chapter pertaining to ambient
11     concentrations is taken primarily from the U.S. Environmental Protection Agency (EPA) Air
12     Quality System (AQS; formerly the  AIRS database).  The AQS contains readily accessible
13     detailed, hourly data that has been subject to EPA quality control and assurance procedures.
14     Data available in AQS were collected from 1979 to 2001.  As discussed in previous versions of
15     the O3 Air Quality Criteria Document or AQCD (U.S. Environmental Protection Agency,  1986,
16     1996), the data available prior to 1979 may be unreliable due to calibration problems and
17     uncertainties.
18          As noted in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996), O3 is the
19     only photochemical oxidant other than nitrogen dioxide (NO2) that is routinely monitored and
20     for which a comprehensive database exists. Data for peroxyacetyl nitrate (PAN), hydrogen
21     peroxide  (H2O2), and other oxidants either in the gas phase or particle phase typically have been
22     obtained only as part of special field studies.  Consequently, no data on nationwide patterns  of
23     occurrence are available for these non-O3 oxidants; nor are extensive data available on the
24     relationships of levels and patterns of these oxidants to those of O3.  However, available data for
25     gas phase and particle phase oxidants will be discussed.
26
27     Characterizing Ambient Ozone Concentrations
28          The "concentration" of a specific air pollutant is typically defined as the amount (mass) of
29     that material per unit volume of air.  However, most of the data presented in this chapter are
30     expressed as "mixing ratios" in terms of a volume-to-volume ratio (parts per million [ppm] or
31     parts per billion [ppb]). Data expressed this way are often referred to as concentrations, both in

       August 2005                              3-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1      the literature and in the text, following common usage. Human exposures are expressed in units
 2      of mixing ratio times time.
 3           Several different types of indicators are used for evaluating exposures of vegetation to O3.
 4      The peak-weighted, cumulative exposure indicators used in this chapter for characterizing
 5      vegetation exposures are SUM06 and SUM08 (the sums of all hourly average concentrations
 6      >0.06 and  0.08 ppm, respectively) and W126 (the sum of the hourly average concentrations that
 7      have been  weighted according to a sigmoid function that is based on a hypothetical vegetation
 8      response [see Lefohn and Runeckles, 1987]). Further discussion of these exposure indices is
 9      presented in Chapter 9.
10           The EPA has established "ozone seasons" during which measurement of ambient O3
11      concentrations for different locations within the United States and the U.S. territories is required
12      (CFR, 2000).  Table AX3-1 shows the O3 seasons during which continuous, hourly averaged O3
13      concentrations must be monitored.  Monitoring is optional outside of these O3 seasons and
14      indeed is conducted during the winter in a number of areas.
15           Data for O3 in ambient air across the United States are  summarized in Section 3.2. The
16      data are summarized for urban, rural, and relatively remote sites. Relatively remote monitoring
17      sites (RRMS) are sites that are not strongly influenced by nearby pollution sources and are
18      located mainly in national parks  in the West. However, this  does not mean that they are free of
19      the effects of regional or local pollution, especially during tourist seasons.  Data for the spatial
20      variability of O3 within urban areas are summarized in Section 3.3. Data for the diurnal and
21      seasonal variability of O3 concentrations are given in Section 3.4.  The long term temporal
22      variability of O3 concentrations is discussed in Section 3.5. Relationships among O3 and other
23      species are discussed in Section 3.6. Information about the occurrence of other oxidants and
24      their relationship to O3 is given in this section.  A discussion of Policy Relevant Background
25      (PRB) O3 concentrations is presented in Section 3.7. PRB O3 concentrations are background O3
26      concentrations used for the purposes of setting the O3 NAAQS. They are used by the EPA to
27      assess risks to human health.  Indoor sources and emissions of O3 are discussed in Section 3.8.
28      Issues related to evaluating human exposure to O3 are summarized in Section 3.9.  Finally, a
29      summary of key points in Chapter 3 is given in Section 3.10.
30
        August 2005                               3-2        DRAFT-DO NOT QUOTE OR CITE

-------
 1     3.2  AMBIENT AIR QUALITY DATA FOR OZONE
 2     Ozone Air Quality at Urban, Suburban, andNonurban Sites
 3          Figure 3-1 shows the mean daily maximum 8-h O3 concentrations and Figure 3-2 shows
 4     the 95th percentile values of the daily maximum 8-h O3 concentrations, based on county wide
 5     averages across the United States from May to September 2000 to 2004. The period from May
 6     to September was chosen because, although O3 is monitored for different lengths of time across
 7     the country, all O3 monitors should be operational during these months.  Data flagged because of
 8     quality control issues were removed with concurrence by the local monitoring agency. Only
 9     days with data for 18 of 24 hours were kept, and a minimum of 115 of 153 days were required in
10     each year. Cut points for the tertile distributions on each map were chosen at the median and
11     95th percentile values. These cut points were chosen as they represent standard metrics for
12     characterizing important aspects of human exposure used by the EPA. Any other percentiles or
13     statistics that are believed to be helpful for characterizing human exposures could also be used.
14     Blank areas on the maps indicate no data coverage. It should be noted that county areas can be
15     much larger in the West than in the East, but monitors are not spread evenly within a county. As
16     a result, the assigned concentration range might not represent conditions throughout a particular
17     county and so large areas in western counties where there are no monitors were blanked out.
18          As shown in Figure 3-1, the median of the county wide, mean daily maximum  8-h O3
19     concentration across the United States is 49  ppb, and the corresponding 95th percentile value is
20     57 ppb. Though the median and 95th percentile values are fairly close, these results cannot be
21     taken to imply that average O3 concentrations lie in a relatively narrow range throughout the
22     United States, because data coverage is not as complete in the West as it is in the East. High
23     mean daily maximum 8-h O3 concentrations are found in California and  states in the Southwest
24     as well as in several counties in the East. As shown in Figure 3-2, the nationwide median of the
25     countywide, 95th  percentile value of the daily maximum 8-h O3 concentration is 73  ppb and 5%
26     of these values are above 85 ppb.  High values for the 95th percentiles are found in California,
27     Texas, and some counties in the East, but not necessarily in the same counties in the East as
28     shown for the mean daily maximum 8-h concentrations in Figure 3-1.
29          Although mean O3  concentrations in Houston, TX were below the nationwide  median, its
30     95th percentile value ranks in the highest 5% nationwide. Conversely, mean O3 concentrations
31     in southwestern states are among the highest in the United States, but values at the upper end of

       August 2005                              3-3         DRAFT-DO NOT QUOTE OR CITE

-------
                  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).
 1     the distribution (e.g., the 95th percentile value) in these states are not among the highest peak
 2     values in the United States.  In other areas where the highest mean O3 concentrations occurred,
 3     such as California; Dallas-Fort Worth, TX; and the Northeast Corridor, the highest peak values
 4     were also observed.
 5          Although countywide averages are shown, it should be noted that considerable spatial
 6     variability can exist within a county, especially within urban areas as described in Section 3.3.
 7     In addition, there can also be differences in the diurnal profile of O3 among monitors within
 8     counties.
 9          Box plots showing the percentile distribution of nationwide O3 concentrations for different
10     averaging periods (1-h daily maximum, 8-h daily maximum and 24-h daily average) are given in
       August 2005
5-4
DRAFT-DO NOT QUOTE OR CITE

-------
              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).
 1     Figures AX3-4 to AX3-6 and numerical values are given in Table AX3-2. The differences
 2     between the 50th and 95th percentile values can be used to provide indications of differences
 3     in O3 levels between "typical" O3 days and "high" O3 days. These differences are approximately
 4     40, 30, and 25 ppb for the daily 1-h and 8-h daily maxima and 24-h average O3 concentrations.
 5     As might be expected, the daily maximum 1-h and 8-h O3 concentrations are highly correlated.
 6           Lehman et al. (2004) have shown that the eastern United States can be divided into five
 7     regions,  each of which exhibit spatial, relatively coherent patterns of O3 properties at nonurban
 8     sites.  Only sites classified as being rural or suburban and with land usage of forest, agriculture,
 9     or residential were included in the analyses.  These criteria were chosen to avoid sites where O3
10     is scavenged by NO that can be found in high concentrations near major sources, such as traffic
       August 2005
5-5
DRAFT-DO NOT QUOTE OR CITE

-------
 1      in urban cores. The five regions, shown in Figure 3-3, are characterized by different patterns
 2      of O3 properties such as temporal persistence and seasonal variability. Figure 3-3 shows
 3      nonurban, monthly average, daily maximum 8-h O3 concentrations in the five regions in the
 4      eastern United States from April to October 1993 to 2002.
 5           Regional differences are immediately apparent.  Highest concentrations among all the
 6      regions are generally found in the Mid-Atlantic region (mean of 52 ppb) with highest values
 7      throughout the percentile distribution except for the overall maximum.  Lowest mean
 8      concentrations (42 ppb) are found in Florida.  In the northern regions (the Northeast, Great
 9      Lakes) and the Mid-Atlantic region, highest median and peak concentrations are found in July,
10      whereas in the Southwest region, highest median concentrations are found in August, with
11      highest peaks in June and September, i.e., outside the warmest summer months.  In Florida,
12      highest monthly averaged median and peak concentrations are found during the spring.  High O3
13      concentrations tend to be most persistent (3-4 days of persistence) in the southern regions, less
14      persistent in the Mid-Atlantic region (2-3 days) and least persistent in the northern regions  (1 or
15      2 days). Analyses, such as these, are not available for the western United States, in part because
16      of the difficulty in defining regions with relatively coherent O3 properties.
17           Box plots showing the percentile distribution of hourly average O3 concentrations for
18      different types of rural sites for 2004 are given in Figures 3-4a (rural-agricultural), 3-4b
19      (rural-forest) and 3-4c (rural-residential or commercial).  Some associated metrics for vegetation
20      exposures are given in Figures AX3-8 to AX3-10. Note that high O3 concentrations are found at
21      sites that are classified as rural, such as Anne Arundel Co., MD; Yosemite NP, CA; and
22      Crestline, CA. Land use designations do not usually give an accurate picture of exposure
23      regimes in rural areas, because the land use characterization of "rural" does not imply that a
24      specific location is isolated from anthropogenic influences. Rather, the characterization refers
25      only to the current use of the land, not to the presence of sources.  Since O3 produced from
26      emissions in urban areas is transported to more rural downwind locations, elevated O3
27      concentrations can occur at considerable distances from urban centers.  In addition, major
28      sources of O3 precursors such  as power plants and highways are located in nonurban areas and
29      also produce O3 in these areas. Due to lower chemical scavenging in nonurban areas, O3 tends to
30      persist longer in nonurban than in urban areas also tending to lead to higher exposures in
31      nonurban areas influenced by anthropogenic precursor emissions.

        August 2005                                3-6        DRAFT-DO NOT QUOTE OR CITE

-------
 "
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).
August 2005
                                             5-7
         DRAFT-DO NOT QUOTE OR CITE

-------
         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-
                                          Rural Agriculture
                     3618
                               3652
                                        3643
                                                 3493
                                                           3621
i
                                                                    3519
                                                                              3486
T
                   Effingham,   Indianapolis,  AnneArundel  Omaha. NE
                     IL         IN      Co., MD
                                            Rural Forest
                                             Beaver
                                             Co., PA
         Fauquier     Dodge
         Co., VA     Co., Wl
                     3420
                                        2676
                                                           3419
                                                                              3312
                 Yosernite NP, CA
                                   Whiteface Mtn., NY
                                         Smoky Mtns. NP, TN
               Shenandoah NP, VA
                         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 k

                  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).
August 2005
                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      Ozone Air Quality Data at Relatively Remote Monitoring Sites (RRMS)
 2           RRMS are sites that are located in the national parks that tend to be less affected by
 3      obvious pollution sources than other sites.  This does not mean that they are completely
 4      unaffected by local pollution, as evidenced by the number of visitors to these national parks.
 5           Box plots showing the percentile distribution of annual hourly averaged O3 concentrations
 6      at four relatively remote monitoring sites (RRMS) are given in Figures 3-5a-d.  It is important to
 7      characterize hourly average O3 concentrations at RRMS so that assessments of the possible
 8      effects of O3 on human health and vegetation use ranges of concentrations in their experiments
 9      that span the range of O3 concentrations found in the U.S.  In many controlled exposure studies
10      examining vegetation, O3 is filtered out of ambient air before it is admitted into the exposure
11      chambers.  As a result, O3 levels of only a few ppb are used as controls.
12           As can be seen from Figures 3-5a-d, annual mean values of the daily maximum 8-h O3
13      concentration have not changed much over the past 10 years of available data. Mean values
14      typically range from about 0.020 ppm to about 0.040 ppm.  Concentrations only rarely exceed
15      0.080 ppm, in contrast to observations at other "rural" sites shown in Figures 3-4a-c.
16           The extent to which distributions found at sites with low maximum hourly average
17      concentrations in the western United States are representative of sites in the eastern and
18      midwestern United States is debatable because of regional differences in sources of precursors
19      and transport patterns.  Given the high density of sources in the eastern and midwestern
20      United States, it is unclear whether a site could be found in either of these regions that would not
21      be influenced by the transport of O3 from nearby urban areas. Thus, with the exception of the
22      Voyageurs NP site in Minnesota, observations at RRMS are limited to those obtained in the
23      western United States.  However, not all national park sites in the West can be considered to
24      be free of strong regional pollution influences, e.g.,  Yosemite NP (CA) as shown in Figure 3-4b.
25      Maps showing the nationwide distribution of various metrics for vegetation exposures are given
26      in Section AX3.2,  Figures AX3-13 to AX3-27.
27
        August 2005                               3-9        DRAFT-DO NOT QUOTE OR CITE

-------
            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).
August 2005
3-10
DRAFT-DO NOT QUOTE OR CITE

-------
 1      3.3  SPATIAL VARIABILITY OF OZONE IN URBAN AREAS
 2           The spatial variability in O3 concentrations in 24 MSAs across the United States was
 3      examined. These MSAs were selected to provide (1) information helpful for risk assessments,
 4      (2) a general overview of the spatial variability of O3 in different regions of the country, and
 5      (3) insight into the spatial  distribution of O3 in cities where health outcome studies have been
 6      conducted.  Statistical analyses of the human health effects of airborne pollutants based on
 7      aggregate population time-series data have often relied on ambient concentrations of pollutants
 8      measured at one or more central sites in a given metropolitan area.  In the particular case of
 9      ground-level O3 pollution, central-site monitoring has been justified as a regional measure of
10      exposure mainly on the grounds that correlations between concentrations at neighboring sites
11      measured over time are usually high. In MSAs with multiple monitoring sites, averages over the
12      monitors have often been used to characterize population exposures.  However, substantial
13      differences in concentrations between monitors can exist even though concentrations measured
14      at the monitoring sites are highly correlated, thus leading to the potential for exposure
15      misclassification error.
16           Metrics for characterizing spatial variability include the use of Pearson correlation
17      coefficients (r), values of the 90th percentile absolute difference in O3 concentrations (P90), and
18      coefficients of divergence (COD)1.  These methods of analysis follow those used for
19      characterizing PM25 and PM10_25 concentrations in Pinto et al. (2004) and in the latest edition of
20      the Paniculate Matter (PM) AQCD (U.S. Environmental Agency, 2004a). However, the
21      calculations were performed on an hourly basis rather than on a 24-h basis. Data were
22      aggregated over the local O3 season as indicated in Table AX3-1.  The length of the O3 season
23      varies across the country.  In several southwestern states, it lasts all year long. In other areas,
24      such as in New England, the mid-Atlantic states, the Midwest and the Northwest, it can be
25      6 months long, but typically it lasts from April through October.
        1 The COD is defined as follows:
       1   P
k=   -V
                                 CODjk=   -        --                            (AX3-1)

        where xu and xik represent the 24-h average PM2 5 concentration for day /' at site y and site k andp is the number of
        observations.

        August 2005                               3-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Table 3-1 shows the urban areas chosen, the range of 24-h average O3 concentrations over
 2      the O3 season, the range of intersite correlation coefficients, the range of P90 differences in O3
 3      concentrations between site pairs, and the range in COD values. A COD of zero implies that
 4      values in both data sets are identical, and a COD of one indicates that two data sets are
 5      completely different. In general, statistics were calculated for partial MSAs.  This was done so
 6      as to obtain reasonable lower estimates of the spatial variability that is present, as opposed to
 7      examining the consolidated MSAs.  However, this could not be readily done for Boston, MA
 8      and New York, NY, so statistics were calculated for those consolidated MSAs. More detailed
 9      calculations for a subset of nine MSAs are given in Figures AX3-28 through AX3-36 in
10      Section AX3.3.
11           As can be seen from Table 3-1, no clearly discernible regional differences were found
12      in the ranges of parameters analyzed.  Additional urban areas would need to be examined to
13      discern broadscale patterns.  The data indicate considerable variability in the concentration
14      fields. Mean O3 concentrations vary within individual urban areas by factors of 1.4 to 4.
15      Intersite correlation  coefficients show mixed patterns (i.e., in some urban areas all pairs of sites
16      are moderately to highly correlated, while other areas show a larger range of correlations).
17      As may be expected, those areas showing a smaller range of seasonal  mean concentrations also
18      show a smaller range of intersite correlation coefficients.  However, there are a number of cases
19      where sites in an urban area may be moderately to highly  correlated, but show substantial
20      differences in absolute concentrations. In many cases, P90 values can  equal or exceed seasonal
21      mean O3 concentrations.
22           It is instructive to compare the metrics for spatial variability shown in Table 3-1 to those
23      calculated for PM25  and PM10_25 in the PM AQCD (U.S. Environmental Agency, 2004).  The
24      values for concentrations and concentration differences are unique to the  individual species, but
25      the intersite correlation coefficients and the COD values can be directly compared. In general,
26      the variability in O3  concentrations is larger than for PM2 5 concentrations and comparable to that
27      obtained for PM10_2 5. Intersite correlation coefficients in some areas (e.g., Philadelphia, PA;
28      Atlanta, GA; Portland, OR) can be very similar for both PM2 5 and for O3. However, there is
29      much greater variability in the concentration fields of O3 as evidenced by the much higher COD
30      values. Indeed, COD values are higher for O3 than for PM2 5 in each of the urban areas
31      examined.  In all of the urban areas examined for O3, some site pairs are always very highly

        August 2005                               3-12       DRAFT-DO NOT QUOTE OR CITE

-------
    Table 3-1. Summary Statistics for the Spatial Variability of O3 (in ppm) in Selected Urban Areas in the United States
OQ
to
O
o









OO
OJ

O
^
?
O
0
o
H
0
0
H
W
0
O
H
W
Urban Area
Boston, MA
New York, NY
Philadelphia, PA
Washington, DC
Charlotte, NC
Atlanta, GA
Tampa, FL
Detroit, MI
Chicago, IL
Milwaukee, WI
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.

-------
 1      correlated with each other (i.e., r >0.9) as seen for PM2 5. These sites also show less variability
 2      in concentration and are probably influenced most strongly by regional production mechanisms.
 3           The above considerations indicate that caution should be observed in using data from the
 4      network of ambient O3 monitors to approximate community-scale human exposures.  A similar
 5      conclusion was reached for PM using data from the PM2 5 FRM network, as indicated in
 6      Section 3.4 of the PM AQCD (U.S. Environmental Protection Agency, 2004a).
 7
 8      3.3.2  Small-scale Horizontal and Spatial Variability in Ozone Concentrations
 9      Ozone concentrations near roadways
10           Apart from the larger scale variability in surface O3 concentrations, there is also significant
11      variability on the micro-scale (< a few hundred meters), especially near roadways and other
12      sources of emissions that react with O3. These sources are not confined to urban areas.  Sources
13      of emissions that react with O3 such as highways and power plants are also found in rural areas.
14      Johnson (1995) described the results of studies examining O3 upwind and downwind of
15      roadways in Cincinnati, OH.  In these studies, O3 upwind of the roadway was about 50 ppb and
16      these values were not found again until distances of about 100 m downwind.  The O3 profile
17      varied inversely with that of NO, as might be expected. For peak NO concentrations of 30 ppb
18      immediately downwind of the road, the O3 mixing ratio was about 36 ppb, or about 70% of the
19      upwind value. The magnitude of the downwind depletion of O3 depends on the emissions of
20      NO, the rate of mixing of NO from the roadway and ambient temperature and so depletions
21      of O3 downwind of roadways are expected, but with variable magnitude. Guidance for the
22      placement of O3 monitors (U.S. Environmental Protection Agency, 1998) states a separation
23      distance that depends on traffic counts.  For example, a minimum separation distance of 100 m
24      from a road with 70,000 vehicles per day (about 3,000 vehicles per hour) is recommended for
25      siting an O3 monitor to avoid interference that would mean a site is no longer representative of
26      the surrounding area. An average rate of about 3,000 vehicles per hour passing by a monitoring
27      site implies a road with rather heavy traffic.  As noted in Section AX3.3.1 for the Lakewood, CA
28      monitoring, O3 levels are lower at sites located near traffic than those located some distance
29      away and the scavenging of O3 by emissions of NO from roadways is a major source of spatial
30      variability in O3 concentrations.  It should also be noted that scavenging of O3 by NO near
31      roadways was more pronounced before the implementation of stringent NOx emissions controls.

        August 2005                              3-14       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Vertical Variations in Ozone Concentrations
 2           In addition to horizontal variability in O3 concentrations, there are also variations in the
 3      vertical profile of O3 in the lowest layers of the atmosphere to consider. The planetary boundary
 4      layer consists of an outer and an inner portion. The inner part of the planetary boundary layer
 5      extends from the surface to about one-tenth the height of the planetary boundary layer. Winds
 6      and transported properties, such as O3, are especially susceptible to interactions with obstacles,
 7      such as buildings and trees in the inner boundary layer (atmospheric surface layer) (e.g., Garratt,
 8      1992).  Inlets to ambient monitors (typically at heights of 3 to 5  meters) are located in, and
 9      human and vegetation exposures occur in this part of the boundary layer.
10           Photochemical production and destruction of O3 occur throughout the planetary boundary
11      layer.  However, O3 is also destroyed on the surfaces of buildings, vegetation, etc.  On most
12      surfaces, O3 is destroyed with every collision. In addition, O3 is scavenged by NO emitted by
13      motor vehicles and soils. These losses imply that the vertical gradient of O3 should always be
14      directed downward.  The magnitude of the gradient is determined by the intensity of turbulent
15      mixing in the  surface layer.
16           Most work characterizing the vertical profile of O3 near the surface has been performed in
17      nonurban areas with the aim of calculating fluxes of O3 and other pollutants through forest
18      canopies and to crops and short vegetation, etc.  Corresponding  data are sparse for urban areas.
19      However, monitoring sites are often set up in open areas such as parks and playgrounds where
20      surface characteristics may be more similar to those in rural areas than to those in the
21      surrounding urban area. The vertical profiles  of O3 measured over low vegetation are  shown in
22      Figure 3-6. These measurements were obtained as part of a field campaign to measure the fluxes
23      of several gas and aerosol phase pollutants using the gradient-flux technique (Horvath et al.,
24      1995).  The labels stable and unstable in the figure refer to atmospheric stability conditions and
25      average represents the overall average.  Ozone concentrations were normalized to their values at
26      4m height. As can be seen from the figure, there was a decrease of about 20% in going from a
27      height of 4 m  down to 0.5 m above the surface during stable conditions, but O3 decreased by
28      only about 7% during unstable conditions. The average decrease was about 10% for all
29      measurements. As might be expected, O3 concentrations at all heights were very highly
30      correlated with one another. Of course, these values represent averages and there is scatter about
31      them.  Under strongly stable conditions, they fall off toward the surface.  However, these

        August 2005                               3-15       DRAFT-DO NOT QUOTE OR CITE

-------
                    4-1
               O)
              "55
              X
 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 unstable
                  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. Horvathetal. (1995).
1      conditions tend to occur mainly during night and the stability regime during the day in urban
2      areas tends more toward instability because of the urban heat island effect. Figure 3-7 shows the
3      vertical profile of O3 measured in a spruce forest by the same group (Horvath et al., 2003). The
4      fall off of O3 in this case is due to uptake by trees, reaction with ambient NO and with NO
5      emitted by the soil in the forest, and reaction with hydrocarbons emitted by the trees in addition
6      to deposition on the surface.
       August 2005
                              3-16
DRAFT-DO NOT QUOTE OR CITE

-------
                      15-
                  D)
                  '5
                      2.5-
                      0.5-
                        0.5
                           + Average (n = 1760)
                           + Stable  (n = 1272)
                           • Unstable (n = 488)
                                   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 unstable and unstable conditions were 36.7 and 33.8 ppb,
                   and the average O3 concentration weighted by stability class was 34.6 ppb at
                   19m.
       Source: Horvath et al. (2003).
 1
 2
 3
 4
 5
 6
 1
 8
 9
10
11
12
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.

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
       August 2005
                                         3-17
DRAFT-DO NOT QUOTE OR CITE

-------
                                                 Urban Sites
                  •=•  0.150
                  Q.
                  C
                  o
                  i  °-100"
                  o
                  C
                  o
                  o
                  0)
                  C
                  o
                  O  0.050
                     0.000 H
                          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).
 1     extrapolating results from one city to another and when attempting to judge the time of day when
 2     the daily 1-h maximum occurs.
 3          Corresponding data for 8 hour average O3 data are shown in Figure 3-9.  As can be seen
 4     from Figure 3-9, daily maximum eight hour O3 concentrations tend to occur from about 10 a.m.
 5     to about 6 p.m. As can be seen from Figure 3-9, they can also occur at slightly different times
 6     and the variation in the 8-h averages is smoother than for the 1-h averages. The minima in the
 7     8 h averages tend to occur starting at about midnight.
 8
 9     Diurnal Patterns in EPA's 12 Cities
10          The diurnal variability of hourly averaged O3 in the twelve urban areas considered for
11     inclusion in EPA's human health exposure assessment  risk assessment for the current review is
12     illustrated in Figures 3-10a-l for April to October.  Daily maximum 1-h concentrations tend to
       August 2005
3-18
DRAFT-DO NOT QUOTE OR CITE

-------
                                                  Urban Sites
                     0.125 -i
                     0.100 -
                  Q.
                  O  0.075
                  c
                  0)
                  O
                  c
                  O  0.050
                  0)
                  c
                  O
                  N
                  O
                     0.025 -
                     0.000 ^
                          22 23 00 01 02 03 04 05 06 07 08
                                                   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 hour 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).
 1     occur in mid-afternoon.  However, as can be seen from the figures, the diurnal patterns vary
 2     from city to city, with high values (>0.100 ppm) also occurring either late in the evening as in
 3     Boston, past midnight as in Los Angeles and Sacramento, or midmorning as in Houston.
 4     Typically, high values such as these are found during the daylight hours in mid to late afternoon.
 5     The reasons for the behavior of O3 during the night at the above-mentioned locations are not
 6     clear. Measurement issues may be involved or there may be physical causes such as transport
 7     phenomena, as discussed in Chapter 2. As discussed in Chapter 2, and in greater detail in
 8     Section AX2.3.3, nocturnal low level jets are  capable of producing secondary O3 maxima at
 9     night.
10          The diurnal variability of O3 averaged over 8 hours in the same twelve urban areas is
11     shown in Figures 3-1 la-1.  The diurnal patterns of O3 are broadly similar between 1-h averages
12     and 8-h averages. A typical pattern shows the 8-h daily maximum occurring from about 10 a.m.
       August 2005
3-19
DRAFT-DO NOT QUOTE OR CITE

-------
        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-W\,
  —  0125
                                                     — 0-125 •
        22 23 00 01 02 03 04 05 060708 09 10 11 12 13 14 15 16 17 18 1920 21 22 230001
                            hour
          2223 00 01 02030405 06 070809 '0 11 12 13 14 15 16 17 18 19 20 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).
August 2005
3-20         DRAFT-DO NOT QUOTE OR CITE

-------
         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).
August 2005
                                         3 -21         DRAFT-DO NOT QUOTE OR CITE

-------
        a. Boston-Worcester-Manchester, MA-NH
          b. New York-Newark-Bridgeport, NY-NJ-CT-PA
                                                     E  0.100-
                                                     o
                                                    '•C  0.075
        22 23 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 '7 18 19 20 21 22 23 00 01

                            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
         d. Washington-Baltimore-Northern Virginia, DC-MD-VA-WV
  8  0,025-
                                                    E  0.100-
                                                    o
                                                    *=  0.075

                                                                                              TTT
        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
          97 93 0001 0? 03040506070809 10 11 1? 13 14 15 16 1718 19 ?0 ?1 79 ?3 00 01

                             hour
        e. Atlanta-Sandy Springs-Gainesville, GA-AL
  E  0.100-
  O
  *=  0.075 •
          f. Cleveland-Akron-Elyria, OH
                                                    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 hour 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.  The hour

                   refers to the start of the 8-h averaging period.


Source:  Fitz-Simons et al. (2005).
August 2005
3 -22        DRAFT-DO NOT QUOTE OR CITE

-------
         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 hour 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.  The hour
                   refers to the start of the 8-h averaging period.

Source:  Fitz-Simons et al. (2005).
August 2005
3-23         DRAFT-DO NOT QUOTE OR CITE

-------
 1     to about 6 p.m., with some deviations from these times.  For example, as shown in Figures 3-1 la
 2     for Boston and 3-1 Ik for Sacramento, the highest 8-h daily maximum values occur starting in
 3     mid-afternoon and extending into late evening. These results suggest that transport processes are
 4     playing the dominant role in determining the timing of the highest daily maxima in these areas.
 5           On days with high 1-h daily maximum concentrations (e.g., >0.12 ppm) the maxima tend
 6     to occur in a smaller time window centered in the middle of the afternoon, compared to days in
 7     which the maximum is lower. For example, on the high O3 days the 1-h maximum occurs from
 8     about 11 a.m. to about 6 p.m.  However, on days in which the 1-h daily maximum is
 9     <0.080 ppm, the daily maximum can occur at any time during the day or night, with only a  50%
10     probability that it  occurs between 1 and 3 p.m., in each of the 12 cities.  (The time of day when
11     the daily maximum  1-h O3 concentration occurs is illustrated for four of the cities in Figures
12     AX3-45a-d.).  Photochemical reactions in combination with diurnal emissions patterns are
13     expected to produce mid-afternoon peaks in urban areas. These results suggest that transport
14     from outside the urban airshed plays the major role for determining the timing of the daily
15     maxima for low peak O3 levels. This pattern is typical for the Los Angeles-Long Beach-
16     Riverside, CA area even on high O3 days.
17           The same general patterns emerge for the timing of the 1-h daily maximum O3
18     concentration as are found for the daily maximum 8-h average O3 concentration. As mentioned
19     above, the daily maximum 8-h O3 concentrations are generally found between the hours of
20     10 a.m. and 6 p.m. However, there are significant fractions of the time when this is not the case,
21     e.g., for high values in Houston, TX and Los Angeles, CA, or in general for lower values at any
22     of the cities examined.  (The time of day when the daily maximum 8-h average O3
23     concentrations occurs is shown for four cities in Figures AX3-46a-d.).  Although the 8-h
24     average O3 concentration is highly correlated with the daily maximum 1-h average O3
25     concentration, there are situations where the daily maximum 8-h average O3 concentration might
26     be driven by very  high values in the daily maximum  1-h average O3 concentration as illustrated
27     in Figure 3-1 Oj for Houston, TX.  In cases such as these, the predicted 8-h average may
28     overestimate the short-term O3 concentration later in the day.
29           The patterns of diurnal variability for both 1-h and 8-h averages have remained quite stable
30     over the 15-year period from 1990 to 2004, with times of occurrence of the daily maxima
31     varying by no more than an hour from year to year in each of the 12 cities.

       August 2005                               3-24        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Weekday/Weekend Differences
 2           Differences in the diurnal behavior of O3 have been observed in a number of cities (e.g.,
 3      Heuss et al., 2003). Figures 3-12a-h show the contrast in the patterns of hourly averaged O3 in
 4      the greater Philadelphia, Atlanta, Houston and Los Angeles areas from weekdays to weekends.
 5      Daily maximum concentrations occur basically at the same time on either weekdays or
 6      weekends. Differences are apparent in the hourly concentrations, especially in the extreme
 7      values. Weekday/weekend differences in 8-h average O3 concentrations are shown in Figures
 8      3-13a-h. As can be seen from a comparison of the weekend versus weekday patterns, there is a
 9      tendency for the lowest values to be higher on weekends than on weekdays. Lower traffic
10      volumes, in particular diesel truck traffic, lead to less NO emissions and titration of O3 on
11      weekends. The spike in values shown for Houston in midmorning shown in Figure 3-12f
12      resulted from the release of highly reactive hydrocarbons from the petrochemical industry
13      (which could occur on  any day of the week).  Otherwise, the maximum O3 concentrations could
14      be seen to occur on the weekdays as they do in Philadelphia and Atlanta, in contrast to
15      Los Angeles. Indeed, the diurnal pattern in Houston is similar to that observed in Atlanta on
16      weekdays, indicating some overall similarity in the sources of O3.
17
18      Spatial Variability in Diurnal Patterns in Urban Areas
19           Daily maxima in  either the  1-h or 8-h averages do not necessarily occur at the same time of
20      day at each site in the 12 cities, and the diurnal pattern observed at individual  sites can vary from
21      the composites shown in Figures 3-8 and 3-9. Differences between sites are not only related to
22      distance; they also depend on the presence of nearby sources, such as highways. For example, in
23      the Los Angeles basin, daily 1-h maxima are reached in the late afternoon in Riverside relatively
24      close to sites in which the maximum is reached much earlier.
25           The general pattern that emerges from the site-to-site variability within the urban areas
26      examined is that peaks in 1-h average concentrations are higher and tend to occur later at
27      downwind sites than in the urban cores.  To the extent that monitoring sites are either near to or
28      remote from sources of precursors in urban/suburban areas, the behavior of O3 will follow these
29      basic patterns. Similar relations are found for the 8-h average O3 concentrations. Differences in
30      diurnal patterns between sites in urban cores and sites downwind of urban cores are illustrated
        August 2005                              3-25        DRAFT-DO NOT QUOTE OR CITE

-------
             a. Philadelphia-Camden-Vineland, PA-NJ-DE-MD
                             (week day)
                                                        b. Philadelphia-Camden-Vineland, PA-NJ-DE-MD
                                                                       (week end)
        E
        a.
        a.
        o
        o
                                                 ~ 0,125
                                                 Q.
                                                 a
                                                        _o
                                                        I
                                                        £ 0.075 •
                                                  o
                                                  O
                                                         o
                                                         8
              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)
                                                        22230001 02 03 04 05 06 07 OB 09 10 11 12131415161718192021 22230001
                                                                           hour

                                                        d. Atlanta-Sandy  Springs-Gainesville, GA-AL
                                                                      (week end)
        B.
        D.
        O
        O
                                                 -~ 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
                                                E
                                                g 0,150
                                                       _o
                                                       I
                                                       £  0,100 •
                                                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

             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.
       Source: Fitz-Simons et al. (2005).
1      Northern Hemisphere during late winter and spring versus summer as for urban sites or for
2      nonurban sites heavily affected by regional pollution sources.
3           High O3 values are also found at some of the 12 cities outside of summer. The seasonal
4      behavior of O3 varies across the 12 cities.  In most northern cities, the extreme values of the daily
5      maximum 8-h average O3 concentration are a little more than half of those during the
6      warm season, the ratios of the medians are more similar as can be judged by comparison of
       August 2005
                                            3-27
DRAFT-DO NOT QUOTE OR CITE

-------
              a. Philadelphia-Camden-Vineland, PA-NJ-DE-MD

                             (week day)
          b. Philadelphia-Camden-Vineland, PA-NJ-DE-MD

                          (week end)
        E 0,100
        a.
        o
        S5 0075
        2
    E t)-ioo
    a.
    &
    c
    o
    S5 0.075
    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
          22230001 02 03 04 05 06 07 OB 09 10 11 12131415161718192021 22230001

                             hour
              c. Atlanta-Sandy Springs-Gainesville, GA-AL

                            (week day)
          d. Atlanta-Sandy Springs-Gainesville, GA-AL

                        (week end)
        £ 0.100
        B.
        D.
        O
        5 0.075
        5 0.050

        O
    £ 0.100
    Q.
    a.
                                                          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.

                          The hour refers to the start of the 8-h averaging period.


       Source:  Fitz-Simons et al. (2005).
1      Figures 3-1 la-1 with Figures 3-14a-l.  Differences are even smaller for the southern cities.


2      Indeed, some of the highest O3 values are found in the Houston CSA outside of summer


3      (Figure 3-14J).
       August 2005
3-28         DRAFT-DO NOT QUOTE OR CITE

-------
             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
              b. New York-Newark-Bridgeport, NY-NJ-CT-PA
o.oao •
g. 0060- T
c T






- — ™
TTTTI
T





_
mi
IT T T I T T

\
r"| _L _L _L

n
IM M LJ
nnilll
,2 T T
1 1 1
g 0.040 •
u
c
nnynnnrn 1 lllllll^nPi
— § 0020- _ H n
Iiiinni ° llfflmllll
rinnn.







nr^iil 1 1 1
• - - _ nMr
Infflulytlt
          22 230001 02 03 04 050607 OS 00 1011 12 13 1*1 15 1S 17 18 19 2021 2223 00 01
                             hour
              22230CO' 020304050607 080910 11 12 13 14 1516 17 IS 192021 2223 DO 01
                                 hour
          c. Philadelphia-Camdcn-Vineland. PA-NJ-DE-ME
              d, Washington-Baltimore-Northern Virginia, DC-MD-VA-WV
    o
    o
        o
        o
          22 23 00 01 02 03 04 05 OB 07 08 09 10 11 12 '3 14 15 16 17 18 19 20 21 22 23 00 01
                             hour
              22230001 020304050607 0809 10 11 12 13 14 15 16 17 18 19 20 21 22 230001
                                 hour
          e. Atlanta-Sandy Springs-Gainesville, GA-AL
              f. Cleveland-Akron-Elyria, OH
          TTTT
TTTT
          22 23 00 C1 02 03 34050607 0809 10 11 12 13 14 15 13 17'8 192021 2223 0001
                             hour
              22 23 OC 01 02 0304050607 3809 10 11 12 13 14 1516 17 1S 192021 2
                                 hour
Figure 3-14a-f.   Diurnal variability in 8 hour averaged 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).
August 2005
    3-30         DRAFT-DO NOT QUOTE OR CITE

-------
        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 ?? 73 00 01
                            hour
                          2??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 192021 22 23 00 01
                                              hour
Figure 3-14g-l.   Diurnal variability in 8 hour averaged 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).
August 2005
                3-31         DRAFT-DO NOT QUOTE OR CITE

-------
                                     Rural (CASTNET) Sites
             0.200 i
           £
           D-  0.150
           a.
           o
             0.100 -
           c
           O
           o
           
-------
 1     (<0.100 ppm) and daily 1-h maxima at CASTNET sites have exceeded 0.150 ppm.  The diurnal
 2     variations in 8-h average O3 concentrations are also much smaller at the CASTNET sites than at
 3     the urban sites.  Note also that the maxima in 8-h average O3 concentrations are higher at the
 4     CASTNET sites than at the urban sites.
 5
 6
 7     3.5   TRENDS IN OZONE CONCENTRATIONS
 8          Year-to-year variability in the nationwide May to September, mean daily maximum 8-h O3
 9     concentrations are shown in Figure 3-17. The corresponding year-to-year variability in the 95th
10     percentile concentrations is shown in Figure 3-18. Data flagged because of quality control issues
11     were removed with concurrence by the local monitoring agency.  Only days with data for 18 of
12     24 hours were kept, and a minimum  of 115 of 153 days were required in each year.  Missing
13     years were filled in using simple linear interpolation, as done in EPA Trends reports. Year-to-
14     year variability in the 95th percentile values of the daily maximum 8-h O3 concentrations are
15     shown in Figure 3-18.  Sites considered in this analysis are shown in the map in Figure AX3-3.
16     As  was shown in Figures 3-1  and 3-2, most sites are located in the East. As can be seen from
17     Figure 3-17, the highest O3 concentrations have tended to decrease over the past 15 years, while
18     there has been little change in O3 concentrations near the center of the distribution.  This is
19     consistent with observations in Europe (Volz-Thomas et al., 2003). Mean O3 concentrations
20     were slightly lower in 2003 and 2004 than in earlier years. The summer of 2003 was slightly
21     cooler than normal in the East (Levinson and Waple, 2004) and the summer of 2004 was much
22     cooler than normal in the East (Levinson, 2005) accounting in part for the dip in O3 during these
23     two years. Observations of O3 at a number of sites in the Northern Hemisphere likewise do not
24     show convincing evidence of strong  upward trends during the 1990s (Oltmans  et al., 1998).
25     There may even have been a slight increase in O3 concentrations at the lower end of the
26     distribution throughout the monitoring period.  This would be consistent with data obtained in
27     Europe, showing that O3 minima increased during the 1990s. Reduced titration of O3 by reaction
28     with NO in response to reductions in NOX emissions may be responsible in large measure for this
29     finding.  The concentration of Ox (NO2 + O3) shows little if any increase at all (Volz-Thomas
30     et al., 2003).
       August 2005                              3-33       DRAFT-DO NOT QUOTE OR CITE

-------
                                    Nationwide Trends, May to September
                              Mean of Daily Maximum 8-Hour Values, 1990 - 2004
                  0.12-
                  0.11 -
              -p  0.10-
              c
               Q.
               Q.
               E
               o
              4M
               E
              *^
               o
               u
               c
               o
              o
               
-------
                                       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).
 1     However, several monitoring sites were moved during the period from 1990 to 2004. Sites were
 2     moved at Acadia NP in 1996, Joshua Tree NP in 1993, Mammoth Cave NP in 1996, Voyageurs
 3     NP in 1996, and Yellowstone NP in 1996. These moves often resulted in offsets in O3 and so
 4     trends for these locations have not been calculated (cf, Section AX3.6,  Table AX3-9).  As noted
 5     in The Ozone Report—Measuring Progress through 2003 (U.S. Environmental Protection
 6     Agency, 2004b), O3 trends in national parks in the South and the East are similar to nearby urban
 7     areas and reflect the regional nature of O3 pollution. For example, O3 trends in Charleston, SC
 8     and Charlotte, NC track those in nearby Cowpens NP and Cape Romaine NP in South Carolina;
 9     O3 in Knoxville and Nashville, TN tracks O3 in Great Smoky NP;  O3 in Philadelphia, PA and
10     Baltimore, MD tracks Brigantine NP in New Jersey; and New York, NY and Hartford, CT track
11     O3 in Cape Cod NS. The situation is not as clear in the West, where national parks are affected
       August 2005
                          3-35
DRAFT-DO NOT QUOTE OR CITE

-------
 1      differently by pollution sources that are located at varying distances away (e.g., Lassen Volcanic
 2      National Park and Yosemite National Park, CA). However, data obtained at these sites still
 3      provide valuable information about the variability in regional background concentrations,
 4      especially since the West has not been broken down into regions as has been done by Lehman
 5      et al. (2004) for the East and shown in Figure 3-3. Comparison of Figures 3-19a-h and 3-20a-h
 6      (in conjunction with Table AX3-9) shows that  O3 concentrations near the center of the
 7      distribution do not necessarily track those at the upper end, as pointed out earlier for nationwide
 8      composite data.
 9          Caution should be exercised  in using trends calculated at national parks to infer
10      contributions from distant sources  either inside or outside of North America, because of the
11      influence of regional pollution. For example, using a 15-year record of O3 from Lassen Volcanic
12      NP and data from two aircraft campaigns, and  observations spanning 18 years from five U.S.
13      west coast marine boundary layer sites, Jaffe et al. (2003) have estimated that the amount of O3
14      in air arriving from the Eastern Pacific in spring has increased by approximately 10 ppb from the
15      mid-1980s to the present. This positive trend might be due to increases of emissions of O3
16      precursors in Asia. Positive trends in O3 were  found during all seasons.  Although the Lassen
17      Volcanic NP site is not close to any major emission sources or urban centers, maximum hourly
18      average O3 concentrations of >0.080 ppm (during April-May) and >0.100 ppm (during the
19      summer) occur at Lassen Volcanic NP. Thus,  although there is evidence that O3 levels may be
20      increasing at some rural locations,  there is also evidence that O3 levels at other locations have
21      either not increased or have decreased over the same period.
22
23
24      3.6 RELATIONSHIPS  BETWEEN OZONE  AND OTHER SPECIES
25      Correlations between Ozone and other Species
26          In order to understand relationships among atmospheric species, an important distinction
27      must be made between primary (directly emitted) species and secondary (photochemically
28      produced) species. In general, it is likely that primary species will be highly correlated with
29      other primary species, and that secondary species will be highly correlated with other secondary
30      species.  By contrast, primary  species are less likely to be correlated with secondary species.
31      Secondary reaction products tend to correlate with each other, but there is considerable variation.

        August 2005                              3-36       DRAFT-DO NOT QUOTE OR CITE

-------
                May to September Mean of Daily Maximum 8-Hour Values, 1990 - 2004
        1990   1992   19
                         1996   19BB

                            Year
                                    2000   2002   2004
                                                         1990   1992
                                                                 1996   1998

                                                                   Year
                                                                                      2000   2002   2004
     0.09 :


     0.08 -


     0.07 -


     o.oe -


     0.05 -


     0.04 -


     0.03 •


     0.02 -


     0.01
e, Theodore Roosevelt NP
                                                       0.09
        1990   1992
                         1996   1998

                            Year
                                    2000   2002
                                             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
       0.11 -,
       0.10-
       0.09-
       0.08-
       0.07-
       0.06-
       0.05-
       0.04-
       0.03-
       0.02 \
a. Brigantine NWR
                           1996   1988
                              Year
       0.11
       0.10-
       0.09-
       0.08-
       0.07-
       0.06-
       0.05-
       0.04-
       0.03-
       0.02 -V
           c. Great Smoky Mountains NP
                           1996   1998
                              Year
       0.11
       0.10-
       0.09-
       0.08-
       0.07-
       0.06
       0.05-
       0.04-
       0.03-
       0.02
       0.11
           e. Theodore Roosevelt NP
                           1996   1998
                              Year
       0.10-
       0.09-
       0.08-
       0.07-
       0.06-
       0.05
       0.04-
       0.03-
       0.02-V
           g. Glacier NP
                           1998   1998
                              Year
                                         0.10-
                                         0.09-
                                         0.08 •
                                         0.07-
                                         0.06-
                                         0.05 •
                                         0.04-
                                         0.03-
                                             d. Voyageurs NP
                                         0.02
                                           1

                                         0.11
                                         0.10
                                         0.09-
                                         0.08-
                                         0.07 -
                                         0.06-
                                         0.05
                                         0.04-
                                         0.03-
                                                          0.02
                                                              f. Yellowstone NP
                                                             1996   1998
                                                                Year
                                         0.10 •
                                         0.09-
                                         0.08 •
                                         0.07-
                                         0.06-
                                         0.05-
                                         0.04
                                         0.03 -
                                                          0.02
                                                              h. Chirlcahua NM
 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).

August 2005
                                  3-38
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Some species (e.g., O3 and organic nitrates) are closely related photochemically and are highly
 2      correlated. Others (e.g., O3 and H2O2) show a more complex correlation pattern.  Further details
 3      are given in Annex AX3 in Section AX3.7.
 4           Relationships between primary and secondary components are illustrated by considering
 5      data for O3 and PM2 5. Ozone and PM2 5 concentrations observed at a monitoring  site in Fort
 6      Meade, MD are plotted as binned means for different intervals in Figure 3-21, based on data
 7      collected between July 1999 and  July 2001.  As can be seen from the figure, PM2 5 tends to be
 8      negatively correlated with O3 to the left  of the inflection point (at about 30 ppbv O3) and tends to
 9      be positively correlated with O3 to the right of the inflection point.  Data to the left of the
10      minimum in PM25 were collected mainly during the cooler months of the year, while data to the
11      right of the minimum were collected during the warmer months. This situation arises because
12      PM2 5 contains a large secondary  component during the summer and has a larger primary
13      component during winter. During the winter, O3 comes mainly from the free troposphere, above
14      the planetary boundary layer and, thus, may be considered a tracer for relatively clean air, and it
15      is titrated by NO in the polluted boundary layer.  Unfortunately, data for PM25 and O3 are
16      collected concurrently at relatively few U.S. sites throughout an entire year.  So these results,
17      while highly instructive, are not readily  extrapolated to areas where appreciable photochemical
18      activity occurs throughout the year. Ito  et al. (2005) examined the  relation between PM10 and O3
19      on a seasonal basis in several urban areas (cf, Figure 7-24). Although PM10 contains
20      proportionately more primary material than does PM2 5, relations similar to those  shown in
21      Figure 3-21 are found, reflecting  the dominant contribution of PM2 5 to PM10.
22
23      Other Oxidants
24           Measurements of gas phase peroxides in the atmosphere were reviewed by Lee et al.
25      (2000). Ground level measurements of H2O2 taken during the 1970s indicated values of 180 ppb
26      in Riverside, CA and 10 to 20 ppb during smog episodes in Claremont and Riverside, with
27      values approaching 100 ppb in forest fire plumes. However, later surface measurements always
28      found much lower values. For example, in measurements made in Los Angeles and nearby areas
29      in the 1980s, peak values were always less than about 2 ppb and in a methods intercomparison
30      study in Research Triangle Park,  NC in June 1986, concentrations were <2.5 ppbv.  Higher
        August 2005                              3-39       DRAFT-DO NOT QUOTE OR CITE

-------
                    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).
 1     values ranging up to 5 ppb were found in a few other studies in Kinterbish, Alabama and
 2     Meadview, Arizona. Several of these studies found strong diurnal variations (typically about a
 3     factor of three) with maximum values in the mid-afternoon and minimum values in the early
 4     morning. Mean concentrations of organic hydroperoxides at the surface at Niwot Ridge, CO in
 5     the summer of 1988 and State Park, GA during the summer of 1991 were all less than a few ppb.
 6          Aircraft measurements of hydroperoxide (H2O2, CH3OOH and HOCH2OOH)
 7     concentrations were made as part of the Southern Oxidants Study intensive campaign in
 8     Nashville, TN in July 1995 (Weinstein-Lloyd et al., 1998).  The median concentration of total
 9     hydroperoxides in the boundary layer between 1100 and 1400 CDT was about 5 ppbv, with more
10     than 50% contribution from organic hydroperoxides.  Median O3 was about 70 ppbv at the same
11     time. The concentrations of the hydroperoxides depended strongly on wind direction with values
12     about 40% lower when winds originated from the N/NW as opposed to the S/SW suggesting that
13     local source areas were important.
       August 2005
               3-40
                   DRAFT-DO NOT QUOTE OR CITE

-------
 1          Peroxyacetylnitrate (PAN) is produced during the photochemical oxidation of a wide range
 2     of VOCs in the presence of NOX. It is removed by thermal decomposition and also by uptake to
 3     vegetation (Sparks et al., 2003; Teklemariam and Sparks, 2004).  PAN is the dominant member
 4     of the broader family of peroxyacylnitrates (PANs) which includes as other significant
 5     atmospheric components peroxypropionyl nitrate (PPN) of anthropogenic origin, and
 6     peroxymethacrylic nitrate  (MPAN) produced from oxidation of isoprene.  Measurements and
 7     models show that PAN in  the United States includes major contributions from both
 8     anthropogenic and biogenic VOC precursors (Horowitz et al., 1998; Roberts et al., 1998).
 9     Measurements in Nashville during the 1999 summertime Southern Oxidants Study (SOS)
10     showed PPN and MPAN amounting to 14% and 25% of PAN respectively (Roberts et al., 2002).
11     Measurements during the  TexAQS 2000 study in Houston indicated PAN concentrations of up to
12     6.5 ppbv (Roberts et al., 2003).  PAN measurements in southern California during the
13     SCOS97-NARSTO study  indicated peak concentrations of 5-10 ppbv, which can be contrasted to
14     values of 60 to 70 ppbv measured back in  1960 (Grosjean, 2003).  Vertical profiles measured
15     from aircraft over the U.S. and off the Pacific coasts show PAN concentrations above the
16     boundary layer of only a few hundred pptv, although there are significant enhancements
17     associated with long-range transport of pollution plumes including from Asia (Kotchenruther
18     et al., 2001a; Roberts et al., 2004). Decomposition of this anthropogenic PAN as it subsides
19     over North America can lead to significant O3 production, enhancing the O3 background
20     (Kotchenruther et al., 200Ib; Hudman et al., 2004).
21          Oxidants are also present in airborne cloud droplets, rain drops and particulate matter.
22     Measurements of hydroperoxides, summarized by Reeves (2003), are available mainly for
23     hydrometeors, but are sparse for ambient particles.  Venkatachari et al. (2005a) sampled the
24     concentrations of total reactive oxygen species (ROS) in particles using a cascade impactor in
25     Rubidoux, CA during July 2003. Although the species constituting ROS were not identified, the
26     results were reported in terms of equivalent H2O2 concentrations. Unlike O3 and gas phase H2O2
27     which show strong diurnal variability (i.e., about a factor of three variation between afternoon
28     maximum and early morning minimum), the diurnal variation of particle phase ROS was found
29     to be much weaker (i.e., less than about 20%) at least for the time between 8 a.m. and midnight.
30     Because the ROS were measured in the  fine aerosol size fraction, which has a lifetime with
31     respect to deposition of much greater than a day, little loss is expected but their concentrations

       August 2005                             3-41        DRAFT-DO NOT QUOTE OR CITE

-------
 1      might also be expected to increase because of nighttime chemistry, perhaps involving NO3
 2      radicals. The ROS concentration, about 7 x 1CT9 M/m3 (expressed as equivalent H2O2), was at
 3      most 1% that of O3 (6.2 to 38 x 1CT7 M/m3 or 15 to 90 ppb), with highest values at night. In a
 4      companion study conducted in Queens, NY during January and early February 2004,
 5      Venkatachari et al. (2005b) found much lower concentrations of ROS of about 1.2 x 10~9 M/m3.
 6      However, O3 levels were also substantially lower,  but ROS concentrations were still less than
 7      1% those of O3. It is of interest to note that gas phase OH concentrations measured at the same
 8      time ranged from about 7.5 x lOVcm3 to about 1.8 x 106/cm3, implying the presence of
 9      significant photochemical activity even in New York City during winter.
10
11      Co-occurrence of Ozone with Other Pollutants
12           The characterization of co-occurrence patterns under ambient conditions is important for
13      relating human health and vegetation effects under ambient conditions to controlled research
14      results as described in Annex AX3.8.  Several attempts have been made to characterize gaseous
15      air pollutant mixtures. The previous 1996 O3 AQCD discussed various patterns of pollutant
16      mixtures of SO2, NO2, and O3. Pollutant combinations can occur at or above a threshold
17      concentration at either the same or different times.
18           The 1996 O3 AQCD noted that studies of the joint occurrence of gaseous NO2/O3 and
19      SO2/O3 reached two conclusions:  (1) hourly simultaneous and daily simultaneous-only
20      co-occurrences are fairly rare (when both pollutants were present at an hourly average
21      concentration >0.05 ppm) and (2) when co-occurrences  are present, complex-sequential and
22      sequential-only co-occurrence patterns predominate. Year-to-year variability was found to be
23      insignificant.
24           Using 2001 hourly data for O3 and NO2 and for O3 and SO2, the co-occurrence patterns for
25      the data are similar to those of previous studies.  As shown in Figure 3-22, fewer than
26      10 co-occurrences of O3 and NO2 were found for most of the collocated monitoring sites.
27      Likewise, Figure 3-23 shows that fewer than 10  co-occurrences of O3 and SO2 were found for
28      most of the collocated monitoring sites analyzed.
29
30
        August 2005                              3-42       DRAFT-DO NOT QUOTE OR CITE

-------
    (1$
    0»

   W 25 -
   4-

    2 20
    OS
           0*   Cw   Cw  C«   01   01   CM   Cw   Oi   Sp   01   O>   CM
                            Number of Co-Occurrences (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.
      60
      50 -|
   W
      30 H
   0)
   A
      20-
       10-1
               I  .
           O)
                    o>
                    CM
                    O
                    CM
O)
CO
O
CO
                            CD
O
ID
    CD
    CD
O
CD
                                        0)
        CJ)
        CO
O
00
                                                 CD
O
CD
        O   ?
             CD
             CM
O)
OO
o>
in
                                                     o
                                                     O
                                      o
                                      CM
                                  O
                                  OO
                                      o
                                      in
                           Number of Co-Occurrences (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.
August 2005
               3-43
                  DRAFT-DO NOT QUOTE OR CITE

-------
 1          Since 1999, monitoring stations across the United States have routinely measured 24-h
 2     average concentrations for PM2 5. Daily co-occurrence of PM2 5 and O3 over a 24-h period was
 3     also characterized. Because PM2 5 data are mostly summarized as 24-h average concentrations in
 4     the AQS database, a daily co-occurrence of O3 and PM25 was subjectively defined as an hourly
 5     average O3 concentration >0.05 ppm and aPM25 24-h concentration >40 |ig/m3 (corresponding
 6     to the EPA Air Quality Index, Level of Concern for PM2 5) occurring during the same 24-h
 7     period. Using 2001 data from the AQS database, the daily co-occurrence of PM25 and O3 was
 8     infrequent (Figure 3-24). Only limited data are available on the co-occurrence of O3 and other
 9     pollutants (e.g., acid precipitation and  acidic cloudwater). In most cases, routine monitoring data
10     are not available from which to draw general conclusions.
               160
               140-
             Sg 120-
               100-
             03
             £3
             E
             3
                80-
                40-
                20-
                 0
                    o>
                            O
                            cM
O
co
O
^
O
m
O
co
OO
h-co
O
cn
                                                             oo
                                                             o   t-
                                    o
                                    c\i
                                    o
                                    ro
                                    o
                                    -=3-
                                    o
                                    uo
                                    Number of Co-Occurrences (Hours)
       Figure 3-24. The co-occurrence pattern for O3 and PM2 5 using 2001 data from AQS.


 1     3.7   POLICY RELEVANT BACKGROUND OZONE CONCENTRATIONS
 2          Background O3 concentrations used for NAAQS-setting purposes are referred to as Policy
 3     Relevant Background (PRB) O3 concentrations. Policy Relevant Background concentrations are
 4     those concentrations that would occur in the United States in the absence of anthropogenic
 5     emissions in continental North America (defined here as the United States, Canada, and
 6     Mexico). Policy Relevant Background concentrations include contributions from natural sources
       August 2005
               3-44
                      DRAFT-DO NOT QUOTE OR CITE

-------
 1      everywhere in the world and from anthropogenic sources outside these three countries.  For the
 2      purposes of informing decisions about O3 NAAQS, EPA assesses risks to human health and
 3      environmental effects from O3 levels in excess of PRB concentrations. Issues concerning the
 4      methodology for estimating PRB O3 concentrations are described in detail in Annex AX3,
 5      Section AX3.9.
 6           Contributions to PRB O3 include photochemical actions involving natural emissions of
 7      VOCs, NOX, and CO as well as the long-range transport of O3 and its precursors from outside
 8      North America and the stratospheric-tropospheric exchange (STE) of O3. Processes involved in
 9      STE are described in detail in Annex AX2.3. Natural sources of O3 precursors include biogenic
10      emissions, wildfires, and lightning. Biogenic emissions from agricultural activities are not
11      considered in the formation of PRB O3.
12           Springtime maxima are observed at relatively remote (Annex AX3 and Figures 3-25a,b)
13      national park sites, located mainly in the western United States and at a number of other
14      relatively unpolluted monitoring sites throughout the Northern Hemisphere.  The major issues
15      concerning the calculation of PRB O3 center on the capability of the current generation of global -
16      scale, three-dimensional, CTMs to simulate the causes of high O3 concentrations observed at
17      monitoring sites in relatively unpolluted areas of the United States from late winter through
18      spring (i.e., February through June). The issues raised do not affect interpretations of the causes
19      of summertime O3 episodes as strongly.  Summertime O3 episodes are mainly associated with
20      slow- moving high-pressure systems characterized by limited mixing between the planetary
21      boundary layer and the free troposphere, as noted in Annex AX2, Section AX2.3.
22           A large number of case studies document the occurrence of STE mainly during winter and
23      spring in mid- and high-latitudes in Europe, Asia, and North America. These studies were based
24      on aircraft, satellite, and ground-based measurements. Considerable uncertainty exists in the
25      magnitude of the exchange; however, these studies have found that STE occurs throughout the
26      year, but with a distinct preference for the transport of O3 directly to the middle and lower
27      troposphere during late winter and spring. Transport to the upper troposphere occurs throughout
28      the year.
29           Springtime maxima in tropospheric O3 observed at high latitudes are also associated with
30      the winter buildup of O3 precursors and thermally labile reservoir species,  such  as PAN and
31      other reactive nitrogen species.  These pollutants originate from all continents in the Northern

        August 2005                              3-45        DRAFT-DO NOT QUOTE OR CITE

-------
                            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 (2003a).

            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 (2003a).
August 2005
3-46
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Hemisphere. Ozone precursor concentrations reach a maximum in late March; and as sunlight
 2      returns to the Arctic, photochemical reactions generate tropospheric O3 (Section AX3.9.1).
 3      The contribution of Asian sources to the U.S. levels is also largest during spring, reflecting the
 4      efficient lifting of Asian pollution ahead of cold fronts originating in Siberia and transport by
 5      strong westerly winds across the Pacific (e.g., Hudman et al., 2004).  The longer lifetime of O3
 6      during spring also contributes to springtime maxima (Wang et al., 1998).
 7           Estimates of PRB concentrations cannot be obtained solely by examining measurements of
 8      O3 obtained at RRMS in the United States (Annex AX3, Section AX3.2.3) because of the
 9      long-range transport from anthropogenic source regions within North America.  It should also be
10      noted that it is impossible to determine sources of O3 without ancillary data that could be used as
11      tracers of sources or to calculate photochemical production and loss rates. The current definition
12      of PRB implies that only CTMs can be used to  estimate the range of PRB values. On the
13      synoptic and larger spatial scales at least, all evidence indicates that global CTMs are adequate
14      tools to investigate the factors controlling tropospheric O3; and three-dimensional CTMs, as
15      typified by Fiore et al.  (2003) appear to offer the best methodology for estimating PRB
16      concentrations that cannot be measured directly (Annex AX3, Section AX3.9.2), at least for
17      averaging periods of longer than one hour.
18           Previous estimates of background O3 concentrations, based on different concepts of
19      background, are given  in Table 3-2.  Results from global three-dimensional CTMs, where the
20      background is estimated by zeroing anthropogenic emissions in North America (Table 3-8) are
21      on the low end of the 25 to 45 ppbv range.  Lefohn et al. (2001) have argued that frequent
22      occurrences of O3 concentrations above 50 to 60 ppbv at remote northern U.S. sites in spring are
23      mainly stratospheric in origin.  Fiore et  al. (2003) used a global CTM to determine the origin of
24      the high-O3 events reported by Lefohn et al. (2001), and to conduct a more general quantitative
25      analysis of background O3 as a function of season, altitude, and local O3 concentration.
26           Figure 3-26 shows a comparison between observations obtained at CASTNet sites and
27      model results of Fiore et al. (2003).  They classified the CASTNet monitoring sites into
28      low-lying sites (generally <1.5 km) and elevated sites (>1.5 km). All elevated sites are in the
29      West. Results were then aggregated to construct the cumulative probability distributions shown
30      in Figure 3-26 for the 58 low-altitude sites  and  the 12 high-altitude sites as well as for the three
31      seasons. The calculated mean background at the surface sites in spring is 27 ppbv, compared to

        August 2005                              3-47        DRAFT-DO NOT QUOTE OR CITE

-------
          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
                              1990-1998     Harvard Forest
                              Summer 1995   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)


35 (fall) - 45 (spring)8
                                15-30 (East)11
                                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).
1      23 ppbv in summer and fall. At these sites, the background is highest for O3 concentrations near

2      the center of the distribution, and it declines as total surface O3 concentrations increase, for

3      reasons summarized below and discussed by Fiore et al. (2002).  The observed O3 concentration

4      thus serves a surrogate for meteorological variability (i.e., stagnant versus ventilated conditions),

5      such that the background O3 is smaller on days when total O3 is highest. At the elevated sites,

6      the calculated mean background is 36 ppbv in spring versus 30 ppbv in the summer and fall.

7      Background concentrations in the fall resemble those in summer but show less variability and do

8      not exceed 40 ppbv anywhere in this analysis.
       August 2005
                               3-48
             DRAFT-DO NOT QUOTE OR CITE

-------
          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).
August 2005
                          3-49       DRAFT-DO NOT QUOTE OR CITE

-------
 1           Major conclusions from the Fiore et al. (2003) study (discussed in detail in Annex AX3,
 2      Sections AX3.9.3 and AX3.9.4) are:
 3         •   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.
 4         •   PRB O3 concentrations can be represented as a function of season, altitude, and total
              surface O3 concentration, as illustrated in Figure 3-26.
 5         •   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.
 6         •   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.
 7           Appropriate background concentrations should thus be allowed to vary as a function of
 8      season, altitude, and total  O3 level.  The diamonds in Figure 3-26 can be applied for this purpose.
 9      In particular, the depletion of the background during high-O3 events should be taken into account
10      (i.e., background O3 is depleted by reactions in the atmosphere and by deposition to the surface
11      but is not replenished at a significant rate in the stable, polluted boundary layer). This depletion
12      is shown in the right-hand panels of Figure 3-26 for the highest O3 values. Note that the model
13      is generally able to reproduce the overall frequency  distributions in Figure 3-26. Typically,
14      models produce distributions flatter than are observed.  Underpredictions, especially at the upper
15      end of the frequency distribution during the warmer months, are likely related to sub-grid-scale
16      processes that the model cannot resolve explicitly.  The highest observed O3 concentrations in all
17      three seasons and at  all  altitudes are associated with regional pollution (i.e., North American
18      anthropogenic emissions), rather than stratospheric influence.
19           Chemistry transport models should be evaluated with observations given earlier in
20      Chapter 3, in Annex AX3, and to simulate the processes causing the intra-day variability in O3
21      concentrations shown in Figure 3-27 in addition to those summarized in Chapter 2. The diurnal
22      patterns shown in Figure 3-27 do not fit the smooth  pattern shown in Figure 3-15 and indicate
        August 2005                               3-50        DRAFT-DO NOT QUOTE OR CITE

-------
         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).
August 2005
3-51
DRAFT-DO NOT QUOTE OR CITE

-------
 1      processes capable of producing rapid rises in O3 at times when substantial photochemical
 2      activity is not present and may indicate stratospheric effects.  Higher resolution models capable
 3      of spatially and temporally resolving stratospheric intrusions  and capable of resolving O3
 4      variability on hourly timescales have not been applied to this  problem.  Ebel et al. (1991) have
 5      demonstrated that regional-scale CTMs could be used to study individual stratospheric
 6      intrusions. As an example of the utility of different types of models, Zanis et al. (2003) were
 7      able to forecast, observe, and model a stratospheric intrusion  (maximum penetration depth was
 8      to slightly >2 km altitude) that occurred from June 20 to 21, 2001, over a large swath of central
 9      Europe.  Roelofs et al. (2003) compared results from  six global tropospheric CTMs with lidar
10      observations obtained during that event and concluded that the models qualitatively captured the
11      features of this intrusion. It was also found that the coarser resolution models overestimated
12      transport to lower altitudes.  The use of higher resolution models, perhaps nested inside the
13      coarser resolution models, may have helped  solve this problem. They would also better address
14      issues related to temporal (i.e., 1-h versus 8-h averages) and spatial (i.e., populated versus
15      remote areas) scales needed by policymakers.
16           Although many of the features of the day-to-day variability of O3  at RRMS in the United
17      States are simulated reasonably well by Fiore et al. (2003), uncertainties in the calculation of the
18      temporal variability of O3 originating from different sources on shorter time scales must be
19      recognized.  The uncertainties stem in part from an underestimate in the seasonal variability in
20      the STE of O3 (Fusco and Logan, 2003), the geographical variability of this exchange, and the
21      variability in the exchange between the free troposphere and the planetary boundary layer in the
22      model.
23           Ideally, the predictions resulting from an ensemble of models should be compared with
24      each other and with observations, so that the range of uncertainty inherent in the model
25      predictions can be evaluated.
26
27
28      3.8  OZONE EXPOSURE IN VARIOUS MICROENVIRONMENTS
29           Humans are exposed to O3 and related  photochemical oxidants through the exchange
30      boundary, the skin and the openings into the body such as the mouth, the nostrils, and punctures
31      and lesions in the skin (U.S.  Environmental Protection Agency, 1992; Federal Register, 1986).

        August 2005                              3-52        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Inhalation exposure to O3 and related photochemical oxidants is determined by pollutant
 2      concentrations measured in the breathing zone that is not affected by exhaled air as the
 3      individual moves through time and space. A discussion of the basic terminology associated with
 4      exposure appears in AX3.
 5
 6      Quantification of Exposure
 1           Ambient O3 concentrations vary with time of day (peaking during the latter portion of the
 8      day) and season and among locations.  Consequently, exposure to O3 will change as a function of
 9      time of day and as an individual moves among locations. A hypothetical exposure is
10      demonstrated in Figure 3-28.  The actual dose received also changes during the day and is
11      dependent on the O3 concentration in the breathing zone and the individual's breathing rate,
12      which is, in turn, dependent on the individual's level of exertion.
13           When measuring or modeling exposure to O3 and related photochemical oxidants
14      consideration should be given to the diurnal  weekly (weekday-weekend) and seasonal
15      variability. Peak concentrations lasting for several hours typically occur toward the latter
16      portion of the day during the summer months.  Regional O3 episodes often co-occur with high
17      concentrations of airborne fine particles, making it difficult to assess O3  dynamics and exposure
18      patterns.  Also, while there are few indoor O3 sources, O3 will react with materials and other
19      pollutants in the indoor environment in an analogous fashion to that occurring in the ambient
20      atmosphere, potentially exposing subjects to other more toxic pollutants  (Nazaroff and Weschler,
21      2004; Lee and Hogsett, 1999; Wainman et al., 2000; Weschler and Shields, 1997). (See
22      discussion on O3 chemistry and indoor sources and concentrations later in this chapter.).
23
24      Personal Exposure and Ambient Concentrations
25           The two approaches for measuring personal exposure are (a) the direct approach, using a
26      personal exposure monitor (PEM) consisting of a passive sampler worn around the breathing
27      zone, and (b) the indirect approach, which measures or estimates the O3 concentrations through
28      the use of models or biomarkers. Both approaches are associated with measurement error.
29           Although it is difficult to develop passive monitors for personal  exposure measurements
30      because of problems in identifying chemical or trapping reagents that can react with O3, several
31      modified passive samplers have been developed for use in personal O3 exposure measurements

        August 2005                              3-53       DRAFT-DO NOT QUOTE OR CITE

-------
                   ro
                   +-*
                   C
                   0)
                   o
                   c
                   o
                   O
                                                     Peak
                                                   Exposure
                                            Averaged
                                            Exposure
                                                               Instantaneous
                                                             . j^' Exposure
                                                   Area = Integrated
                                                         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).
 1     (Bernard et al., 1999; Koutrakis et al., 1993; Avol et al., 1998b; Geyh et al., 1997, 1999).  Some
 2     personal exposure measurements using passive samplers show O3 exposures below those O3
 3     concentrations measured at outdoor stationary sites (Delfino et al., 1996; Avol et al., 1998b;
 4     Sarnat et al., 2000; Geyh et al., 2000; Brauer and Brook, 1997). However, other studies have
 5     found strong correlations between O3 measured at stationary sites and personal monitored
 6     concentrations (Liard et al., 1999; Bauer and Brook, 1997; Linn et al., 1996; Lee et al., 2004;
 7     Avol et al., 1998b; O'Neill et al., 2003) when the time spent outdoors, age, gender, and
 8     occupation of the subjects were considered.
 9          The indirect approach determines and measures the concentrations in all of the locations or
10     "microenvironments".  The concept of microenvironments is important in the understanding of
11     human exposure modeling.  Often identified with a perfectly mixed compartment,
12     microenvironments are more recently viewed as a controlled volume, indoors or outdoors, that
13     can be characterized using a set of either mechanistic or phenomenological governing equations.
14     This allows for a nonhomogeneous environment, including sources and sinks within the
       August 2005
3-54
DRAFT-DO NOT QUOTE OR CITE

-------
 1      microenvironment. Microenvironments include indoor residences, other indoor locations,
 2      outdoors near roadways, other outdoor locations, and areas within vehicles.
 3
 4      Microenvironmental Concentration and Ozone Exposure Models
 5           Outdoor concentrations of O3 are estimated either through emissions-based mechanistic
 6      modeling, or through ambient-data-based modeling.  Emissions-based models determine the
 7      spatiotemporal fields of the O3 concentrations using precursor emissions and meteorological
 8      conditions as inputs.  (They are described in Annex AX2.). The ambient-data-based models
 9      determine spatial or spatiotemporal distributions of O3 through the use of interpolation schemes.
10      The kriging approach provides standard procedures for generating an interpolated O3 spatial
1 1      distribution for a given period of time (Georgopoulos et al., 1997a,b). The Spatio-Temporal
12      Random Field (STRF) approach has been used to interpolate monitoring data in both space and
13      time (Christakos and Vyas, 1998a,b).  The STRF approach can analyze information on temporal
14      trends which cannot be directly incorporated by kriging.
15           Several approaches are available for modeling microenvironmental concentrations:
16      empirical, mass balance, and detailed computational fluid dynamics (CFD) models.  Empirical
17      relationships provide the basis for future, "prognostic" population exposure models. Mass
18      balance modeling is the most common approach used to model pollutant concentrations in
19      enclosed microenvironments. Mass balance modeling ranges from very simple formulations,
20      assuming ideal (homogeneous) mixing and only linear physicochemical transformations with
21      sources and sinks, to models that account for complex multiphase chemical and physical
22      interactions and nonidealities in mixing. Mass balance models take into account the effects of
23      ventilation, filtration, heterogeneous removal, and direct emission as well as photolytic, thermal,
24      and chemical reactions.  The simplest form of the model is represented by the following
25      differential equation:
26
                                             ^        S     ,-•<
27
28     where dCIN is the indoor pollutant concentration (mass/volume), dt is time in hours, v is the air
29     exchange rate, COUT is the outdoor pollutant concentration (mass/volume), Vis the volume of the
30     microenvironment, and S is the indoor source emission rate.  When the model was used to
       August 2005                             3-55       DRAFT-DO NOT QUOTE OR CITE

-------
 1      estimate indoor O3 concentrations, indoor concentrations were found to be 33% of outdoor O3
 2      concentrations (Freijer and Bloemen, 2000). A more in-depth discussion of the mass balance
 3      model has been reported in Nazaroff and Cass (1986).  The pNEM/O3 model, discussed later in
 4      this chapter, includes a sophisticated mass balance model for indoor and vehicle
 5      microenvironments (Johnson, 2003). CFD models take into account the complex, multiphase
 6      processes that affect indoor concentrations of interacting gas phase pollutants, such as the
 7      interactions of O3 with indoor sinks and sources (surfaces, gas releases) and with entrained gas
 8      (Sarwar et al., 2001, 2002; S0rensen and Weschler, 2002).
 9           Exposure modeling is often used in evaluating exposure to large populations over time.
10      The use of models is complicated by the fact that O3 is a secondary pollutant with complex
11      nonlinear and multiscale dynamics in space and time. Ozone is formed in the atmosphere
12      through a series of chemical reactions involving precursor VOCs and NOX. Therefore, O3
13      exposures may be affected by: (1) emission levels and spatiotemporal patterns of VOCs and
14      NOX; (2) ambient atmospheric as well as indoor microenvironmental transport, removal and
15      mixing processes; and (3) chemical transformations that take place over a multitude of spatial
16      scales.  The transformations are dependent on the presence of co-occurring pollutants  and the
17      nature of surfaces interacting with the pollutants.
18           Exposure models may be classified as (1) potential exposure models, typically the
19      maximum outdoor concentrations versus "actual" exposure, including locally modified
20      microenvironmental outdoor and indoor exposures; (2) population versus "specific individual"-
21      based exposure models; (3) deterministic versus probabilistic models; and (4) observation versus
22      mechanistic air quality model-driven estimates of spatially and temporally varying O3
23      concentrations.
24           There are several steps involved in defining exposure models. The steps are based on
25      frameworks described in the literature over the last 20 years and the structure of various existing
26      inhalation exposure models (NEM/pNEM, MENTOR/SHEDS, REHEX, TRJJVI.Expo also known
27      as APEX, AIRPEX, AIRQUIS). The steps include (1) estimation/ determination of the
28      background or ambient levels of O3; (2) estimation/determination of levels and temporal profiles
29      of O3 in various microenvironments; (3) characterization of relevant attributes of individuals or
30      populations under study (age, gender, weight, occupation, other physiological characteristics);
31      (4) development of activity event or exposure event sequences; (5) determination of appropriate
32      inhalation rates during the exposure events; (6) determination of dose; (7) determination of
        August 2005                              3-56       DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
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.
                           Air Exchange
                             Rates and
                             Building
                             Volumes
                       Emission Rates and
                         Use Patterns for
                         Indoor Sources
                       (e.g., gas appliances,
                        passive smoking)
                             Ambient
                             Fixed-Site
                           Concentrations
                                    Seasonal
                                  Considerations
                                   (Temperature)
"Human Activity1
 and Exertion
   Patterns
(Exercise Level
   Patterns)
Population and
 Commuting
   Data


Air Quality
Specification
                                                           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).
1           To estimate the actual O3 dose delivered to the lung, information on the concentration,
2      minute ventilation rate, activity level, and the morphology of the respiratory tract are needed.
3      Limited data have been compiled for ventilation rates for different age groups, both healthy and
4      compromised individuals, at varies levels of activity (Klepeis et al.,  1996, 2001; Avol et al.,
5      1998b; Adams, 1993). Based on the available information, the highest level of outdoor activity
6      occurs during the spring and summer months, during the mid- to late afternoon and early
7      evening—the times when O3 concentrations are highest. Children are likely more susceptible to
8      the effects of O3 than other groups. School-age children spend more time outdoors engaged in
9      high-level activities than do other groups and breath more air in than adults relative to body
       August 2005
                                            3-57
       DRAFT-DO NOT QUOTE OR CITE

-------
 1      surface area, breathing frequency, and heart rate. Asthmatic children spend the same amount of
 2      time outdoors as other more healthy children but the time spent engaged in high levels of activity
 3      are less.
 4           Estimates of activity level have been compiled based on questionnaire data.  The National
 5      Human Activity Pattern Survey (NHAPS), a probability-based telephone survey, was conducted
 6      in the early 1990s. The  survey concluded that outdoor work-related activities were highest
 7      during the springtime and were more frequent during the morning and early afternoon.
 8      Exercise/sports-related activities were highest from noon to 3 p.m. during the summer months.
 9      During the spring months, exercise/sports-related activities were highest from mid- to late
10      afternoon (Klepeis et al., 1996, 2001). A pilot study by Gonzales et al. (2003) evaluated the use
11      of retrospective questionnaires for reconstructing past time-activity and location pattern
12      information. Ozone  concentration estimates using  ambient stationary monitors and estimates
13      derived from diaries  and questionnaires differed slightly.  However, both estimates were greater
14      than O3 personal exposure measurements.
15           Existing comprehensive inhalation exposure models (NEM and pNEM) (Johnson, 2003),
16      (MENTOR/SHEDS) Burke et al., 2001; McCurdy et al., 2000), and the Air Pollutants Exposure
17      model (TRIM.Expo) treat human activity patterns as sequences of exposure events in which each
18      event is defined by a geographic location and microenvironment and then assigned activity diary
19      records from the CHAD (Consolidated Human Activities Database; www.epa.gov/chadnetl)
20      (Glen et al., 1997; McCurdy, 2000; McCurdy et al., 2000). There are now about 22,600 person-
21      days of sequential daily  activity pattern data in CHAD representing all  ages and both genders.
22      The data for each subject consist of one or more days of sequential activities, in which each
23      activity is defined by start time, duration, activity type (140 categories), and microenvironment
24      classification (110 categories). Activities vary from 1 min to 1 h in duration.  Activities longer
25      than 1 h are subdivided into clock-hour durations to facilitate exposure modeling.  A distribution
26      of values for the ratio of oxygen uptake rate to body mass (referred to as metabolic equivalents
27      or METs) is provided for each activity type listed.  A table listing the activity patterns included
28      in CHAD appears in AX3.
29           pNEM divides  the population of interest into  representative cohorts based on the
30      combinations of demographic characteristics (age, gender, and employment), home/work
31      district, and residential cooking fuel.  TRIM.Expo and MENTOR/SHEDS generate a population
32      demographic file  containing a user-defined number of person-records for each census tract of the
        August 2005                              3-58        DRAFT-DO NOT QUOTE OR CITE

-------
 1      population based on proportions of characteristic variables (age, gender, employment, and
 2      housing) obtained for the population of interest, and then assigns the matching activity
 3      information from CHAD to each individual record of the population based on the characteristic
 4      variables.
 5           The TRIM.Expo model is capable of simulating individual movement through time and
 6      space to provide estimates of exposure to a given pollutant in various microenvironments (e.g.,
 7      indoor, outdoor, and in-vehicle microenvironments). One of the key strengths of the
 8      TRIM.Expo model is  its ability to estimate hourly exposures and doses for all simulated
 9      individuals in a sampled population. However, TRIM.Expo is limited in that uncertainties in the
10      predicted distributions (e.g., age, activity data, commuting patterns, personal  activities) have not
11      been addressed.
12           MENTOR/SHEDS is capable of simulating individuals exposures in various
13      microenvironments (outdoors, residence, office, school, store, restaurant, bar, and vehicles)
14      using spatial concentration data for each census tract. The indoor and in-vehicle pollutant
15      concentrations are calculated using specific equations for the microenvironment and ambient
16      pollutant concentration relationship. Randomly selected characteristics for a fixed number of
17      individual are selected to match demographics within the census tract for age, gender,
18      employment status, and housing type.  Smoking prevalence statistics by gender and age is
19      randomly selected for each individual in the simulation. Diaries for activity patterns are matched
20      for the simulated individual by demographic characteristics (Burke et al., 2001).
21           An important source of uncertainty in existing exposure modeling involves the creation of
22      multiday, seasonal, or year-long exposure activity sequences based on 1- to 3-day activity data
23      for any given individual from CHAD.  Currently, appropriate longitudinal data are not available
24      and the existing models use various rules to derive longer-term activity sequences utilizing 24-h
25      activity data from CHAD.
26           Of the above models, only NEM/pNEM have been used extensively in O3 exposure
27      modeling. The pNEM probabilistic model builds on the earlier NEM deterministic exposure
28      model.  The model takes into consideration the temporal and spatial distribution of people and
29      O3 in the area of consideration, variations in O3 concentrations in the microenvironment, and the
30      effects of exercise-increased ventilation on O3 uptake.  There are three versions of the pNEM/O3
31      model:  (1) general population (Johnson et al., 1996a), (2) outdoor workers (Johnson et al.,
32      1996b), and (3) outdoor children (Johnson  et al., 1996c, 1997). The pNEM models have been
        August 2005                              3-59        DRAFT-DO NOT QUOTE OR CITE

-------
 1      applied to nine urban areas and a summer camp. The models used activity data from the
 2      Cincinnati Activity Diary Study (CADS) along with time-activity data from several other
 3      studies.  Data from stationary monitoring sites were used to estimate outdoor O3 exposure.
 4      Indoor O3 decay was assumed to be proportional to the indoor O3 concentration. An algorithm
 5      assigned the EVR associated with each exposure event. The EVR for the outdoor children
 6      model was generated using a module based on heart rate data by Spier et al.  (1992) and Linn
 7      etal. (1992).
 8
 9      Characterization of Exposure
10           The use of ambient air monitoring stations is the most common surrogate for assigning
11      exposure in epidemiological studies.  Since the primary source of O3 exposure is the ambient air,
12      monitoring concentration data would provide the exposure outdoors while exercising, a potential
13      important exposure to evaluate in epidemiological  studies. Monitored concentrations are useful
14      for a relative assignment of exposure with time if the concentration were uniform across the
15      region; the time-activity pattern were the same across the population; and the housing
16      characteristics, such as ventilation rates and the O3 sinks contributing to its indoor decay rates,
17      were constant for the study area. Since these factors vary by population and location there will
18      be errors in the magnitude of the total exposure and in the relative total exposure assignment
19      based solely on ambient monitoring data.
20           Personal O3 exposure measurements have been made for potentially susceptible
21      populations (children, outdoor workers, the elderly, and individuals with chronic obstructive
22      pulmonary disease).  Children and outdoor workers have somewhat higher exposures than other
23      individuals because they spend more time outdoors engaged in moderate and heavy exertion.
24      Children are also more active outside and, therefore, have a higher minute ventilation rate than
25      most adults (Klepeis et al.,  1996, 2001). Available exposure studies suggest trends in exposure
26      magnitude for some populations, however, additional exposure studies are needed to generalize
27      differences in exposure between the general population and potentially susceptible populations.
28      Table 3-3 summaries the findings of available exposure studies.
        August 2005                              3-60        DRAFT-DO NOT QUOTE OR CITE

-------
                              Table 3-3.  Personal Exposure Concentrations
        Location, Population, Sample Duration
    Personal Exposure Mean3
          (range) (ppb)
                       Reference
        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

        Southern California, Subjects 10-38 years       24
           Spring
           Fall

        Montpellier, France, Adults, Hourly           16
           Winter
           Summer

        Souther 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
          12 ± 12 (0-84)
           10 weekend
           12 weekday
             (ND-9)
            (ND-12)
             (2-44)
 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)
                       Delfinoetal. (1996)
                       Brauer and Brook
                       (1997)
                              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.
1           Ozone concentrations in various microenvironments under a variety of environmental

2      conditions have been reported in the literature.  In the absence of an indoor O3 source,

3      concentrations of O3 indoors are lower than that found in the ambient air. Ozone concentrations

4      in microenvironments were found to be primarily controlled by  ambient O3 concentrations and

5      the AER:  they increase with increasing AER.  To a lesser extent, O3 concentrations in

6      microenvironments are influenced by the ambient temperature, time of day, indoor

7      characteristics (e.g., presence of carpeting), and the presence of other pollutants in the

8      microenvironment. Table 3-4 describes the findings of the available studies.
       August 2005
3-61
     DRAFT-DO NOT QUOTE OR CITE

-------
 1     Factors Affecting Ozone Concentrations
 2           Ozone and other photochemical oxidants are formed in the ambient air from the reaction of
 3     sunlight with vehicle emissions, gasoline fumes, solvent vapors, and power plant and industrial
 4     emissions (See Chapter 2 for a discussion of O3 atmospheric chemistry). Ozone enters the
 5     indoor environment primarily through infiltration from outdoors through building components,
 6     such as windows, doors, and ventilation systems. There are also a few indoor sources of O3
 7     (photocopiers, facsimile machines, laser printers, and electrostatic air cleaners and precipitators)
 8     (Weschler, 2000).  Generally O3 emissions from office equipment and air cleaners are low
 9     except under improper maintenance conditions. Reported O3 emissions from office equipment
10     range from 1300 to 7900 |ig/h (Leovic et al., 1996,  1998). Most air cleaners (particulate
11     ionizers) emitted no or only a small amount (56 to 2757 |ig/h) of O3 during operation (Niu et al.,
12     2001). Emissions from O3 generators can range from tens to thousands of micrograms per hour
13     (Weschler, 2000; U.S. Environmental Protection Agency, 1996).
14           Other photochemical oxidants (peroxyacyl nitrates; PAN and PPN) have no known direct
15     emission sources indoors. PAN may be formed in the indoor environment from the reaction of
16     the OH- or NO3 with acetaldehyde to form the acetyl radical, CH3CO (Grosjean et al., 1996).
17     The acetyl radical then reacts with oxygen to for an acetylperoxy radical which reacts with NO2
18     to form PAN. Peroxyacyl nitrates primarily  occur in the indoor environment from infiltration
19     through the building envelop and through openings in the building envelopment.
20           The concentration  of O3 in indoor environments is dependent on the outdoor O3
21     concentration, the AER or outdoor infiltration, indoor circulation rate, and O3 removal processes
22     through contact with indoor surfaces and reactions with other indoor pollutants. Since O3
23     concentrations are generally higher during the warmer months, indoor concentrations will likely
24     be highest during that time period. (See earlier discussion on ambient concentrations of O3.).
25           Air exchange rates vary depending on temperature differences, wind effects, geographical
26     region, type of heating/mechanical ventilation system, and building type (Weschler and Shields,
27     2000; Colome et al., 1994). The balance of the flow of air in and out of a microenvironment is
28     greatest in a residential building when a window or door is open (Johnson et al., 2004; Howard-
29     Reed et al., 2002). The opening of windows or doors is dependent on the building occupancy,
30     season, housing density, the presence of air conditioning, and wind speed (Johnson and Long,
31     2004). When windows and doors are closed, the dominant mechanism controlling AERs is

       August 2005                              3-62        DRAFT-DO NOT QUOTE OR CITE

-------
                                 Table 3-4.  Indoor/Outdoor Ozone Concentrations in Various Microenvironments
OQ
 to
 o
 o
                                             Indoor/Outdoor
        Location and Ventilation Conditions   Concentrations
                             Comments
                                                                     Reference
          New England States (9)
           Fall
        Mexico City, School
             Windows/Doors Open (27)
             Windows/Doors Closed Cleaner
             Off (41)
             Windows/Doors Closed Cleaner
             On (47)

        Mexico City
             Homes
             Schools
20 ppb/40 ppb
0 to 247 ppb/
64 to 361 ppb
5 ppb/27ppb (7 d)
7 ppb/37 ppb (14 d)

22ppb/56to733ppb
Schools represented a variety of environmental conditions - varying          NESCAUM
ambient O3 concentrations, sources, geographic locations, population         (2002)
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.

Study conducted over 4 d period during winter months. Two-minute         Gold et al.
averaged measurements were taken both inside and outside of the             (1996)
school every 30 min from 10 a.m. to 4 p.m. Estimated air exchange
rates were 1.1, 2.1, and 2.5 IT1 for low, medium, and high flow rates.
Ozone concentrations decreased with increasing relative humidity.
Ozone monitoring occurred between September and July.  Study             Romieu
included 3 schools and 145 homes. Most of the homes were large and        et al. (1998)
did not have air conditioning. 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 was monitored at the schools sites
from 8 a.m. to 1 p.m. daily for 14 consecutive days.  Homes
were monitored for continuous 24-h periods for 7 and 14 consecutive days.
        Boston, MA, Homes (9)
             Winter - continuously

             Summer - continuously
        Los Angeles, Homes (239)
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 examined the potential for O3 to react with VOCs to form acid
aerosols.  Carbonyls were formed.  No clear trend of O3 with AERs.          Reiss et al.
The average AER was 0.9 IT1 during the winter and 2.6 IT1 during the         (1995)
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 home from February to December. Concentrations based on        Avol et al.
24-h average O3 concentrations indoors and outdoors. Low outdoor          (1998a)
concentrations resulted in low indoor concentrations. However, high
outdoor concentrations resulted in a range of indoor concentrations.

-------
                    Table 3-4 (cont'd).  Indoor/Outdoor Ozone Concentrations in Various Microenvironments
3Q
to
O
o


Location and Ventilation Conditions
Burbank, CA
Telephone Switching Station
Munich Germany
Office
Gymnasium
Classroom
Residence
Bedroom
Livingroom
Indoor/Outdoor
Concentrations
0.2/2 1.1 ppb
0.4/0. 9 ppb
0.49/0.92 ppb
0.54/0.77 ppb
0.47/1.0 ppb
0.74/1.0 ppb
Comments
Major source of O3 was transport from outdoors. From early spring to late
fall O3 concentrations peaked during the early afternoon and approach zero
at sunset. AER ranged from 1.0 to 1.9 h'1.
Indoor concentrations were dependent on the type of ventilation.

Reference
Weschler
etal. (1994)
Jakobi and
Fabian
(1997)

Montpellier, France, Homes (110)
Southern CA, Homes
     Upland
     Mountains
Krakow, Poland, Museums
     Cloth Hall
     Matejko
     Wawel Castle
     National
Buildings, Greece
     Thessalonki
     Athens
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
 Ozone measurements were made over 5-d periods in and outside             Bernard et al.
 of 21 homes during the summer and winter months. The winter              (1999)
 I/O ratio was 0.31 compared to 0.46 during the summer months.

Ozone measurements were taken at 119 homes (57 in Upland and 62 in        Geyh et al.
towns located in the mountains) during April and May. Concentrations        (2000)
were based on average monthly outdoor concentrations and average weekly    Lee et al.
indoor concentrations. Indoor based on the home location, number of         (2002)
bedrooms, and the presence of an air conditioner.

Ozone continuously monitored at five museums and cultural centers.          Salmon et al.
Monitoring conducted during the summer months for 21 to 46 h or 28 to       (2000)
33 days at each of the sites. The indoor concentration was found to be
dependent on the ventilation rate, i.e., when the ventilation rate was high the
indoor O3 concentrations approached that of ambient O3. Rooms sequestered
from the outdoor air or where air was predominantly recycled through
charcoal filters the O3 levels indoors were greatly reduced.

There was no heating/air conditioning system in the building at            Drakou et al.
Thessaloniki. Windows were kept closed during the entire monitoring      (1995)
period.  Complete air exchange took place every 3 h. The air
conditioning system in continuous use at the Athens site recirculated
the air.  Complete air exchange was estimated to be 1 h. Monitoring
lasted for 30 days at each site but only the 7 most representative days
were used.

-------
                           Table 3-4 (cont'd). Indoor/Outdoor Ozone Concentrations in Various Microenvironments
OQ
 to
 o
 o
        Location and Ventilation
        Conditions
Indoor/Outdoor
Concentrations
Comments
Reference
        Patrol cars, NC
        University of CA
            Photocopy room

        Home/office
            O3 generators
11.7/28.3 ppb




<20 to 40 ppb/—


14 to 200 ppb/—
Patrol cars were monitored Mon. through Thurs. between the hours of      Riediker
3 p.m. to midnight on 25 occasions during the months of Aug., Sept.,       et al. (2003)
and Oct. Outdoor O3 concentrations were taken from ambient
monitoring station. Air inside the patrol car was recirculated cool air.

Room volume was 40 m3. Ozone concentrations increased               Black etal.
proportionately with increasing use of photocopier.                      (2000)

Room volume was 27 m3. Doors and windows  were closed.              Steiber et al.
Heating/air conditioning and mechanical ventilation systems were off.      (1995)
Ozone generator was operated for 90 min. High O3 concentrations noted
when O3 generator used at high setting.  AER was 0.3 h~'.

-------
 1     infiltration through unintentional openings in the building envelope. Williams et al. (2003a,
 2     2003b) reported AERs of 0.001 to 4.87 hl in 37 homes in Research Triangle Park, NC. Chan
 3     et al. (2005) compared air leakage measurements for 70,000 houses. Older and smaller houses
 4     had higher normalized leakage areas than newer and larger houses. Meng et al. (2004) also
 5     attributed higher AERs to the age of the housing stock. AERs for homes in Houston, TX  and
 6     Elizabeth, NJ were averaged for all four seasons, the highest AER, 1.22 h"1, was noted for homes
 7     in Elizabeth, NJ where the homes were older. Evaluations of AERs for residential structures was
 8     reported by Murray and Burmaster (1995) and includes AERs for 2,844 residential structures in
 9     four different climatic regions by season (winter, spring, summer, and fall).  The AER for all
10     seasons across all regions was 0.76 h"1 (arithmetic mean) (Region 1:  IN, MN, MT, NH, NY1,
11     VT, WI; Region 2:  CO, CT, IL, NJ, NY2, OH, PA, WA; Region 3:  CA3, MD, OR, WA;
12     Region 4: AZ, CA4, FL, TX).  The AERs were generally higher during the warm seasons, when
13     ambient O3 concentrations are highest. Data for the warmest region during the summer months
14     may not be representative of all homes because measurements were made in southern California
15     where windows are open and air conditioning is not used.
16          Average mean (median) AERs of 2.45 (2.24),  1.35 (1.09), and 2.22 (1.79) h"1 were
17     reported by Lagus Applied Technology, Inc. (1995) for schools, offices, and retail
18     establishments in California. Mean AERs for schools, offices, and retail establishments in
19     Oregon and Washington were 0.32, 0.31, and 1.12 h"1 (Turk et al., 1989)—considerably less than
20     that reported by Lagus Applied Technology. Park et al. (1998) reported mean AERs ranging
21     from 1.0 to 47.5 h"1 for stationary vehicles under varying ventilating conditions. Where
22     available, AERs for other studies are included in Table 3-10.
23          The most important removal process for O3 in the indoor environment is deposition on and
24     reaction with indoor surfaces. The rate of deposition is material-specific. The removal rate will
25     depend on the indoor dimensions, surface coverings, and furnishings. Smaller rooms generally
26     have larger surface-to-volume ratio (A/V) and remove O3 faster than larger rooms. Fleecy
27     materials, such as carpets, have larger surface-to-volume ratios and remove O3 faster than
28     smooth surfaces (Weschler,  2000).  However, the rate of O3 reaction with carpet diminishes with
29     cumulative O3 exposure (Morrison and Nazaroff, 2000, 2002). Weschler (2000) compiled the O3
30     removal rates for a variety of microenvironments. Generally, the removal rates ranged between
31     3.0 and 4.3 kd (A/VyiT1.  The highest removal rate, 7.6 kd (A/VyiT1, was noted for a clean room
32     (Weschler etal, 1989).
       August 2005                              3-66        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Ozone chemical reactions in the indoor environment are analogous to those reactions
 2      occurring in the ambient air (See discussion on atmospheric chemistry in Chapter 2). Ozone
 3      reacts with unsaturated VOCs in the indoor environment, primarily terpenes or terpene-related
 4      compounds from cleaning products, air fresheners, and wood products. The reactions are
 5      dependent on the O3 indoor concentration, the indoor temperature and, in most cases, the air
 6      exchange rate/ventilation rate. Some of the reaction products may more negatively impact
 7      human health and artifacts in the indoor environment than their precursors (Wolkoff et al., 1999;
 8      Wilkins et al., 2001; Weschler et al., 1992; Weschler and Shields, 1997; Rohr et al., 2002;
 9      N0jgaard et al., 2005). Primary reaction products are Criegee biradicals, nitrate radicals, and
10      peroxyacetyl radicals.  Secondary reaction products are hydroxy, alkyl, alkylperoxy,
11      hydroperoxy, and alkoxy radicals. Reactions with alkenes can produce aldehydes, ketones, and
12      organic acids (Weschler and Shields, 2000; Weschler et al., 1992).
13           Hydroxyl radicals formed from the reaction of O3 with VOCs, nitric oxide and
14      hydroperoxy, and other intermediate products can react with various nitrogen compounds, sulfur
15      dioxide, carbon monoxide and other compounds to produce significantly more toxic compounds
16      (Sarwar et al., 2002; Orzechowska and Paulson, 2002; Pick et al., 2003, 2004; Van den Bergh
17      et al., 2000; Fan et al., 2003; Wilkins et al., 2001; Clausen et al., 2001; Rohr et al., 2002, 2003;
18      Poupard et al., 2005; Blondeau et al., 2005).  The reaction between O3 and terpenes also has been
19      shown to increase the  concentration of indoor particles (Weschler and Shields, 1999, 2003;
20      Weschler, 2004; Clausen et al., 2001; Fan et al, 2003; Wainman et al., 2000), possibly from
21      further reactions of the hydroxy radical with terpenes (Sarwar et al., 2002).
22           Decomposition and formation of PAN in the indoor environment are influenced by NO2
23      and NO. Decomposition of PAN is expected to be a relatively fast  process when indoor O3
24      levels are low and when motor vehicle emissions are large or there  is an indoor source of NOX
25      (Weschler and Shields, 1997).
26
27      Factors Affecting the Relationship between Ambient Concentrations and
28      Personal Exposures to O3
29           Ambient O3 concentrations vary with the time of day, season  of the year, and among
30      locations. Personal exposure to O3 is influenced by the microenvironmental concentration and
31      the amount of time spent in each microenvironment. Because the majority of the population
32      spends on average nearly 90% of their time in an indoor microenvironment, the majority of the

        August 2005                              3-67       DRAFT-DO NOT QUOTE OR CITE

-------
 1      O3 exposure will occur in the indoor environment.  Since there are few indoor sources of O3, O3
 2      ambient concentration may be the most important factor that affects average population exposure
 3      in the indoor environment.
 4           Indoor O3 concentrations also are affected by several other factors and mechanisms.
 5      Studies have shown that in addition to the ambient O3 concentrations, indoor O3 concentrations
 6      are influenced by the air exchange rate or outdoor infiltration, increasing with increasing air
 7      exchange.  Once indoors, the O3 concentration is affected by the indoor circulation rate and O3
 8      removal through contact with indoor surfaces and reactions with other indoor pollutants.
 9           In some instances, ambient O3 monitors are located in areas outside the breathing zone.
10      Studies on the effect of elevation on O3 concentrations found that concentrations increased with
11      increasing elevation (Vakeva et al., 1999; Johnson, 1997). Also, since O3 monitors are
12      frequently located on rooftops in urban settings, the concentrations measured there may
13      overestimate the exposure to individuals outdoors in streets and parks, locations where people
14      exercise and their maximum O3 exposure is more likely to occur.
15           In epidemiologic studies investigating acute and chronic health outcomes using ambient
16      monitoring data from  stationary monitoring sites, O3 exposure assessment was affected by the
17      distance between home and the monitoring site, gender, time-activity patterns (e.g., percentage
18      of time spent outdoors, type of outdoor activity, time of day during outdoor activity), and indoor
19      air exchange rates (e.g., ventilation conditions, home characteristics) (Geyh et al., 2000; Lee
20      et al., 2002, 2004;  Liu et al., 1995, 1997; Chang et al., 2000; Chan et al., 2005; O'Neill et al.,
21      2003; Brauer and Brook,  1997; O'Neill et al., 2003). People that work outdoors tend to be
22      exposed to higher levels of O3 (Brauer and Brook,  1997; O'Neill et al., 2003). Geyh et al. (2000)
23      observed higher indoor and personal O3 concentrations in a southern California community with
24      2% air-conditioned homes compared to a community with 93% air-conditioned homes during
25      the summer (high O3) months, but showed no difference in O3 levels during the winter (low O3)
26      months. Lee et al. (2004) observed that personal O3 exposure was positively correlated with
27      outdoor time (r = 0.19, p < 0.01)  and negatively correlated with indoor time (r = -0.17,
28      p < 0.01).  Additional factors that affected indoor O3 levels were air conditioning, window fans,
29      and window opening.  The O3 exposure assessment study by Liu et al. (1995) found that after
30      adjusting for time spent in various indoor and outdoor microenvironments (e.g., car with
31      windows open, car with windows closed, school, work, home, outdoors near home, outdoors

        August 2005                              3-68        DRAFT-DO NOT QUOTE OR CITE

-------
 1      other than near home), mean 12-hour ambient O3 concentrations explained 32% of the variance
 2      in personal exposure in the summer.
 3           In a southern California study by Avol et al. (1998b), boys were found to spend more time
 4      outdoors and be more physically active than girls. Another southern California study found that
 5      boys were outdoors 30 minutes longer than girls, and had higher personal O3 exposure during
 6      both high and low O3 months (Geyh et al., 2000).
 7           The announcement of smog alerts or air quality indices may influence personal exposures
 8      to O3 by causing individuals to alter behaviors (avoidance behavior).  Neidell (2004), in his
 9      evaluation of the effect of pollution on childhood asthma, examined the relationship between the
10      issuance of smog alerts or air quality indices for several counties in California and hospital
11      admissions for asthma in children under age 18 years (not including newborns).  Smog alerts are
12      issued in California on days when O3 concentrations exceed 200 ppb. There was a significant
13      reduction in the number of asthma-related hospital admissions in children ages 1 to 12 years on
14      smog alert days, indicating that avoidance behavior might be present on days of high O3
15      concentrations. Changes in population behavior as a function of concentration complicate the
16      estimation of health effects from population-based studies; thus, it may be desirable to include
17      sensitivity analyses that eliminate high O3 days, particularly in areas where avoidance behavior
18      is expected.
19
20      Potential Sources of Error Resulting from the Use of Ambient Ozone Concentrations in
21      Epidemiological A nalyses
22           There is no clear consensus among exposure analysts as to how well stationary monitor
23      measurements of ambient O3 concentrations represent a surrogate for personal O3 exposure.  The
24      approaches available for assessing exposure in air pollution epidemiology studies, the
25      microenvironmental (indirect) approach and the personal sampling (direct) approach (Navidi
26      et al., 1999;  Ott, 1982, 1985), are associated with measurement error. To determine personal
27      exposure using the microenvironmental approach, the concentrations of the various
28      microenvironments are multiplied by the time spent in each microenvironment. Both the
29      concentration and time component contribute to the measurement error.  There is no time
30      component to the measurement error in the personal  sampling approach, however, the  estimation
31      of exposure  using personal monitoring devices contributes to measurement error, especially in
32      the case of O3. Passive badges are commonly used for monitoring O3 integrated personal

        August 2005                              3-69        DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposure. Their sensitivity to wind velocity, badge placement, and interference with other
 2      copollutants may result in measurement error.
 3           Results from the error analysis models developed by Navidi et al. (1999) indicated that
 4      neither the microenvironmental nor personal sampling approach gave reliable health effect
 5      estimates when measurement errors were uncorrected. The nondifferential measurement error
 6      biased the effect estimates toward zero under the model assumptions.  However, if the
 7      measurement error was correlated with the health response, a bias away from the null could
 8      result. The use of central ambient monitors to estimate exposure also biased the estimates
 9      toward the null.  Since most people spend the majority of their time indoors, where O3 levels
10      tend to be much lower than outdoor ambient levels, using ambient concentrations to determine
11      exposure generally overestimates true personal O3 exposure, resulting in effect estimates biased
12      toward the null.
13           Several studies have examined the relationship between measured ambient O3
14      concentrations from fixed monitoring sites and personal O3 exposure (Avol et al., 1998a; Brauer
15      and Brook, 1995, 1997; Chang et al., 2000; Delfmo et al., 1996;  Lee et al., 2004;  Liard et al.,
16      1999; Linn et al., 1996; Liu et al., 1995, 1997; O'Neill et al., 2003; Sarnat et al., 2001).  In a
17      Baltimore, MD study of older adults, individuals with COPD, and  children, 24-h  average
18      ambient O3 concentrations from a monitoring site were not found to be significantly associated
19      with personal O3 exposure (Sarnat et al., 2001). The mixed regression effect estimates were
20      P = 0.01 (t = 1.21) and P =  0.00 (t =  0.03), for  summer and winter, respectively. Chang et al.
21      (2000) compared one-hour personal  and ambient O3 measurements in older adults in various
22      microenvironments using activity data from the National Human Activity Pattern Survey study
23      (Klepeis, 1999). There was no correlation between personal and ambient O3 concentrations in
24      the indoor residence  (r = 0.09 and r = 0.05, for summer and winter, respectively), although a
25      moderate correlation was found in other indoor environments  such as restaurants, hospitals, and
26      shopping malls (r = 0.34 in summer, r = 0.46 in winter).  In comparison, the correlation  in
27      outdoor environments (near and away from roads) was moderate to high (0.68 < r < 0.91) and
28      statistically significant. Slopes for the relationship between personal and ambient O3
29      concentrations were not reported in this study.
30           Brauer and Brook (1995, 1997) observed that the daily averaged personal O3 measurements
31      and ambient concentrations were well-correlated after stratifying groups by time spent outdoors.
32      Clinic workers (n = 25; 24-hour samples), teenage camp counselors (n = 25; 24- hour samples),
        August 2005                               3-70        DRAFT-DO NOT QUOTE  OR CITE

-------
 1      and farm workers (n = 15; 6-14 h work shift samples) spent 0 to 25%, 7.5 to 45%, and 100% of
 2      their monitored time outdoors, respectively. The personal to ambient O3 concentration ratios
 3      were significantly different for the clinic workers (0.28) and farm workers (0.96).  Ambient O3
 4      concentrations and time spent outdoors explained more of the variability in the personal O3
 5      measurements for outdoor farm workers compared to the clinical workers. However, the
 6      Spearman correlation coefficients were comparable, 0.60 and 0.64 for the clinic workers and
 7      farm workers, respectively, indicating that the variability of nonambient O3 exposures was
 8      similar in the two groups.  A study by O'Neill et al. (2003) examined 107 pairs of ambient and
 9      personal O3 measurements from 39 outdoor workers in Mexico City using a longitudinal analysis
10      method.  Two to seven personal measurements were collected on each of the 26 monitoring
11      days, which were averaged then compared with the ambient concentrations. They estimated that
12      a 1 ppb increase in ambient O3 concentration was associated with a 0.56 ppb (95% CI:  0.43,
13      0.69) increase in personal O3 concentration. In a Paris,  France study by Liard et al. (1999),
14      adults  (n = 55) and children (n = 39) wore passive O3 monitors for 4 consecutive days during
15      three periods.  For each period, all adults wore the O3 monitors over the same 4 days. Likewise,
16      all  children wore monitors over the same 4 days for each of the three periods,  but on different
17      days from the adults. The ambient O3 concentrations from the stationary monitoring sites did not
18      explain a high percentage of the variance of personal O3 exposure (nonsignificant  [value not
19      stated] in adults and 21% in children).  However, when personal measurements from all subjects
20      were aggregated for each of the six periods, the 4-day mean personal O3 exposure  was found to
21      be  highly correlated with the corresponding mean ambient concentration (r = 0.83, p < 0.05).
22      Similarly, a study of Los Angeles school children by Linn et al. (1996) found  that  daily 24-h
23      average ambient O3 concentrations from a central site were well-correlated (r  = 0.61) with daily
24      averaged personal O3 exposures.
25          The low correlation observed between personal O3 exposures and ambient O3
26      concentrations in the study by Sarnat et al. (2001) suggests that O3 concentrations  measured at
27      central ambient monitors do not explain the variance of individual personal exposures.
28      However, daily averaged personal exposures from the aggregate population have been found to
29      be  correlated with monitored ambient O3 concentrations, which is of greater relevance in time-
30      series  studies. Although there are correlations between aggregate personal and monitored
31      ambient O3 concentrations, the absolute personal concentrations may be considerably lower than
32      the monitored ambient O3 concentrations.
        August 2005                              3-71        DRAFT-DO NOT QUOTE OR CITE

-------
 1           In summary, results indicate that the relationship between ambient O3 concentrations and
 2     personal exposure will vary depending on factors such as O3 concentrations, time spent in the
 3     various microenvironments, and activity levels, creating potential measurement errors. The
 4     expectations based on statistical modeling considerations are that these exposure measurement
 5     errors or uncertainties will reduce the statistical power of the O3 health effects analysis, making
 6     it difficult to detect a true underlying association between the correct exposure metric and the
 7     health outcome studied.  However, until more data on O3 exposure become available, the use of
 8     monitored ambient O3 concentrations as a surrogate for exposures is not expected to change the
 9     principal conclusions from O3 epidemiologic studies using community average health and
10     pollution data.
11
12     Exposure to Related Photochemical Oxidants
13           A variety of related photochemical oxidants produced outdoors, such as PAN and
14     peroxypropionyl nitrate (PPN), can infiltrate into indoor environments. These compounds are
15     thermally unstable and decompose to peroxacetyl radicals and NO2. Exposure to related
16     photochemical oxidants has not been measured, nor are these compounds routinely monitored at
17     stationary monitoring sites. Available monitored concentrations of related photochemical
18     oxidants may be found in Annex AX3.
19
20
21     3.9 SUMMARY OF KEY POINTS
22           The median of the daily maximum 8-h O3 concentration averaged over May to September
23     is about 0.049 ppm from 2000 to 2004.  The daily maximum 1-h O3 concentrations could have
24     been much higher in large urban areas or in areas downwind of large urban areas.  For example,
25     in Houston, TX, the daily maximum 1-h O3 concentrations have approached 0.20 ppm during
26     this period.
27           Daily maximum 8-h average O3 concentrations are lower than the maximum 1-h O3
28     concentrations, but they are highly correlated. Within individual MSAs, O3 concentrations tend
29     to be well correlated across monitoring sites. However, there can be substantial variations in
30     O3  concentrations.  Ozone in city centers tends to be lower than in regions either upwind or
31     downwind because of titration by NO emitted by motor vehicles.

       August 2005                               3-72        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Ozone concentrations tend to peak in early- to mid-afternoon in areas where there is strong
 2      photochemical activity and later in the day in areas where transport is more important in
 3      determining the O3 abundance. Summertime maxima in O3 concentrations occur in areas in the
 4      United States where there is substantial photochemical activity involving O3 precursors emitted
 5      from human activities. Monthly maxima can occur anytime from June through August.
 6      However, springtime maxima are observed in national parks, mainly in the western United States
 7      and at a number of other relatively unpolluted monitoring sites throughout the Northern
 8      Hemisphere. For example, the highest O3 concentrations at Yellowstone NP tend to occur
 9      during April and May. Generally, monthly minima O3 concentrations tend to occur from
10      November through February at polluted sites and during the fall at relatively remote sites.
11           Nationwide, daily maximum 8-h O3 concentrations have decreased at the upper end of the
12      distribution from  1990 to 2004.  However, the daily maximum 8-h O3 concentrations toward the
13      center of the distribution have not reflected these changes. Trends have not been consistent at
14      national park sites; with downward trends observed at some sites and upward or no trends
15      observed at others. At some sites, trends reversed direction in going from the 98th to the 95th
16      percentile values.
17           Sufficient data are not available for other atmospheric oxidants (e.g., H2O2, PAN) and
18      oxidation products (e.g., HNO3, H2SO4) to relate concentrations of O3 to these species for use in
19      time series studies. Data for these species are only obtained as part of specialized field studies.
20      In general, secondary  species, such as HNO3, H2SO4, H2O2, and PAN, are expected to be at least
21      moderately correlated with O3. On the other hand, primary species are expected to be more
22      highly correlated with each other than with secondary species, provided that the primary species
23      originate from common sources.  Concentrations of other oxidants are much lower than for O3
24      and range from < 1% for oxidants in particles to several  percent for gas phase species. The
25      relationship of O3 to PM25 is complex, because PM is not a distinct chemical species but is a mix
26      of primary and secondary species. PM25 concentrations were positively  correlated with O3
27      during summer, but negatively correlated with O3 during winter at Ft. Meade, MD.  PM10
28      concentrations show similar relations with O3.
29           Co-occurrences of O3 (defined when both pollutants are present at an hourly average
30      concentration of >0.05 ppm) with NO2 and SO2 are rare. For example, there were fewer than
31      10  co-occurrences with either NO2 or SO2 in 2001. The number of co-occurrences for O3 and
32      PM2 5 (defined as  an hourly average O3 concentration >0.05 ppm and a 24-h average PM25
        August 2005                              3-73       DRAFT-DO NOT QUOTE OR CITE

-------
 1      concentration >40 |ig/m3 occurring during the same 24-h period) also tended to be infrequent
 2      (<10 times) at most sites, but there were up to 20 such co-occurrences at a few sites.
 3          Policy relevant background O3 concentrations are used for assessing risks to human health
 4      associated with O3 produced from anthropogenic sources in continental North America. Because
 5      of the nature of the definition of PRB concentrations, they cannot be directly derived from
 6      monitored concentrations, instead they must be derived from modeled estimates.  Current model
 7      estimates indicate that ambient air PRB concentrations in the United States are generally
 8      0.015 ppm to 0.035 ppm. They decline from spring to summer and are generally <0.025 ppm
 9      under conditions conducive to high O3 episodes.  However, PRB concentrations can be higher,
10      especially at elevated sites during spring, due to enhanced contributions from hemispheric
11      pollution and stratospheric exchange.
12          Ozone exposure changes as a function of time of day, season, and microenvironment.
13      Ambient O3 concentrations are generally higher during warmer seasons and during the weekday,
14      peaking during the later portion of the day. Ozone concentrations in indoor microenvironments
15      are generally lower than those concentrations encountered in the ambient air.  There are few
16      indoor sources of O3.  Ozone occurs in indoor microenvironments primarily through infiltration
17      through the building envelop and through windows, doors, and ventilation systems. The indoor
18      O3 concentration is dependent on the outdoor concentration, the AER, indoor circulation rate,
19      and removal processes.  Consequently, measured and modeled exposures should take into
20      consideration O3 diurnal weekly and seasonal variability and varying microenvironmental
21      concentrations.
22          Once indoors, O3 reacts with indoor surfaces, including surface coverings and furnishings.
23      Ozone also will react with VOCs in indoor environments, primarily terpenes or terpene-related
24      compounds.  Ozone reactions with pollutants indoors are analogous to those reactions occurring
25      in the ambient air, potentially exposing subjects to compounds significantly more toxic than O3.
26      The reaction products include Criegee biradicals, nitrate radicals, peroxyacetyl radicals, and
27      hydroxy, alkyl, alkyperoxy, hydroperoxy, and alkoxy radicals.  The hydroxy radical will react
28      with various nitrogen compounds, sulfur dioxide, carbon monoxide, and other compounds.  The
29      formation of submicron particles has been attributed to the reaction of O3 and the hydroxy
30      radical with terpene and terpene-related compounds.
31          The available approaches for measuring personal O3 exposure include the direct approach,
32      using a PEM, and the indirect approach, which measures or models exposure in the
        August 2005                              3-74       DRAFT-DO NOT QUOTE OR CITE

-------
 1      microenvironments the individual encounters. Both approaches are associated with
 2      measurement errors.
 3           There are difficulties in identifying chemical trapping agents for PEMs that can react with
 4      O3, and PEMs are sensitive to wind velocity, badge placement, and interference with other
 5      copollutants. Some studies using PEMs have shown personal O3 exposures below those
 6      concentrations measured at stationary monitoring sites, while other studies have found strong
 7      correlations between O3 measured at stationary monitoring sites and personal monitored
 8      concentrations.
 9           The use of measured O3 concentrations from stationary ambient monitoring sites as
10      surrogates for personal exposure may be affected by the O3 ambient concentration,  percentage of
11      time spent outdoors, and type of outdoor activity. Epidemiologic studies investigating health
12      outcomes using data from stationary monitoring sites found O3 exposure to be affected by the
13      distance between the subjects' location and the stationary monitor, individual activity patterns,
14      and the O3 concentration in the microenvironment.
15           The use of exposure models to evaluate O3 exposure to large populations over time is
16      complicated by the fact that O3 is a secondary pollutant with complex nonlinear and multiscale
17      dynamics in space and time.  The existing comprehensive inhalation exposure models (NEM,
18      pNEM, MENTOR/SHEDS, TREVI.Expo) treat human activity patterns as sequences of exposure
19      events. Estimates of activity levels are assigned from CHAD, the Consolidated Human
20      Activities Database.
21           Ambient O3 concentrations are  estimated using emissions-based mechanistic models or
22      ambient-data-based models. Models for estimating microenvironmental concentrations include
23      the empirical, mass balance, and detailed CFD models. Mass balance modeling is the most
24      common modeling approach to estimating concentrations in enclosed microenvironments.  The
25      pNEM/O3 population exposure model, the model used more extensively in O3 exposure
26      modeling, includes a sophisticated mass balance model for indoor and vehicle
27      microenvironments.  There are three versions of the pNEM/O3 model:   the general population,
28      outdoor workers, and outdoor children.
29           Results from O3 exposure studies indicate that the relationship between ambient O3
30      concentrations and personal exposure/dose will vary depending on O3 concentrations and time
31      spent in the various microenvironments, particularly the time spent outdoors where O3
32      concentrations tend to be higher, and the personal activity  level.  Consequently, the O3
        August 2005                             3-75        DRAFT-DO NOT QUOTE OR CITE

-------
1     exposure/dose may differ from the concentrations measured at stationary monitoring sites.
2     However, until more data on O3 exposure become available, the use of monitored ambient O3
3     concentrations as a surrogate for exposures is not expected to change the principal conclusions
4     from O3 epidemiologic studies using community average health and pollution data.
      August 2005                              3-76        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES

  2       Adams, W. C. (1993) Measurement of breathing rate and volume in routinely performed daily activities [final
  3             report]. Sacramento, CA: California Environmental Protection Agency, Air Resources Board; contract no.
  4             A033-205.
  5       Altshuller, A. P.; Lefohn, A. S. (1996) Background ozone in the planetary boundary layer over the United States.
  6             J. Air Waste Manage. Assoc. 46: 134-141.
  7       Avol, E. L.; Navidi, W. C.; Colome, S. D. (1998a) Modeling ozone levels in and around southern California homes.
  8             Environ. Sci. Technol. 32: 463-468.
  9       Avol, E. L.; Navidi, W. C.; Rappaport, E. B.; Peters, J. M. (1998b) Acute effects of ambient ozone on asthmatic,
10             wheezy, and healthy children. Cambridge, MA: Health Effects Institute; research report no. 82.
11       Bernard, N. L.; Gerber, M. J.; Astre, C. M.; Saintot, M.  J. (1999) Ozone measurement with passive samplers:
12             validation and use for ozone pollution assessment in Montpellier, France. Environ. Sci. Technol. 33:
13              217-222.
14       Black, D. R.; Harley, R. A.; Hering, S. V.; Stolzenburg, M. R.  (2000) A new, portable, real-time monitor. Environ.
15             Sci. Technol. 34: 3031-3040.
16       Blondeau, P.; lordache, V.; Poupard, O.; Genin, D.; Allard, F.  (2005) Relationship between outdoor and indoor air
17             quality in eight French schools. Indoor Air 15: 2-12.
18       Brauer, M.; Brook, J. R. (1995) Personal and fixed-site ozone measurements with a passive sampler. J. Air Waste
19             Manage. Assoc. 45: 529-537.
20       Brauer, M.; Brook, J. R. (1997) Ozone personal exposures and health effects for selected groups residing  in the
21             Fraser Valley. In: Steyn, D. G.; Bottenheim, J. W., eds.  The Lower Fraser Valley Oxidants/Pacific  '93 Field
22             Study. Atmos. Environ. 31: 2113-2121.
23       Burke, J. M.; Zufall, M. J.; Ozkaynak, H. (2001) A population exposure model for paniculate matter: case study
24             results for PM2 5 in Philadelphia, PA. J. Exposure Anal.  Environ. Epidemiol. 11: 470-489.
25       Chan, C.-C.; Wu, T.-H. (2005) Effects of ambient ozone exposure on mail carriers' peak expiratory flow rates.
26             Environ. Health Perspect.  113: 735-738.
27       Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H. H. (2000) Hourly personal exposures to fine particles and
28             gaseous pollutants—results from Baltimore,  Maryland.  J. Air Waste Manage. Assoc. 50: 1223-1235.
29       Chen, L.-W. A. (2002) Urban fine paniculate matter: chemical composition and possible origins (dissertation).
30             College Park, MD:  University of Maryland, Department of Chemical Physics. Available from: University
31             Microfilms, Ann Arbor, MI; AADAA-I3078297.
32       Christakos, G.; Vyas, V. M. (1998a) A composite space/time approach to studying ozone distribution over eastern
33             United States. Atmos. Environ. 32: 2845-2857.
34       Christakos, G.; Vyas, V. M. (1998b) A novel method for studying population health impacts of spatiotemporal
35             ozone distribution.  Soc. Sci. Med. 47: 1051-1066.
36       Clausen, P. A.; Wilkins, C. K.; Wolkoff, P.; Nielsen, G. D. (2001) Chemical and biological evaluation of a reaction
37             mixture of R-(+)-limonene/ozone: formation of strong airway irritants. Environ. Int. 26: 511-522.
3 8       Code of Federal Regulations. (2000) Appendix D to part 58—Network design for state and local air monitoring
39             stations (SLAMS),  national air monitoring stations (NAMS), and photochemical assessment monitoring
40             stations (PAMS). C. F. R.  40: pt. 58, app. D.
41       Colome, S. D.; Wilson, A. L.; Tian, Y. (1994) California residential indoor air quality study. Volume 2. Carbon
42             monoxide and air exchange rate: an univariate and multivariate analysis. Chicago, IL: Gas Research Institute;
43             report no. GRI-93/0224.3.
44       Cooper, O. R.; Moody, J. L. (2000) Meteorological controls on ozone at an elevated eastern United States regional
45             background monitoring site. J. Geophys.  Res. [Atmos.]  105: 6855-6869.
46       Delfino, R. J.; Coate, B. D.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Koutrakis, P. (1996) Daily asthma severity in
47             relation to personal ozone  exposure and outdoor fungal  spores. Am. J. Respir. Crit. Care Med. 154: 633-641.
48       Drakou, G.; Zerefos, C.; Ziomas, I. (1995) A preliminary study on the relationship between outdoor and indoor air
49             pollution levels. Fresenius' Environ. Bull. 4:  689-694.
50       Ebel, A.; Hass, H.; Jakobs, J. H.; Laube, M.; Memmesheimer,  M.; Oberreuter, A.  (1991) Simulation of ozone
51             intrusion caused by a tropopause fold and cut-off low. Atmos. Environ. Part A25: 2131-2144.
52       Fan, Z.; Lioy, P.; Weschler, C.; Fiedler, N.; Kipen,  H.; Zhang, J. (2003) Ozone-initiated reactions with mixtures of
53             volatile organic compounds under simulated indoor conditions. Environ. Sci.  Technol. 37: 1811-1821.
54       Federal Register. (1986) Guidelines for estimating exposures. F. R. (September 24) 51:  34,042-34,054.
55       Pick, J.; Pommer, L.; Nilsson, C.; Andersson, B. (2003) Effect of OH radicals, relative humidity, and time on the
56             composition of the products formed in the ozonolysis of a-pinene. Atmos. Environ. 37: 4087-4096.


         August 2005                                     3-77         DRAFT-DO  NOT QUOTE  OR CITE

-------
  1       Pick, I; Nilsson, C.; Andersson, B. (2004) Formation of oxidation products in a ventilation system. Atmos. Environ.
  2             38: 5895-5899.
  3       Fiore, A. M; Jacob, D. I; Bey, I.; Yantosca, R. M; Field, B. D.; Fusco, A. C.; Wilkinson, J. G. (2002) Background
  4             ozone over the United States in summer: origin, trend, and contribution to pollution episodes. J. Geophys.
  5             Res. (Atmos.) 107(D15):  10.1029/2001JD000982.
  6       Fiore, A.; Jacob, D. J.; Liu, H.; Yantosca, R. M.; Fairlie, T. D.; Li, Q. (2003) Variability in surface ozone
  7             background over the United States: implications for air quality policy. J. Geophys. Res. (Atmos.)
  8             108(024): 10.1029/2003 JD003855.
  9       Fitz-Simons, T.; McCluney, L.; Rizzo, M. (2005) Analysis of 2004 ozone data for the ozone NAAQS review
10             [memorandum to Dr. Joseph Pinto]. Research Triangle Park, NC: U.S. Environmental Protection Agency;
11             August 22.
12       Freijer, J. I.; Bloemen, H. J. T. (2000) Modeling relationships between indoor and outdoor air quality. J. Air Waste
13             Manage. Assoc.  50: 292-300.
14       Fusco, A. C.; Logan, J. A. (2003) Analysis of 1970-1995 trends in tropospheric ozone at Northern Hemisphere
15             midlatitudes with the GEOS-CHEM model. J. Geophys. Res. (Atmos.) 108:  10.1029/2002JD002742.
16       Garratt, J. R. (1992) The atmospheric boundary layer. Cambridge, United Kingdom: Cambridge University Press.
17             (Houghton,  J. T.; Rycroft, M. J.; Dessler, A. J., eds. Cambridge atmospheric and space science series).
18       Georgopoulos, P. G.; Arunachalam, S.; Wang, S. (1997a) Alternative metrics for assessing the relative effectiveness
19             of NOX and  VOC emission reductions in controlling ground-level ozone. J. Air Waste Manage. Assoc.
20             47:838-850.
21       Georgopoulos, P. G.; Walia, A.; Roy, A.; Lioy, P. J. (1997b) Integrated exposure and dose modeling and analysis
22             system. 1. Formulation and testing of microenvironmental and pharmacokinetic components. Environ. Sci.
23             Technol. 31: 17-27.
24       Geyh, A. S.; Wolfson, J. M.; Koutrakis, P.; Mulik, J. D.; Avol, E. L. (1997) Development and evaluation of a small
25             active ozone sampler. Environ. Sci. Technol. 31: 2326-2330.
26       Geyh, A. S.; Roberts, P. T.;  Lurmann, F. W.; Schoell, B. M.; Avol, E. L. (1999) Initial field evaluation of the
27             Harvard active ozone sampler for personal ozone monitoring. J. Exposure Anal. Environ. Epidemiol. 9:
28             143-149.
29       Geyh, A. S.; Xue, J.; Ozkaynak, H.; Spengler, J. D. (2000) The Harvard Southern California chronic ozone exposure
30             study: assessing ozone exposure of grade-school-age  children in two southern California communities.
31             Environ. Health Perspect. 108: 265-270.
3 2       Glen,  G.; Lakkadi, Y.; Tippett, J. A.; del Valle-Torres, M. (1997) Development of NERL/CHAD:  the National
33             Exposure Research Laboratory consolidated human activity database. Research Triangle Park, NC: U.S.
34             Environmental Protection Agency, Office of Research and Development; contract no. 68-D5-0049.
35       Gold,  D. R.; Allen, G.;  Damokosh, A.; Serrano, P.; Hayes, C.; Castillejos, M. (1996) Comparison  of outdoor and
36             classroom ozone exposures for school children in Mexico City. J. Air Waste Manage. Assoc. 46: 335-342.
37       Gonzales, M.; Ngo, L.; Hammond, S. K.; Tager, I. (2003) Validation of a questionnaire and microenvironmental
3 8             model for estimating past exposures to ozone. Int. J. Environ. Health Res. 13: 249-260.
39       Grosjean, D. (2003) Ambient PAN and PPN in southern California from 1960  to the SCOS97-NARSTO. Atmos.
40             Environ. 37(suppl. 2): S221-S238.
41       Grosjean, E.; Grosjean, D.; Fraser, M. P.; Cass, G. R. (1996) Air quality model evaluation data for organics: 3.
42             Peroxyacetyl nitrate and peroxypropionyl nitrate in Los Angeles air Environ. Sci. Technol.  30: 2704-2714.
43       Heuss, J. M.; Kahlbaum, D. F.; Wolff, G. T. (2003) Weekday/weekend ozone differences: what can we learn from
44             them? J. Air Waste Manage. Assoc. 53: 772-788.
45       Hirsch, A. I.; Munger, J. W.; Jacob, D. J.; Horowitz, L. W.;  Goldstein, A. H. (1996) Seasonal variation of the ozone
46             production efficiency per unit NOx at Harvard Forest, Massachusetts. J. Geophys. Res. [Atmos.]
47             101: 12,659-12,666.
48       Horowitz, L. W.; Liang, J.; Gardner,  G. M.; Jacob, D. J . (1998) Export of reactive nitrogen from North America
49             during summertime: sensitivity to hydrocarbon chemistry. J. Geophys. Res. [Atmos.]
50             103(D11): 13451-13476.
51       Horvath, L.; Nagy, Z.; Weidinger, T.; Artz, R.;  Luke, W. T.; Valigura, R.; Pinto, J. P.; Womack, J. (1995)
52             Measurement effluxes of trace gases (O3, NOX, SO2,  CO2, HNO3),  paniculate sulfate and nitrate, water
53             vapour over short vegetation by gradient and eddy correlation techniques in Hungary. EGS  XX. General
54             Assembly; April; Hamburg, Germany. Ann. Geophys. 13(suppl. 2): C490.
55       Horvath, L.; Pinto, J.; Weidinger, T. (2003) Estimate of the dry deposition of atmospheric nitrogen and sulfur
56             species to spruce forest. Idojaras (Q. J. Hung. Meteorol. Serv.) 107: 249-255.
         August 2005                                    3-78         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Howard-Reed, C.; Wallace, L. A.; Ott, W. R. (2002) The effect of opening windows on air change rates in two
  2             homes. J. Air Waste Manage Assoc. 52: 147-159.
  3       Hudman, R. C.;  Jacob, D. J.; Cooper, O. C.; Evans, M. J.; Heald, C. L.; Park, R. J.; Fehsenfeld, F.; Flocke, F.;
  4             Holloway, J.; Hubler, G.; Kita, K.; Koike, M.; Kondo, Y.; Neuman, A.; Nowak, J.; Oltmans, S.; Parrish, D.;
  5             Roberts, J. M.; Ryerson, T. (2004) Ozone production in transpacific Asian pollution plumes and implications
  6             for ozone air quality in California. J. Geophys. Res. (Atmos.) 109(D23): 10.1029/2004JD004974.
  7       Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and
  8             meta-analysis. Epidemiology  16: 446-457.
  9       Jaffe, D.; Price,  H.; Parrish, D.; Goldstein, A.; Harris, J. (2003) Increasing background ozone during spring on the
10             west coast of North America.  Geophys. Res. Lett.  30: 10.1029/2003GL017024.
11       Jakobi, G.; Fabian, P. (1997) Indoor/outdoor concentrations of ozone and peroxyacetyl nitrate (PAN). Int. J.
12             Biometeorol. 40: 162-165.
13       Johnson, T. R. (1995) Recent advances in the estimation of population exposure to mobile source pollutants. J.
14             Exposure Anal. Environ. Epidemiol. 5: 551-571.
15       Johnson, T. (1997) A pilot study in Los Angeles to measure personal ozone exposures during scripted activities.
16             Washington, DC: American Petroleum Institute, Health and Environmental Sciences Department;
17             API publication no. DR 218.
18       Johnson, T. (2003) A guide to selected algorithms, distributions, and databases used in exposure models developed
19             by the Office of Air Quality Planning and Standards. Research Triangle Park, NC: U.S. Environmental
20             Protection Agency, Office of Research and Development; EPA grant no. CR827033. Available:
21             http://www.epa.gov/ttn/fera/data/human/report052202.pdf [9 April, 2004].
22       Johnson, T.; Long, T. (2004) Determining the frequency of open windows in residences: a pilot study in Durham,
23             North Carolina during varying temperature conditions. J. Exposure  Anal. Environ. Epidemiol.:
24             10.1038/sj.jea.7500409.
25       Johnson, T.; Capel, J.; McCoy, M. (1996a) Estimation of ozone exposures experienced by urban residents using a
26             probabilistic version of NEM  and 1990 population data. Research Triangle Park, NC: U.S. Environmental
27             Protection Agency, Office of Air Quality Planning and Standards; contract no. 68-DO-0062.
28       Johnson, T.; Capel, J.; McCoy, M.; Mozier, J. W. (1996b) Estimation of ozone exposures experienced by outdoor
29             workers in nine urban areas using a probabilistic version of NEM. Research Triangle Park,  NC: U.S.
30             Environmental Protection Agency, Office of Air Quality Planning and Standards; contract no. 63-D-30094,
31             work assignment nos. 0-1 and 1-4.
32       Johnson, T.; Capel, J.; Mozier, J.; McCoy, M. (1996c) Estimation of ozone exposures experienced by outdoor
33             children in nine urban areas using a probabilistic version of NEM. Research Triangle Park,  NC: U.S.
34             Environmental Protection Agency, Office of Air Quality Planning and Standards; contract no. 63-D-30094.
35       Johnson, T.; Mozier, J.; Capel, J. (1997) Supplement to "Estimation of ozone exposures experienced by outdoor
36             children in nine urban areas using a probabilistic version of NEM (April 1996)". Research Triangle Park, NC:
37             U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards.
38       Johnson, T.; Mihlan, G.; LaPointe, J.; Fletcher, K.; Capel, J. (1999) Estimation of carbon monoxide exposures and
39             associated carboxyhemoglobin levels in Denver residents using pNEM/CO (version 2.0) [draft]. Research
40             Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning  and Standards;
41             March 15.
42       Johnson, T.; Myers, J.; Kelly, T.; Wisbith, A.; Ollison, W. (2004) A pilot study using scripted ventilation conditions
43             to identify key factors affecting indoor pollutant concentrations and air exchange rate in a residence. J.
44             Exposure Anal. Environ. Epidemiol. 14:  1-22.
45       Klepeis, N. E. (1999) An introduction to the indirect exposure assessment  approach: modeling human exposure
46             using microenvironmental measurements and the recent National Human Activity Pattern Survey. Environ.
47             Health Perspect. Suppl. 107(2):  365-374.
48       Klepeis, N. E.; Tsang, A. M.; Behar, J. V. (1996) Analysis of the national  human activity pattern survey (NHAPS)
49             respondents from a standpoint of exposure assessment. Washington, DC: U.S. Environmental Protection
50             Agency, Office of Research and Development; report no. EPA/600/R-96/074.
51       Klepeis, N. E.; Nelson, W. C.; Ott. W. R.; Robinson, J. P. Tsang, A. M.; Switzer, P.; Behar, J. V.;  Hern, S. C.;
52             Engelmann, W. H. (2001)  The National Human Activity Pattern Survey (NHAPS): a resource for assessing
53             exposure  to environmental pollutants. J. Exposure Anal. Environ. Epidemiol. 11: 231-252.
54       Kotchenruther, R. A.; Jaffe, D. A.; Beine, H. J.; et al. (2001a) Observations of ozone and related species in the
55             northeast Pacific  during the PHOBEA campaigns. 2. Airborne observations. J. Geophys. Res. (Atmos.)
56             106(07):  7463-7483.
         August 2005                                    3-79         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Kotchenruther, R. A.; Jaffe, D. A.; Jaegle, L. (200Ib) Ozone photochemistry and the role of peroxyacetyl nitrate in
  2             the springtime northeastern Pacific troposphere: results from the photochemical ozone budget of the eastern
  3             north Pacific atmosphere (PHOBEA) campaign. J. Geophys. Res. (Atmos.) 106(D22): 28731-28742.
  4       Koutrakis, P.; Sioutas, C.; Ferguson, S. T.; Wolfson, J. M; Mulik, J. D.; Burton, R. M. (1993) Development and
  5             evaluation of a glass honeycomb denuder/filter pack system to collect atmospheric gases and particles.
  6             Environ. Sci. Technol. 27: 2497-2501.
  7       Lagus Applied Technology, Inc. (1995) Air change rates in non-residential buildings in California. Sacramento, CA:
  8             California Energy Commission; contract no. 400-91-034; July.
  9       Lee, E. H.; Hogsett, W. E. (1999) Role of concentrations and time of day in developing ozone exposure indices for a
10             secondary standard. J. Air Waste Manage. Assoc. 49:  669-681.
11       Lee, M.; Heikes, B. G.; O'Sullivan, D. W. (2000) Hydrogen peroxide and organic hydroperoxide in the troposphere.
12             Atmos. Environ. 34: 3475-3494.
13       Lee, K.; Xue, J.; Geyh, A. S.; Ozkaynak, H.; Leaderer, B. P.; Weschler, C. J.; Spengler, J. D. (2002) Nitrous acid,
14             nitrogen dioxide, and ozone concentrations in residential environments. Environ. Health Perspect. 110:
15             145-150.
16       Lee, K.; Parkhurst, W. J.; Xue, J.; Ozkaynak, H.; Neuberg, D.; Spengler, J. D. (2004) Outdoor/indoor/personal ozone
17             exposures of children in Nashville, Tennessee. J. Air Waste Manage. Assoc. 54: 352-359.
18       Lefohn, A. S.; Runeckles, V. C. (1987) Establishing standards to protect vegetation - ozone exposure/dose
19             considerations. Atmos. Environ. 21:  561-568.
20       Lefohn, A. S.; Oltmans, S. J.; Dann, T.; Singh, H. B. (2001) Present-day variability of background ozone in the
21             lower troposphere. J. Geophys. Res.  [Atmos.] 106: 9945-9958.
22       Lehman, J.; Swinton, K.; Bortnick, S.; Hamilton, C.; Baldridge, E.; Ender, B.; Cox, B. (2004) Spatio-temporal
23             characterization of tropospheric ozone across the eastern United States. Atmos. Environ. 38: 4357-4369.
24       Leovic, K. W.; Sheldon, L. S.; Whitaker, D. A.; Hetes, R. G.; Calcagni, J. A.; Baskir, J. N. (1996) Measurement of
25             indoor air emissions from dry-process photocopy machines. J. Air Waste Manage. Assoc. 46: 821-829.
26       Leovic, K.; Whitaker, D.; Northeim, C.; Sheldon, L. (1998) Evaluation of a test method for measuring indoor air
27             emissions from dry-process photocopiers. J. Air Waste Manage. Assoc. 48: 915-923.
28       Levinson, D. H., ed. (2005) State of the climate in 2004. Bull. Am. Meteorol. Soc. 86: S1-S86.
29       Levinson, D. H.; Waple,  A. M., eds.  (2004) State of the climate in 2003. Bull. Am. Meteorol. Soc. 85: S1-S72.
30       Liang, J.; Horowitz, L. W.; Jacob, D. J.; Wang, Y.; Fiore, A. M.; Logan, J. A.; Gardner, G. M.; Munger, J.  W.
31             (1998) Seasonal budgets of reactive nitrogen species and ozone over the United States, and export fluxes to
32             the global atmosphere. J. Geophys. Res. (Atmos.) 103: 13,435-13,450.
33       Liard, R.; Zureik, M.; Le Moullec, Y.; Soussan, D.; Glorian, M.; Grimfeld, A.; Neukirch, F.  (1999) Use of personal
34             passive samplers for measurement of NO2, NO, and O3 levels in panel studies. Environ. Res. 81: 339-348.
35       Lin, C.-Y.; Jacob, D. J.; Munger, J. W.; Fiore, A. M. (2000) Increasing background ozone in surface air over the
36             United States. Geophys. Res. Lett. 27: 3465-3468.
37       Linn, W.  S.; Shamoo, D. A.; Hackney, J. D. (1992) Documentation of activity patterns in 'high-risk' groups exposed
38             to ozone in the Los Angeles area. In: Tropospheric ozone and the environment II: effects, modeling and
39             control: papers from an Air & Waste Management Association international specialty conference; November,
40             1991; Atlanta, GA. Pittsburgh, PA: Air & Waste Management Association; pp. 701-712. (A&WMA
41             publication TR-20).
42       Linn, W.  S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996)
43             Short-term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal.
44             Environ. Epidemiol. 6: 449-472.
45       Liu, L.-J. S.; Koutrakis, P.; Leech, J.; Broder, I. (1995) Assessment of ozone exposures in the greater metropolitan
46             Toronto area. J. Air Waste Manage. Assoc. 45: 223-234.
47       Liu, L.-J. S.; Delfino, R.; Koutrakis, P. (1997) Ozone exposure assessment in a southern California community.
48             Environ. Health Perspect. 105: 58-65.
49       McCurdy, T. (2000) Conceptual basis for multi-route intake dose modeling using an energy expenditure approach.
50             J. Exposure Anal.  Environ. Epidemiol.  10: 86-97.
51       McCurdy, T.; Glen, G.; Smith, L.; Lakkadi, Y. (2000) The National Exposure Research Laboratory's Consolidated
52             Human Activity Database. J. Exposure Anal. Environ. Epidemiol. 10: 566-578.
53       Meng, Q. Y.; Turpin, B. J.; Korn, L.; Weisel, C. P.; Morandi, M.; Colome, S.; Zhang, J.; Stock, T.; Spektor, D.;
54             Winer, A.; Zhang, L.; Lee, J. H.; Giovanetti, R.; Cui, W.; Kwon, J.; Alimokhtari, S.; Shendell, D.; Jones, J.;
55             Farrar, C.; Maberti, S. (2005) Influence of ambient (outdoor) sources on residential indoor and personal PM25
56             concentrations: analyses of RIOPA data. J. Exposure Anal. Environ. Epidemiol. 15: 17-28.
         August 2005                                    3-80         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Morrison, G. C.; Nazaroff, W. W. (2000) The rate of ozone uptake on carpets: experimental studies. Environ. Sci.
  2             Technol. 34: 4963-4968.
  3       Morrison, G. C.; Nazaroff, W. W. (2002) Ozone interactions with carpet: secondary emissions of aldehydes.
  4             Environ. Sci. Technol. 36: 2185-2192.
  5       Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
  6             parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
  7       Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
  8             effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
  9       Nazaroff, W. W.; Cass, G. R. (1986) Mathematical modeling of chemically reactive pollutants in indoor air.
10             Environ. Sci. Technol. 20: 924-934.
11       Nazaroff, W. W.; Weschler,  C. J. (2004) Cleaning products and air fresheners: exposure to primary and secondary
12             air pollutants. Atmos.  Environ. 38: 2841-2865.
13       Neidell, M. J. (2004) Air pollution, health, and socio-economic status: the effect of outdoor air quality on childhood
14             asthma. J. Health Econ. 23: 1209-1236.
15       Niu, J.; Tung, T. C. W.; Burnett, J.  (2001) Ozone emission rate testing and ranking method using environmental
16             chamber. Atmos. Environ. 35: 2143-2151.
17       Nejgaard, C.; Sondergaard, S. B.; Madsen, J. L.; Moller, S. (2005) [Does the administration of dipyridamole affect
18             the ventilatory capacity of patients with chronic obstructive lung disease?] Ugeskr. Laeg. 167:  1750-1753.
19       Northeast States for Coordinated Air Use Management (NESCAUM). (2002) Indoor/outdoor school air monitoring
20             project. Boston, MA. Available: http://www.nescaum.org/pdf/schoolmonitoring.pdf [29 October, 2003].
21       Oltmans, S. J.; Lefohn, A. S.; Scheel, H. E.; Harris, J. M.; Levy, H., II; Galbally, I. E.; Brunke, E.-G.; Meyer, C. P.;
22             Lathrop, J. A.; Johnson, B. J.; Shadwick, D. S.; Cuevas, E.; Schmidlin, F. J.; Tarasick, D. W.; Claude, H.;
23             Kerr, J. B.; Uchino, O.; Mohnen, V. (1998) Trends of ozone in the troposphere. Geophys. Res. Lett. 25:
24             139-142.
25       O'Neill, M. S.; Ramirez-Aguilar, M.; Meneses-Gonzalez, F.; Hernandez-Avila, M.; Geyh, A. S.; Sienra-Monge, J. J.;
26             Romieu, I. (2003) Ozone exposure among Mexico City outdoor workers. J. Air Waste Manage. Assoc.
27             53: 339-346.
28       Orzechowska, G. E.; Paulson, S. E. (2002) Production of OH radicals from the reactions of C4-C6 internal alkenes
29             and styrenes with ozone in the gas phase. Atmos. Environ. 36: 571-581.
30       Ott, W. R. (1982) Concepts of human exposure to  air pollution. Environ. Int. 7:  179-196.
31       Ott, W. R. (1985) Total human exposure: an emerging science focuses on humans as receptors of environmental
32             pollution. Environ. Sci. Technol. 19: 880-886.
33       Park, J.-H.; Spengler, J.  D.; Yoon, D.-W.; Dumyahn, T.; Lee, K.; Ozkaynak, H.  (1998) Measurement of air
34             exchange rate of stationary vehicles and estimation of in-vehicle exposure. J. Exposure Anal. Environ.
35             Epidemiol. 8: 65-78.
36       Pinto, J. P.; Lefohn, A. S.; Shadwick, D. S. (2004) Spatial variability of PM25 in urban areas in the United States.
37             J. Air Waste Manage.  Assoc. 54: 440-449.
38       Poupard, O.; Blondeau, P.; lordache, V.; Allard, F. (2005)  Statistical analysis of parameters influencing the
39             relationship between outdoor and indoor air quality  in schools. Atmos. Environ. 39: 2071-2080.
40       Reeves, C. E.; Penkett, S. A.  (2003) Measurements of peroxides and what they tell us. Chem. Rev.  103: 5199-5218.
41       Reiss, R.; Ryan, P. B.; Koutrakis, P.; Tibbetts, S. J. (1995) Ozone reactive chemistry on interior latex paint. Environ.
42             Sci. Technol. 29: 1906-1912.
43       Riediker, M.; Williams, R.; Devlin, R.; Griggs, T.; Bromberg, P. (2003) Exposure to paniculate matter, volatile
44             organic compounds, and other air pollutants inside patrol cars. Environ. Sci. Technol. 37: 2084-2093.
45       Roberts, J. M.; Williams, J.; Baumann, K.; Buhr, M. P.; Goldan, P. D.; Holloway, J.; Htibler, G.; Kuster, W. C.;
46             McKeen, S. A.; Ryerson, T. B.; Trainer, M.; Williams, E. J.; Fehsenfeld,  F. C.; Bertman, S. B.; Nouaime, G.;
47             Seaver, C.; Grodzinsky, G.; Rodgers,  M.; Young, V. L. (1998) Measurements of PAN, PPN, and MPAN
48             made during the 1994 and 1995 Nashville Intensives of the Southern Oxidant Study: implications for regional
49             ozone production from biogenic hydrocarbons. J. Geophys. Res. [Atmos.] 103: 22,473-22,490.
50       Roberts, J. M.; Flocke, F.; Stroud, C. A.; Hereid, D.; Williams, E.; Fehsenfeld, F.; Brune, W.; Martinez, M.; Harder,
51             H. (2002) Ground-based measurements of peroxycarboxylic nitric anhydrides (PANs) during the 1999
52             Southern Oxidants Study Nashville intensive. J. Geophys. Res. [Atmos.]  107(D21): 10.1029/2001JD000947.
53       Roberts, J. M.; Jobson, B. T.; Kuster, W.; Goldan, P.; Murphy, P.; Williams, E.; Frost, G.; Riemer, D.; Apel, E.;
54             Stroud, C.; Wiedinmyer,  C.;  Fehsenfeld, F.  (2003) An examination of the chemistry of peroxycarboxylic
55             nitric anhydrides  and related volatile organic compounds during Texas Air Quality Study 2000 using
56             ground-based measurements. J. Geophys. Res. [Atmos.] 108(D16):  10.1029/2003JD003383.
         August 2005                                    3-81         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Roberts, J. M; Flocke, F.; Chen, G.; de Gouw, I; Holloway, J. S.; Htibler, G.; Neuman, J. A.; Nicks, D. K., Jr.;
  2             Nowak, J. B.; Parrish, D. D.; Ryerson, T. B.; Sueper, D. T.; Warneke, C.; Fehsenfeld, F. C. (2004)
  3             Measurement of peroxycarboxylic nitric anhydrides (PANs) during the ITCT 2K2 aircraft intensive
  4             experiment. J. Geophys. Res. 109(D23S21): 10.1029/2004JD004960.
  5       Roelofs, G. J.; Scheeren, H. A.; Heland, J.; Ziereis, H.; Lelieveld, J. (2003) A model study of ozone in the eastern
  6             Mediterranean free troposphere during MINOS (August 2001). Atmos. Chem. Phys. 3: 1199-1210.
  7       Rohr, A. C.; Wilkins, C. K.; Clausen, P. A.; Hammer, M.; Nielsen, G. D.; Wolkoff, P.; Spengler, J. D. (2002) Upper
  8             airway and pulmonary effects of oxidation products of (+)-a-pinene, J-limonene, and isoprene in B ALB/c
  9             mice. Inhalation Toxicol. 14: 663-684.
10       Rohr, A. C.; Shore, S.  A.; Spengler, J. D. (2003) Repeated exposure to isoprene oxidation products causes enhanced
11             respiratory tract effects in multiple murine strains. Inhalation Toxicol. 15: 1191-1207.
12       Romieu, I.; Lugo, M. C.; Colome, S.; Garcia A. M.; Avila, M. H.; Geyh, A.; Velasco, S. R.; Rendon, E. P. (1998)
13             Evaluation of indoor ozone concentration and predictors of indoor-outdoor ratio in Mexico City. J. Air Waste
14             Manage. Assoc. 48: 327-335.
15       Salmon, L. G.; Cass, G. R.; Bruckman, K.; Haber, J. (2000) Ozone exposure inside museums in the historic central
16             district of Krakow, Poland. Atmos. Environ. 34: 3823-3832.
17       Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal paniculate and gaseous
18             exposures of senior citizens living in Baltimore, MD. J. Air Waste Manage. Assoc. 50: 1184-1198.
19       Sarnat, J. A.; Schwartz, J.; Catalano, P. J.; Suh, H. H. (2001) Gaseous pollutants in paniculate matter epidemiology:
20             confounders or surrogates? Environ. Health Perspect. 109: 1053-1061.
21       Sarwar, M.; Corsi, R.; Kimura, Y.; Allen, D.; Weschler, C. (2001) Hydroxyl radicals in indoor environments. In:
22             Proceedings of the Air & Waste Management Association's 94th Annual Conference & Exhibition; June;
23             Orlando, FL. Pittsburgh, PA: Air & Waste Management Association.
24       Sarwar, G.; Corsi, R.; Kumura, Y.; Allen, D.; Weschler, C. J. (2002) Hydroxyl radicals in indoor environments.
25             Atmos. Environ. 36: 3973-3988.
26       Serensen, D. N.; Weschler, C. J. (2002) Modeling-gas phase reactions in indoor environments using  computational
27             fluid dynamics. Atmos. Environ. 36: 9-18.
28       Sparks, J. P.; Roberts,  J. M.; Monson, R. K. (2003) The uptake of gaseous organic nitrogen by leaves: a significant
29             global nitrogen transfer process. Geophys. Res. Lett. 30(23): 10.1029/2003GL018578.
30       Spier, C. E.; Little, D.  E.; Trim, S.  C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992) Activity patterns in
31             elementary and high school  students exposed to oxidant pollution. J. Exposure Anal. Environ. Epidemiol.
32             2:277-293.
33       Steiber, R. S. (1995) Ozone generators in indoor air settings. Research Triangle Park, NC: U.S. Environmental
34             Protection Agency, National Risk Management Research Laboratory; report no. EPA-600/R-95-154.
35             Available from: NTIS, Springfield, VA; PB96-100201.
36       Teklemariam, T. A.; Sparks, J. P. (2004) Gaseous fluxes  of peroxyacetyl nitrate (PAN) into plant leaves. Ecol.  Soc.
37             Am. Ann. Meeting Abst. 89: 501-502.
38       Trainer, M.; Parrish, D. D.; Buhr, M. P.; Norton, R. B.; Fehsenfeld, F. C.; Anlauf,  K. G.; Bottenheim, J. W.; Tang,
39             Y. Z.; Wiebe, H. A.; Roberts, J. M.; Tanner, R. L.; Newman, L.;  Bowersox, V. C.; Meagher, J. F.; Olszyna,
40             K. J.; Rodgers, M. O.; Wang, T.; Berresheim, H.; Demerjian, K. L.; Roychowdhury, U. K. (1993) Correlation
41             of ozone with NOy in photochemically aged air. J. Geophys. Res. [Atmos.] 98: 2917-2925.
42       Turk, B. H.; Grimsrud, D. T.; Brown, J. T.; Geisling-Sobotka, K. L.; Harrison, J.; Prill, R. J. (1989) Commercial
43             building ventilation rates and particle concentrations. ASHRAE Trans. 95(part 1): 422-433.
44       U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other photochemical oxidants.
45             Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
46             Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA;
47             PB87-142949.
48       U.S. Environmental Protection Agency. (1992) Guidelines for exposure assessment. Washington, DC: Risk
49             Assessment Forum, USEPA 600Z-92/001.
50       U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
51             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
52             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
53             http://cfpub2.epa.gov/ncea/.
54       U.S. Environmental Protection Agency. (1998) Guideline on ozone monitoring site selection. Research Triangle
55             Park, NC: U.S. Environmental Protection Agency, National Center for Environmental Assessment;
56             EPA-454/R-98-002.
         August 2005                                   3-82         DRAFT-DO NOT QUOTE OR CITE

-------
  1       U.S. Environmental Protection Agency. (2003a) Technology Transfer Network: Air Quality System (AQS).
  2             Washington, DC: Office of Air and Radiation. Available: http://www.epa.gov/ttn/airs/airsaqs/
  3             [24 August, 2005].
  4       U.S. Environmental Protection Agency. (2004a) Air quality criteria for paniculate matter. Research Triangle Park,
  5             NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
  6             http://cfpub.epa.gov/ncea/ [9 November, 2004].
  7       U.S. Environmental Protection Agency. (2004b) The ozone report: measuring progress through 2003. Research
  8             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/K04-001. Available:
  9             http://www.epa.gov/air/airtrends/pdfs/2003ozonereport.pdf [12 May, 2005].
10       Vakeva, M; Ha'meri, K.; Kulmala, M; Lahdes, R.; Ruuskanen, I; Laitinen, T. (1999) Street level versus rooftop
11             concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon. Atmos.
12             Environ. 33: 1385-1397.
13       Van den Bergh, V.; Vanhees, I.; De Boer, R.; Compernolle, F.; Vinckier, C. (2000) Identification of the oxidation
14             products of the reaction between a-pinene and hydroxyl radicals by gas and high-performance liquid
15             chromatography with mass spectrometric detection. J. Chromatogr. A 896:  135-148.
16       Venkatachari, P.; Hopke, P. K.; Grover, B. D.; Eatough, D. J. (2005) Measurement of particle-bound reactive
17             oxygen species inRubidoux aerosols. J. Atmos.  Chem. 50: 49-58.
18       Venkatachari, P.; Hopke, P .K.; Brune, W. H.; Ren, X.; Lesher, R.; Mao, J.; Mitchell, M. (2005) Characterization of
19             reactive oxygen species trends in Flushing, New York. Atmos. Environ.: in press.
20       Volz-Thomas, A.; Geiss, H.; Hofzumahaus, A.; Becker, K.-H. (2003) Introduction to special section: photochemistry
21             experiment in BERLIOZ. J. Geophys. Res. [Atmos.] 108(D4): 10.1029/JD002029.
22       Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of submicron
23             particle exposure. Environ. Health Perspect. 108: 1139-1145.
24       Wang, Y.; Logan, J. A.; Jacob, D. J.  (1998) Global simulation of tropospheric O3-NOx-hydrocarbon chemistry 2.
25             Model  evaluation and global ozone budget. J. Geophys. Res. (Atmos.) 103: 10,727-10,755.
26       Weinstein-Lloyd; et al. (1998)  Measurements of peroxides and related species during the 1995 summer intensive of
27             the Southern oxidants study in Nashville, Tennessee. J. Geophys. Res. (Atmos.) 103: 22361-22373.
28       Weschler, C. J. (2000) Ozone in indoor environments: concentration and chemistry. Indoor Air 10: 269-288.
29       Weschler, C. J. (2004) Chemical reactions among indoor pollutants: what we've learned in the new millennium.
30             Indoor Air 14(suppl. 7): 184-194.
31       Weschler, C. J.; Shields, H. C.  (1997) Potential reactions among indoor pollutants. Atmos. Environ. 31: 3487-3495.
32       Weschler, C. J.; Shields, H. C.  (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
33             Environ. 33:2301-2312.
34       Weschler, C. J.; Shields, H. C.  (2000) The influence of ventilation on reactions among indoor pollutants: modeling
35             and experimental observations. Indoor Air. 10: 92-100.
36       Weschler, C. J.; Shields, H. C.  (2003) Experiments probing the influence of air exchange rates on secondary organic
37             aerosols derived from indoor chemistry. Atmos.  Environ. 37: 5621-5631.
38       Weschler, C. J.; Shields, H. C.; Naik, D. V (1989) Indoor ozone  exposures. JAPCA 39: 1562-1568.
39       Weschler, C. J.; Hodgson, A. T.; Wooley, J. D. (1992)  Indoor chemistry: ozone, volatile organic compounds, and
40             carpets. Environ. Sci. Technol. 26: 2371-2377.
41       Weschler, C. J.; Shields, H. C.; Naik, D. V. (1994) Indoor chemistry involving O3, NO, and NO2 as evidenced by
42             14 months of measurements at a site in southern California. Environ. Sci. Technol. 28:  2120-2132.
43       Wilkins, C.  K.; Clausen, P. A.; Wolkoff, P.; Larsen, S.  T.; Hammer, M.; Larsen, K.; Hansen, V; Nielsen, G. D.
44             (2001)  Formation of strong irritants in mixtures of isoprene/ozone and isoprene/ozone/nitrogen dioxide.
45             Environ. Health Perspect.  109: 937-941.
46       Williams, R.;  Suggs, J.; Rea, A.; Leovic, K.; Vette, A.; Croghan, C.; Sheldon, L.; Rodes,  C.; Thornburg,, J.;
47             Ejire, A.; Herbst, M.; Sanders, W., Jr. (2003a) The Research Triangle Park paniculate matter panel study: PM
48             mass concentration relationships Atmos. Environ. 37: 5349-5363.
49       Williams, R.;  Suggs, J.; Rea, A.;  Sheldon, L.; Rodes, C.; Thornburg, J. (2003b) The Research Triangle Park
50             paniculate matter panel  study: modeling ambient source contribution to personal and residential PM mass
51             concentrations. Atmos. Environ. 37: 5365-5378.
52       Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Hougaard, K. S.; Nielsen, G. D. (1999) Formation of strong airway
53             irritants in a model mixture of (+)-a-pinene/ozone. Atmos. Environ. 33: 693-698.
54       Zanis, P.; Trickl, T.; Stohl, A.; Wernli, H.; Cooper, O.; Zerefos, C.; Gaeggeler, H.; Schnabel,  C.;  Tobler, L.; Kubik,
55             P.  W.; Priller, A.; Scheel, H. E.; Kanter, H. J.; Cristofanelli, P.; Forster, C.; James, P.; Gerasopoulos, E.;
56             Delcloo, A.; Papayannis, A.; Claude, H. (2003) Forecast, observation and modelling of a deep stratospheric
57             intrusion event over Europe. Atmos. Chem. Phys. 3: 763-777'.


         August  2005                                    3-83        DRAFT-DO NOT QUOTE OR CITE

-------
 i      4.  DOSIMETRY, SPECIES HOMOLOGY, SENSITIVITY,
 2            AND ANIMAL-TO-HUMAN EXTRAPOLATION
 3
 4
 5     4.1    INTRODUCTION
 6          The dosimetry of ozone (O3) in humans has been examined in a series of studies published
 7     in the past decade.  These studies further characterize the dose of O3 delivered to various sites in
 8     the respiratory tract (RT).  Ozone, classified as a reactive gas, interacts with surfactant,
 9     antioxidants, and other compounds in the epithelial lining fluid (ELF).  Researchers have
10     attempted to obtain a greater understanding of how these complex interactions affect O3 uptake
11     and O3-induced injury. New work has also been completed evaluating species differences in
12     responses to O3 exposures, which allow more accurate quantitative extrapolation from animals
13     to humans.
14          This chapter is not intended to be a complete overview of O3 dosimetry and animal-to-
15     human comparisons, but rather, it is an update of the dosimetry/extrapolation chapter from the
16     last O3 criteria document (U.S. Environmental Protection Agency, 1996), or 1996 O3 AQCD, and
17     other reviews of the earlier published literature.  The framework for presenting this chapter is
18     first a discussion in Section 4.2 of general concepts of the dosimetry of O3 in the RT.  Bolus-
19     response studies are then presented in Section 4.2.1 followed by general uptake studies in
20     Section 4.2.2. Dosimetry modeling is presented in Section 4.2.3 followed by the summary and
21     conclusions for the dosimetry material in Section 4.2.4. The chapter continues in Section 4.3
22     with a discussion of species comparisons and ends with a discussion of animal-to-human
23     extrapolation. More detailed discussions of the studies are presented in the supporting material
24     to this chapter (Annex AX4). The toxicological effects of O3 in laboratory animals and in vitro
25     test systems are discussed in Chapter 5 and direct effects of O3 in humans are  discussed in
26     Chapter 6. The historical O3 literature is very briefly summarized in this chapter, providing a
27     very concise overview of previous work. The reader is referred to the 1996 O3 AQCD for more
28     detailed discussion of the literature prior to the early 1990s.
29
30
       August 2005                              4-1         DRAFT-DO NOT QUOTE OR CITE

-------
 1      4.2   DOSIMETRY OF OZONE IN THE RESPIRATORY TRACT
 2           Ozone dosimetry refers to the measurement or estimation of the amount of O3 or its
 3      reaction products reaching and persisting at specific sites in the RT following an exposure. The
 4      compound most directly responsible for toxic effects may be the inhaled gas O3 or one of its
 5      chemical reaction products.  Complete identification of the actual toxic agents and their
 6      integration into dosimetry is a complex issue that has not been resolved. Dosimetric studies
 7      attempt to quantify the amount of O3 retained in the lung (i.e., not exhaled) or the dose of O3 or
 8      its active metabolites delivered to target cells or tissues (i.e., dose per cell or tissue surface area).
 9      For comparison, epidemiologic studies may simply consider exposure concentration while
10      clinical studies may consider the total amount of O3 inhaled (product of exposure concentration,
11      duration, and minute ventilation). Hence, dosimetric studies seek to accurately quantify dose to
12      target lung regions or tissues, whereas epidemiologic and clinical studies typically consider
13      exposures.
14           Understanding dosimetry as it relates to O3-induced injury is complex due to the fact
15      that O3 interacts  primarily with the ELF which contains surfactant and antioxidants.  In the upper
16      airways ELF is thick and highly protective  against oxidant injury.  Figure 4-1 illustrates the
17      structure of the lower airways with progression from the large airways to the alveolus. In  lower
18      airways ELF is thinner, has lower levels of antioxidants, and thus, allows more cellular injury.
19      Adding to the complexity is the fact that O3 can react with molecules in the ELF to create  even
20      more reactive metabolites, which can then diffuse within the lung or be transported out of the
21      lung to generate systemic effects.
22           A considerable number of dosimetric studies were summarized in the 1996 O3 AQCD.
23      These studies provided estimates of absorbed O3 in the RT as a whole or in regions such as the
24      upper airways (URT) or lower airways (LRT), defined as being proximal or distal to the tracheal
25      entrance, respectively.  Estimates were obtained for both humans and animals via direct
26      measurement and mathematical modeling.  The mathematical models also estimated O3 doses to
27      specific target sites such as the proximal alveolar region (PAR; first generation distal to the
28      terminal bronchioles) and the centriacinar region (CAR; junction of conducting airways and gas
29      exchange region).
30
        August 2005                               4-2        DRAFT-DO NOT QUOTE OR CITE

-------
              a.
                       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. 231).
August 2005
             DRAFT-DO NOT QUOTE OR CITE

-------
 1           In general, the consensus of experimental and modeling studies summarized in the
 2      1996 O3 AQCD supported the following conclusions: (1) for the URT, animal and human
 3      studies suggested that O3 uptake is greater in the nose than the mouth but the effect of flow on
 4      uptake was equivocal; (2) for the LRT, predicted tissue doses (O3 flux to liquid-tissue interface)
 5      were very low in the trachea, increased to a maximum in the terminal bronchioles or first airway
 6      generation in the pulmonary region, and rapidly decreased with distal progression; (3) increasing
 7      tidal volume (VT) increases  O3 uptake, whereas, increasing flow or breathing frequency (fB)
 8      decreases O3 uptake; (3) increasing flow shifts O3 uptake to the smaller peripheral airways, i.e.,
 9      toward the CAR;  and (4) similarly, the effect of exercise is to significantly increase the
10      pulmonary region total dose (mass of O3) and the CAR dose (mass per unit surface area).
11           Some cross-species in vivo comparisons were described in the  1996  O3 AQCD.
12      For instance, comparing bronchoalveolar lavage (BAL)  cells from rats and humans, it was
13      estimated that a 0.4 ppm O3 exposure in exercising humans gave 4 to 5 times the O3 dose
14      (retained) relative to rats exposed at rest to the same concentration. In vitro dosimetry studies in
15      the 1996 O3 AQCD using isolated lung preparations showed that uptake efficiency is chemical -
16      reaction dependent, indicating the importance of reaction product formation. These reaction
17      products, created mainly by the ozonolysis of polyunsaturated fatty acids, included hydrogen
18      peroxide, aldehydes, and hydroxyhydroperoxides, which are mediators of O3 toxicity. Other
19      products are created by the reaction of O3 with other ELF constituents, all of which must be
20      considered in understanding the dosimetry of O3.
21           The next two sections (4.2.1 and 4.2.2) review the available new experimental studies
22      on O3 dosimetry, all of which were conducted by Ultman and colleagues.  Table AX4-1 in
23      Annex AX4 summarizes theses studies.
24
25      4.2.1  Bolus-Response Studies
26           The bolus-response method has been used by the Ultman group as an approach to explore
27      the distribution of O3 absorption in the airways of humans.  This  non-invasive method consists
28      of an injection of a known volume and concentration of O3 during inspiration.  Ozone uptake is
29      the amount of O3  absorbed during a breath relative to the amount contained in the inhaled
30      bolus. Figure 4-2 illustrates the uptake of a series of O3 boli as a function of volumetric
31      penetration (VP), i.e., the volume between the center of mass of an inhaled bolus and the end of

        August 2005                              4-4         DRAFT-DO NOT QUOTE OR CITE

-------
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).
 1     inspiration. The inspired O3 boli (for which the uptake fractions are illustrated in Figure 4-2)
 2     were 20 ml of 2 ppm O3. Kabel et al. (1994) have previously shown that varying the O3
 3     concentration of inspired boli between 0.4 and 4 ppm does not affect the distribution of uptake as
 4     a function of VP.
 5          The O3 bolus-technique was used by Bush et al. (1996a) to ascertain differences in lung
 6     anatomy and gender that can alter the exposure-dose cascade.  Forced vital capacity (FVC), total
 7     lung capacity (TLC) and anatomic dead space (VD) were determined for ten male and ten female
 8     subjects, who then inhaled to a 20 ml bolus of 3 ppm O3 injected into the airstream. In all
 9     subjects, dosimetry differences could be explained by differences in VD. In a subsequent study,
10     Ultman et al. (2004) showed that the volume at which 50% of an inspired O3 bolus is absorbed
       August 2005
4-5
DRAFT-DO NOT QUOTE OR CITE

-------
 1      was better associated with the volume of the conducting airways than VD (see Figure 4-2).
 2      Bush et al. (1996a) pointed out that the applicability of their results may be limited because of
 3      their assumptions that the intrinsic mass transfer parameter (Ka) was independent of location in
 4      the RT and that there was no mucous resistance. They further suggested that the dependence
 5      of Ka on flowrate and VD be restricted to flowrates < 1000 mL/s until studies at higher rates have
 6      been performed.
 7           Nodelman and Ultman (1999) demonstrated that the uptake distributions of O3 boli were
 8      sensitive to the mode of breathing and to the airflow rate. As flowrates increased from 150 to
 9      1000 mL/s, O3 penetrated deeper into the lung and penetration was further increased by oral
10      relative to nasal breathing.  The authors suggest that the  switch from nasal to oral breathing
11      coupled with increases in respiratory flow as occurs during exercise causes a shift in the O3 dose
12      distribution, allowing O3  to penetrate deeper into the lung, increasing the potential for  damage to
13      bronchiolar and alveolar tissues.
14           More recently, Ultman et al. (2004) measured O3 uptake using the bolus technique in
15      60 young heathy nonsmoking adults (32 M, 28 F).  Bolus were inspired at a rate of 1 mL/s,
16      equivalent to a moderate  exercise rate with a minute ventilation of 30 L/min.  Figure 4-2
17      illustrates the O3 uptake fraction as a function  of VP in a representative subject. Anatomic dead
18      space was measured in 47 of the subjects (24 M, 23 F). In these subjects, the volume at which
19      50% of an inhaled bolus was absorbed (VP500/0) was correlated with VD (r = 0.57, p < 0.001) and
20      the volume of the conducting airways, i.e., VD minus the volume of the upper airways, (r = 0.65,
21      p = 0.001). Both VP50o/0 and VD were significantly greater in males than females, although the
22      volume of the upper airways was not. These findings suggest that in females the smaller
23      airways, and associated larger surface-to-volume ratio, enhance local O3 uptake and cause
24      reduced penetration of O3 into the distal lung.  It is not clear from these findings, however, if the
25      actual anatomical location of VP50o/0 differed between males and females.
26           A few studies have  measured the effect of a continuous pollutant exposure on O3 bolus
27      uptake.  Asplund et al.  (1996) randomly exposed young healthy adults (8 M, 3 F) for 2 h
28      [presumably  at rest] to 0.0 (air), 0.12, or 0.36 ppm O3 on 3 separate occasions separated by at
29      least 1-wk. Ozone bolus  uptake was measured preexposure and subsequently at 30 minute
30      intervals during the exposure. Ozone uptake over the VP range of 70 to 120 ml increased
31      after the air exposure, decreased slightly after the 0.12 ppm O3 exposure, and decreased

        August 2005                               4-6        DRAFT-DO NOT QUOTE OR CITE

-------
 1      more substantially following the 0.36 ppm O3 exposure.  Relative to uptake during the air
 2      exposure, O3 bolus uptake was significantly decreased by 30 minutes of the 0.12 and
 3      0.36 ppm O3 exposures and remained significantly decreased for the duration of these exposures.
 4           Using a similar protocol, Rigas et al. (1997) randomly exposed young healthy adults (6 M,
 5      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.
 6      Ozone bolus uptake (VP range of 70 to 120 ml ) was measured preexposure and every 30 minute
 7      during the exposures.  The results of an F test indicated that exposure duration (30-, 60-, 90-,
 8      120-min) was not a significant factor, but treatment (NO2, SO2, etc.) was (p < 0.01).  Ozone
 9      bolus uptake was increased by 30 minutes during the NO2 and SO2 exposures and decreased
10      during the O3 exposure.  The authors suggested that there may be increased production of
11      an O3-reactive substrate in the ELF due to airway inflammation.  During NO2 and SO2 exposures
12      the substrate was not depleted by these gases and so could react with the O3 bolus. During O3
13      exposure the substrate was depleted, causing the fractional absorption of the O3 bolus to
14      decrease.
15
16      4.2.2   General Uptake Studies
17           Ultman and colleagues have recently completed some general uptake studies to determine
18      the ratio of O3 uptake to the quantity of O3 inhaled. Uptake efficiency was determined at
19      exposures of 0.2 or 0.4 ppm O3 while exercising at a minute volume of approximately 20 L/min
20      for 60 minutes or 40 L/min  for 30 minutes in both men and women (Rigas et al., 2000).  Uptake
21      efficiency ranged from 0.56 to 0.98 and had a statistically significant but weak dependence on
22      concentration, minute volume, and exposure time. Intersubject differences had the largest
23      influence on uptake efficiency, resulting in a variation of approximately 10%.  As the quantity
24      of O3 retained by the RT is equal to uptake efficiency times the quantity of O3 inhaled, relatively
25      large changes in concentration, minute volume, or exposure time may result in relatively large
26      changes in the amount of O3 retained by the RT or absorbed locally. The authors concluded that
27      for exposure times <2 h, inhaled dose (product of O3 concentration, exposure duration, and
28      minute ventilation) is a reasonable predictor of actual uptake as long as there are fixed
29      concentrations of O3 and fixed levels of exercise.  More importantly, similarly exposed
30      individuals vary in the amount of actual dose received.
        August 2005                               4-7         DRAFT-DO NOT QUOTE OR CITE

-------
 1           Santiago et al. (2001) studied the effects of airflow rate (3 to 15 L/min) and O3
 2      concentration (0.1, 0.2, or 0.4 ppm) on O3 uptake in nasal cavities of males and females.
 3      As would be expected, uptake efficiency in the nose was inversely related to the flowrate and the
 4      concentration of O3 in the inlet air.  They computed a gas-phase diffusion resistance of <24% of
 5      overall diffusion resistance which suggested to them that simultaneously occurring diffusion and
 6      chemical reactions in the mucous layer were the limiting factors in O3 uptake. Difference in O3
 7      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
 8      effects of flowrate and concentration on uptake efficiency were statistically significant, but
 9      intersubject differences accounted for approximately half of the total variation in uptake
10      efficiency. Both these general uptake studies, done at environmentally relevant O3
11      concentrations, indicate that inter-individual differences in fractional uptake are extremely
12      important in O3 dose-response relationships.
13           In the research mentioned above, Ultman et al. (2004) also completed continuous exposure
14      studies. The same 60 subjects were exposed continuously for 1 h to either clean air or 0.25 ppm
15      ozone while exercising at a target minute ventilation of 30 L/min. This is the first study to assess
16      ventilatory and dosimetric parameters for an entire hour of exposure.  Additionally they
17      measured bronchial cross-sectional area available for gas diffusion in addition to other
18      ventilatory parameters.  At a fixed minute ventilation of 30 L/min, the uptake fraction of O3
19      decreased with increasing fB (see Figure 4-3) and increased with increasing VT.  The uptake
20      fraction was significantly  greater in males (91.4%) than females (87.1%), which is consistent
21      with the larger fB and smaller VT of the females than males. There was a small but significant
22      reduction in the breath-by-breath uptake of O3 from 90.6% on average for the first  15 minutes to
23      87.3% on average for the last 15 minutes of exposure. Ozone uptake rate correlated with percent
24      changes in individual bronchial cross-sectional area but did not correlate with individual FEVj
25      responses. Neither of these parameters correlated with the penetration volume determined in the
26      bolus  studies mentioned above. The authors concluded that the intersubject differences in forced
27      respiratory responses were not due to differences in O3 uptake. However, these data did partially
28      support the hypothesis that changes in cross-sectional  area available for gas diffusion are related
29      to overall O3 retention.
30
        August 2005                                4-8        DRAFT-DO NOT QUOTE OR CITE

-------
i.u -

Efficiency
o
CD
0)
« 0.8 -
Q.
3


0.7 -
+++ ° Q
+ + + + O
+ + c0 p <8
*
+ +^ + o
o\e +?0

o1- o
0 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 Ultmanetal. (2004).
 1     4.2.3   Dosimetry Modeling
 2          When all of the animal and human in vivo O3 uptake efficiency data are compared, there is
 3     a good degree of consistency across data sets, which raises the level of confidence with which
 4     these data sets can be used to support dosimetric model formulations. Models predict that the
 5     net O3 dose (O3 flux to air-liquid interface) gradually decreases distally from the trachea toward
 6     the end of the TB and then rapidly decreases in the pulmonary region. However, the tissue
 7     dose (O3 flux to liquid-tissue interface) is low in the trachea, increases to a maximum in the
 8     terminal bronchioles and the first generation of the pulmonary region, and then decreases rapidly
 9     distally into the pulmonary region. The increased VT and flow, associated with exercise in
10     humans or CO2-stimulated ventilation increases in rats, shifts O3 dose further into the periphery
11     of the lung, causing a disproportionate increase in distal lung dose.
       August 2005
4-9
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Localized damage to lung tissue has been modeled showing variation of O3 dose among
 2      anatomically equivalent ventilatory units as a function of path length from the trachea with
 3      shorter paths showing greater dose (Overton and Graham,  1995). More recent data indicate that
 4      the primary site of acute cell injury occurs in the conducting airways (Postlethwait et al., 2000).
 5      These data must be considered when developing models that attempt to predict site-specific
 6      locations of O3-induced injury.  The early models computed relationships between delivered
 7      regional dose and response with the assumption that O3 was the active agent responsible for
 8      injury. It is now known that reactive intermediates such as hydrohydroxyperoxides and
 9      aldehydes are important agents mediating the response to O3 (further discussed in Section 5.3.1).
10      Thus, models must consider O3 reaction/diffusion in the ELF and ELF-derived reactions
11      products.
12           Table AX4-2 in the annex presents a summary of new theoretical studies of the uptake
13      of O3 by the RTs (or regions) of humans and laboratory animals that have become available
14      since the 1996 review.  They are discussed below.
15           Overton and Graham (1995) created a rat model combining multiple path anatomic models
16      and one-dimensional convection-dispersion equations which simulates transport and uptake
17      of O3 in airways and airspaces of the modeled TB region.  Predictions from this model
18      realistically detail O3 transport and uptake of different but morphologically equivalent sites.
19      Using computational fluid  dynamics (CFD), Cohen-Hubal et al. (1996) modeled the effect of the
20      mucus layer thickness in the nasal passage of a rat.  Predictions of overall uptake were within the
21      range of measured uptake.  Predicted regional O3 flux was correlated with measured cell
22      proliferation  for the CFD simulation that incorporated two regions, each with a different mucus
23      thickness.  But using bolus-response data described above, Hu et al. (1994) and Bush et al.
24      (2001) estimate a reaction rate constant that is more than 1000 times as large as that used by
25      Cohen-Hubal et al. (1996).
26           With a RT dosimetry model, Overton et al. (1996) investigated the sensitivity of uptake
27      efficiency, proximal alveolar region (PAR) dose, and PAR dose ratio to TB region volume (V^)
28      and TB region expansion in humans and rats. The PAR was defined as the first generation distal
29      to terminal bronchioles and the PAR dose ratio was defined as the ratio  of a rat's predicted PAR
30      dose to a human's predicted PAR dose.  This ratio relates human and rat exposure concentrations
31      so that both species receive the same PAR dose. In rats, the PAR is a region of major damage

        August 2005                              4-10        DRAFT-DO NOT QUOTE OR CITE

-------
 1      from O3. For each species, three values of VXB were used:  a mean value from the literature and
 2      the mean ± twice the SD. For both the rat and human simulations, there were several general
 3      findings: (1) uptake efficiency and PAR dose both increased with decreasing V^, e.g., using the
 4      highest TB region mass transfer coefficient (k^), the PAR dose for VTB - 2SD was five times
 5      greater than the PAR dose for V^ + 2SD, (2) uptake efficiency and PAR dose both decreased
 6      with TB expansion relative to no expansion, 3) PAR dose increased with tidal volume,
 7      4) PAR dose increased with decreasing k^, and 5) uptake efficiency increased with k^.
 8          Bush et al. (2001) modified their single-path model (Bush et al., 1996b) so that simulations
 9      would coincide with experimental uptake efficiency data for O3 and C12 during oral and nasal
10      breathing.  Relative to their original model, the Bush et al.  (2001) model added lung expansion
11      and modified the mass transfer coefficients for both the gas-phase (kg) and the liquid-phase (kj).
12      Consistent with Overton et al. (1996), considering expansion of the TB airways reduced uptake
13      efficiency versus no expansion. As very little inhaled O3 reaches the peripheral lung, it was not
14      surprising that alveolar expansion had minimal affect on uptake efficiency. Ignoring the O3
15      reaction rate constant (k,.), the simulations for O3 and C12 were nearly the same since the gas-
16      phase diffusion coefficients of O3 and C12 are similar. But for a given VP the TB airways of the
17      lung, experimental bolus uptake are always less for O3 than for C12. The authors surmised that
18      the difference between the uptake for these gases could be  explained adequately based solely on
19      the diffusive resistance of O3 in airways surface fluid (modeled by k,.). Qualitatively, model
20      simulations also agreed well with the experimental data of Gerrity et al. (1995).
21          Age- and gender-specific differences in both regional and systemic uptake in humans was
22      modeled using a physiologically-based pharmacokinetic (PBPK ) approach (Sarangapani et al.,
23      2003).  The model estimated that regional (URT, TB, pulmonary) extraction efficiency of O3 is
24      relatively insensitive to age and gender.
25          A recent attempt was made (Mudway and Kelly, 2004) to model O3 dose-inflammatory
26      response using a meta-analysis of 23 exposures in published human chamber studies.  The O3
27      concentrations ranged from 0.08 to 0.6 ppm and the exposure durations ranged from 60 to
28      396 minutes. The analysis showed linear relationships between O3 dose and neutrophilia in
29      bronchoalveolar lavage fluid (BALF).  Linear relationships were also observed between O3 dose
30      and protein leakage into BALF, which suggested to the authors that a large-scale study could
31      determine a possible O3 threshold level for these inflammatory responses.  These recent findings

        August 2005                              4-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      seem consistent with the linear relationship between O3 dose to pulmonary tissues normalized
 2      for body weight and lavage fluid protein in rats, guinea pigs, and rabbits (Miller et al., 1988).
 3
 4      4.2.4   Summary and Conclusions - Dosimetry
 5           Ozone is a highly reactive gas and powerful oxidant with a short half-life. Uptake occurs
 6      in mucous membranes of the RT where O3 reacts with components of the ELF. Uptake
 7      efficiency is chemical-reaction dependent and the reaction products (hydrogen peroxide,
 8      aldehydes, and hydroxyhydroperoxides) created by ozonolysis of polyunsaturated fatty acids
 9      mediate O3 toxicity.  The 1996 O3 AQCD reported that uptake of O3 in rat is about 0.50 and in
10      humans at rest is about 0.8 to 0.95. In humans, about 0.07 of the O3 is removed in the
11      larynx/trachea, about 0.50 in the head, and about 0.43 in the lungs, where the primary site of
12      damage was believed to be the CAR. Increasing flow shifted O3 uptake distally toward smaller
13      airways of the lung.  Studies in humans showed that increasing minute ventilation with exercise
14      (by increasing both breathing frequency and tidal volume) causes only a small decrease in
15      uptake efficiency by  the total RT. The nasal passages appeared to absorb more O3 than the oral
16      passages. Comparing BAL cells, a 0.4 ppm exposure in exercising humans showed 4 to 5 times
17      the retained dose of O3 relative to rats exposed at rest to the same concentration.
18           New research on O3 uptake has been performed in humans but not in laboratory animals.
19      Bolus-response studies demonstrated that a previous continuous exposure to O3 decreases the
20      absorption of a bolus of O3, probably due to depletion of compounds able to absorb O3.
21      Continuous exposure to NO2 and SO2 increased absorption of a bolus of O3.  These data are of
22      some relevance to environmental exposures where humans may receive differing concentrations
23      of O3 depending on time of day.  Verifying prior work, the bolus-response method was used to
24      demonstrate that O3 bolus uptake is sensitive to the mode of breathing and to the airflow rate.
25      As flow is increased  from 150 to 1000 mL/s, O3 boli penetrated deeper into the lung and
26      penetration was further increased by oral  versus nasal breathing. This suggests that the switch
27      from nasal to oral breathing coupled with increases in respiratory flow as occurs during exercise
28      causes a shift in regional O3 dose deeper into the lung, increasing the potential of damage to
29      bronchiolar and alveolar tissues. The finding that O3 uptake is inversely related to airflow also
30      agrees with earlier animal studies.
        August 2005                              4-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1           New general uptake study data demonstrate that exercising men and women receiving
 2      0.2 or 0.4 ppm O3 at 20 L/min for 60 minutes or 40 L/min for 30 minutes absorb 0.56 to 0.98.
 3      The absorbed fraction or FA is affected only by large changes in concentration, minute volume,
 4      and exposure time.  This suggests that for exposure times <2 h, inhaled dose is a reasonable
 5      predictor of actual uptake as long as there are fixed concentrations of O3 and fixed levels of
 6      exercise.  Individuals exposed to similar concentrations vary considerably in the amount of
 7      actual dose received. This intersubject variability was also demonstrated in a study of O3 uptake
 8      in nasal cavities of men and women. The FA in the nose was inversely related to the flowrate
 9      and the concentration of O3 suggesting that simultaneously occurring diffusion and chemical
10      reactions in the mucous layer were the limiting factors in O3 uptake. Both these general uptake
11      studies, done at environmentally relevant O3 concentrations, indicate that inter-individual
12      differences in fractional uptake, which can range from 0.25 to 0.97, are extremely important
13      in O3 dose-response relationships.
14           The consistency of uptake data generated in animal and human studies allow a high level
15      of confidence in their use in dosimetry modeling. Early models predicted that net O3 dose to
16      ELF and tissue gradually decreases distally from the trachea toward the end of the TB and then
17      rapidly decreases in the pulmonary region. Exercise-induced or CO2-stimulated increases in VT
18      and flow, shift O3 dose further into the periphery of the lung, causing a disproportionate increase
19      in distal lung dose.  Localized damage to lung tissue has been modeled showing variation of O3
20      dose among anatomically equivalent ventilatory units  as a function of path length from the
21      trachea with shorter paths showing greater damage.
22           New models have produced some refinements of earlier models such as:  (1) the use of
23      mucus resistance and thickness in describing O3 dosimetry and determining the patterns
24      of O3-induced lesions; (2) the shape of the dose versus generation plot along any path from the
25      trachea to alveoli is independent of path, with the tissue dose decreasing with increasing
26      generation index; (3) simulations sensitive to conducting airway volume but relatively
27      insensitive to characteristics of the respiratory airspace; (4) the importance of TB region
28      expansion; (5) the importance of dose received in the PAR both inter-individual differences and
29      extrapolations based on dose; and (6) revaluation of mass transfer coefficients for conducting
30      airways. Additionally, more recent data indicate that the primary site of acute cell injury  occurs
        August 2005                              4-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1      in the conducting airways and that reactive intermediates in the ELF, rather than O3 itself, are
 2      responsible for pulmonary injury.  These data must be considered when developing new models.
 3
 4
 5      4.3   SPECIES HOMOLOGY, SENSITIVITY, AND ANIMAL-TO-HUMAN
 6            EXTRAPOLATION
 7          Basic similarities exist across human and other animals species with regard to basic
 8      anatomy, physiology, biochemistry, cell biology, and disease processes.  However, there are
 9      obviously some species differences that have the potential to affect both the patterns of O3
10      uptake in the respiratory tract as well as responses. For instance, primates are oronasal breathers
11      with a dichotomous branching lung structure, whereas, rodents are obligate nasal breathers with
12      a monopodial branching lung structure (Miller et al., 1993).  Even when comparing nasal
13      breathing, differences in the nasal structure between primates and rodents can affect both the site
14      and amount of gaseous uptake in this region  (DeSesso, 1993; Morgan et al., 1989). Cellular
15      profiles also differ between species as a function of location in the respiratory tract (Miller et al.,
16      1993; Plopper et al., 1989; Stone et al., 1992).
17          The homology as it exists creates similarities in acute O3-induced effects, especially in the
18      respiratory tract and in  lung defense mechanisms. Rodents appear to have a slightly higher
19      tachypneic response to  O3, which is clearly concentration-dependent in most species and shows
20      parallel exacerbation when hyperventilation  (e.g., exercise or CO2) is superimposed.  What is not
21      known is whether this is evidence of pulmonary irritant sensitivity, perhaps as a prelude to
22      toxicity, or whether tachypnea is a defensive action taken by the respiratory system to minimize
23      distal lung  O3 deposition. Airway or lung  resistance in humans is not affected appreciably by
24      acute exposure to O3, except under conditions of heavy exercise; animals appear to need high-
25      level exposures or special preparations that bypass nasal scrubbing.  Dynamic lung compliance
26      (Cdyn) has been shown to have small magnitude decreases in response to O3 in some studies
27      across species, but it is  thought that these changes are of little biological significance for ambient
28      exposures.  Spirometric changes, the hallmark of O3 response in humans, occur in rats, but to a
29      lesser degree. It is unclear,  however, the degree to which anesthesia (rat) and the comparability
30      of hyperventilation induced by CO2 (rat) or exercise (human) may influence this difference in
        August 2005                              4-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1      responsiveness.  Collectively, the acute functional response of laboratory animals to O3 appears
 2      quite homologous to that of the human.
 3           Examination of BAL constituents show that the influx of inflammatory cells and protein
 4      from the serum is influenced by species, but perhaps to less extent than by ventilation and
 5      antioxidant status. Adjustment for these factors can modulate responses to approximate animal
 6      responses to those of humans.  Unfortunately, these influential factors are rarely measured and,
 7      even less often, controlled. Increases in protein levels in BALF with O3 exposures in guinea pigs
 8      are also a factor in the species' susceptibility to the effects of O3. Species comparisons of
 9      acute O3 exposures to mice, guinea pigs, rats, hamsters, and rabbits found that guinea pigs were
10      the most responsive (to >0.2 ppm); rabbits were the least responsive (2.0 ppm only); and rats,
11      hamsters, and mice were intermediate (effects at > 1.0 ppm). Rats and humans have subtle
12      species-specific differences in inflammatory responses to O3 in terms of the timing of PMN
13      influx in the nasal and bronchoalveolar regions.
14           When humans are exposed to O3 repeatedly for several consecutive days, lung function
15      decrements subside, and normal spirometric parameters are regained (see Section 6.6).  This
16      phenomenon of functional attenuation also has been demonstrated in rats, not only in terms of
17      spirometry, but also in terms of the classic tachypneic ventilatory response.  Full or partial
18      attenuation of some BAL parameters also appears to occur in both rats and humans, but exposure
19      scenario appears to play a role; other cellular changes do not attenuate (see Section 6.9.4).
20      Existing epidemiologic studies provide only suggestive evidence that persistent or progressive
21      deterioration in lung function is associated with long-term oxidant-pollutant exposure (See
22      Chapter 7). With chronic, repeated exposures to >0.12 ppm O3, however, laboratory animals
23      demonstrate changes in lung structure, function, and biochemistry that are indicative of airway
24      irritation and inflammation with the possible development of chronic lung disease (U.S.
25      Environmental Protection Agency, 1996). Based on the apparent hornology of these responses
26      between humans and laboratory animals, animal studies appear to provide a means for assessing
27      such chronic health concerns.
28           A species' susceptibility to the effects of O3 exposure may be due, in part, to biochemical
29      differences among species. Evidence for this is provided by differences in activity of SD rat and
30      rhesus monkey CYP moonoxygenases elicited by O3 exposure (Lee et al., 1998).  Additional
31      characterization of species- and region-specific CYP enzymes will create a better understanding

        August 2005                              4-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1      of the differences in response to O3.  This will allow more accurate extrapolation from animal
 2      exposures to human exposures and toxic effects.
 3           Antioxidant metabolism varies widely among species, which can greatly influence the
 4      effects of O3 (discussed in greater detail in 5.2.1.3). The guinea pig appears to be the species
 5      most susceptible to O3. Early studies ranked mice > rats > guinea pigs in order of antioxidant
 6      responsiveness to O3 challenge. Guinea pigs have been shown to have lower basal levels of
 7      GSH transferase activity, lower activity of GSH peroxidases, and lower levels of vitamin E
 8      compared to rats. These lower levels of antioxidants combined with increases in protein levels
 9      in BALF (discussed above) with O3 exposures likely explain, at least in part, the species'
10      susceptibility to the effects of O3.
11           Because cytokine and chemokine responses are so important in an animal's defense
12      against O3 exposure, comparisons of differences in species expression and activity of these
13      inflammatory mediators is necessary. Arsalane et al. (1995) compared guinea pig and human
14      AM recovered in BALF and subsequently exposed in vitro to 0.1 to 1 ppm for 60 minutes.
15      Measurement of inflammatory cytokines showed a peak at 0.4 ppm in both species.  Guinea pig
16      AM had an increase in IL-6 and TNFa while human AM had increases in TNFa, IL-lb,  IL-6 and
17      IL-8. This exposure also caused an increase in mRNA expression for TNFa, IL-lb, IL-6 and
18      IL-8 in human cells. At 0.1 ppm exposures, only TNFa secretion was increased. These data
19      suggest similar cytokine responses in guinea pigs and humans, both qualitatively and
20      quantitatively.
21           Species differences in morphological responses to O3 exposure have been  characterized by
22      Dormans et al. (1999), as discussed in previous sections. Dormans et al. (1999) continuously
23      exposed rats, mice, and male guinea pigs to filtered air, 0.2, or 0.4 ppm O3 for 3, 7, 28, and
24      56 days. The animals exposed for 28 days were examined at 3, 7, or 28 days PE. Depending
25      on the endpoint studied, the species varied in sensitivity. Greater sensitivity was shown in the
26      mouse as determined by biochemical endpoints, persistence of bronchiolar epithelial
27      hypertrophy, and recovery time.  Guinea pigs were more sensitive in terms of the inflammatory
28      response though all three species had increases in the inflammatory response after three days that
29      did not decrease with exposure. These data on inflammation are in general agreement with
30      Hatch et al., (1986), discussed above. In all species, the longest exposure to the highest O3
31      concentration caused increased collagen in ductal septa and large lamellar bodies in Type II

        August 2005                              4-16        DRAFT-DO NOT QUOTE  OR CITE

-------
 1      cells, but that response also occurred in rats and guinea pigs at 0.2 ppm. No fibrosis was seen at
 2      the shorter exposure times and the authors question whether fibrosis occurs in healthy humans
 3      after continuous exposure.  The authors do not rule out the possibility that some of these
 4      differences may be attributable to differences in total inhaled dose or dose actually reaching a
 5      target site. Overall, the authors rated mice as most susceptible, followed by guinea pigs and rats.
 6           Comparisons of airway effects in rats, monkeys and ferrets resulting from exposures of
 7      1.0 ppm O3 for 8 h (Sterner-Kock et al. 2000) demonstrated that monkeys and ferrets had similar
 8      inflammatory responses and epithelial  necrosis. The response of these two species was more
 9      severe than that seen in rats. These data suggest that ferrets are a good animal model for O3-
10      induced airway effects due to the similarities in pulmonary structure between primates and
11      ferrets. However, the mechanisms of O3 effects at these high concentrations may differ from
12      those at more realistic levels.
13           A number of species, including nonhuman primates, dogs, cats, rabbits, and rodents, have
14      been used to study the effects of O3 exposure on airway bronchoconstriction. A commonly used
15      model of bronchospasm utilizes guinea pigs acutely exposed to high O3 concentrations (2 to
16      3 ppm) to induce airway hyperreactivity (AHR). As mentioned earlier, the model is helpful for
17      determining mechanistic aspects of AHR, but is not really relevant for extrapolation to potential
18      airway responses in humans exposed to ambient levels of O3.  Additionally, guinea pigs have
19      been shown to have AHR in other studies that is very similar to asthmatic humans, but the utility
20      of guinea pig data is somewhat limited by their disparity from other animal models.
21           The rat is a key species used in O3 toxicological studies, but the rat has both behavioral
22      and physiological  mechanisms that can lower core temperature in response to acute exposures,
23      thus limiting extrapolation of rat data to humans.  Iwasaki et al. (1998) evaluated cardiovascular
24      and thermoregulatory responses to O3 at exposure of 0.1, 0.3, and 0.5 ppm O3 8 hrs/day for
25      4 consecutive days. A dose-dependent disruption of HR and Tco was seen on the first and second
26      days of exposure, which then recovered to control  values. Watkinson et al. (2003) exposed
27      rats to 0.5 ppm O3 and observed this hypothermic response, which included lowered HR,
28      lowered Tco, and increased inflammatory components in BALF.  The authors suggested that the
29      response is an inherent reflexive pattern that can possibly attenuate O3 toxicity in rodents. They
30      discuss the cascade of effects created by decreases in Tco, which include: (1) lowered metabolic
31      rate, (2) altered enzyme kinetics, (3) altered membrane function, (4)  decreased oxygen

        August 2005                              4-17       DRAFT-DO NOT QUOTE OR CITE

-------
 1      consumption and demand, (5) reductions in minute ventilation, which would act to limit the dose
 2      of O3 delivered to the lungs. These effects are concurrent with changes in HR which lead to:
 3      (1) decreased CO, (2) lowered BP, and (3) decreased tissue perfusion, all of which may lead to
 4      functional deficits.  The hypothermic response has not been observed in humans except at very
 5      high exposures, which complicates extrapolation of effects in rats to humans.
 6           The importance of animal studies derives from their utilization in  determining cause-effect
 7      relationships between exposure and health outcome, but the animal data must be integrated with
 8      epidemiological studies and controlled human clinical studies. Animal  studies can corroborate
 9      both clinical and epidemiology  studies and further provide important data that is impossible to
10      collect in human studies. Toxic pulmonary and extrapulmonary effects following O3 exposure
11      have been well-studied in rodents, nonhuman primates, and a few other species; so,
12      extrapolation, both qualitative and quantitative, to human exposures and consequent health
13      effects is possible. Quantitative extrapolation, required to determine what specific exposure is
14      likely to cause an effect in humans, is theoretically founded on the equivalency of mechanisms
15      across species.  At the molecular level, O3 acts on the carbon-carbon double bond in
16      polyunsaturated fatty acids and  on sulfhydryl groups in proteins, both of which are found within
17      cell membranes in animals and humans.  At higher levels of cellular organization, cells affected
18      in animals (e.g., AMs, Type 1 cells) have similar functions in humans, and organ systems (e.g.,
19      respiratory system) have major interspecies similarities.  However, interspecies differences do
20      occur and complicate extrapolation.
21           Quantitative extrapolation, which involves a combination of dosimetry and species
22      sensitivity, still requires more research before it can be fully realized. Knowledge of dosimetric
23      animal-to-human extrapolation is more advanced than that of species-sensitivity, but
24      extrapolation models have not been completely validated, and therefore, significant uncertainties
25      remain.  Mathematical modeling of O3 deposition in the lower respiratory tract (i.e., from the
26      trachea to alveoli) of several animal species and humans shows that the pattern of regional dose
27      is similar, but that absolute values differ. In  spite of structural and ventilatory differences
28      between species, the greatest predicted tissue dose is to the CAR. Even though the CAR of rats
29      has very rudimentary respiratory bronchioles, compared to well-developed ones in primates, the
30      CAR of both rats and nonhuman primates respond similarly to O3.
        August 2005                               4-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Experimental measurement of delivered O3 doses estimate that total respiratory uptake is
 2      -47% in laboratory animals and -87% in exercising humans, while nasopharyngeal removal is
 3      -17% in rats and -40% in humans. The previous O3 AQCD (U.S. Environmental Protection
 4      Agency, 1996) provided the first quantitative animal-to-human extrapolation of morphological
 5      changes in the proximal alveolar region using rat and monkey studies.  The extrapolation
 6      predicted that a 9-year-old child would have a 20% or 75% increase in PAR tissue thickness if
 7      their sensitivity to O3 was equal to that of a rat or monkey, respectively. Adults would have
 8      15 or 70% increase, suggesting the potential for chronic effects in humans.  In spite of the
 9      significant uncertainties, this extrapolation raises concern about the potential for chronic effects
10      in humans
11           Experiments using 2 h exposures to 0.4 ppm 18O3 suggested that exercising (15 min
12      intervals, rest and exercise at 60 L/min) humans received a 4- to 5-fold higher 18O3
13      concentrations  in BAL than resting rats (Hatch et al., 1994).  That level of exposure increased
14      BAL protein and PMNs in humans, while a concentration of 2.0 ppm in rats was necessary for
15      similar effects.  Caveats in the interpretation of 18O3 studies include: (1) only a very small
16      portion of the labeled compound is recoverable to assess incorporation; and (2) if species being
17      compared differ in physiocochemical factors  controlling mass transfer and downstream O3
18      metabolism, it could cause significant differences in the amount of inhaled 18O3 that is detected
19      during subsequent tissue analysis. Further, species differences in pulmonary anatomy,
20      ventilation, antioxidants, and susceptibility all influence dose, repair processes, and tolerance to
21      subsequent O3 exposure.  Important differences between exercising humans and resting rats that
22      can affect tissue O3 dose include: (1) increased ventilation and O3 delivery with exercise;
23      (2) decreased pulmonary ventilation and body temperature during O3 exposure in rats;
24      (3) diminished  dose received in rats due to their burying their noses in their fur during exposure;
25      and (4) increased concentration of antioxidants in ELF in rats compared to humans. These
26      antioxidants are important for converting O3 to inactive products before toxicity occurs (Kari
27      et al., 1997; Gunnison and Hatch, 1999; Plopper et al.,  1998), though this quenching is not
28      quantitative.  These and possibly other differences between rats  and humans suggest that a
29      2 ppm exposure in nonexercising rats approximates a 0.4 ppm exposure in exercising humans.
30      Further comparisons of exercising human exposure to 0.1 ppm for 6 hours (Devlin et al., 1991)
        August 2005                              4-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1     and resting rat exposure to 0.3 ppm show inflammatory and permeability changes in humans but
 2     not rats.
 3
 4     4.3.1   Summary and Conclusions:  Species Homology, Sensitivity, and
 5             Animal-to-Human Extrapolation
 6          Comparisons of acute exposures in rats and humans suggest that, though both species have
 7     similar qualitative responses to O3 exposure, there are interspecies mechanistic disparities that
 8     necessitate careful comparisons of dose-response relationships. There is no perfect nonhuman
 9     species with which to model O3 toxicity. All have limitations that must be considered when
10     attempting to extrapolate to human exposures. Awareness of these limitations, even at the level
11     of subtle strain differences within a test species, is extremely important.  The currently available
12     data suggest that LOELs in resting rats are approximately 4- to 5-fold higher than for exercising
13     humans for toxicological endpoints including BAL protein and BAL PMNs.  Studies comparing
14     species-specific differences in O3-induced effects showed that guinea pigs were the most
15     susceptible, rabbits the least susceptible, and rodents intermediate in susceptibility. The recent
16     work being done utilizing various mouse strains with differing sensitivities to O3 will help us to
17     understand the extremely complex inter-individual differences in human sensitivity to O3.
18
       August 2005                              4-20        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES

  2       Arsalane, K.; Cosset, P.; Vanhee, D.; Voisin, C.; Hamid, Q.; Tonnel, A.-B.; Wallaert, B. (1995) Ozone stimulates
  3             synthesis of inflammatory cytokines by alveolar macrophages in vitro. Am. J. Respir. Cell Mol. Biol.
  4             13:60-68.
  5       Asplund, P. T.; Ben-Jebria, A.; Rigas, M. L.; Ultman, J. S. (1996) Longitudinal distribution of ozone absorption in
  6             the lung: effect of continuous inhalation exposure. Arch. Environ. Health 51: 431-438.
  7       Bush, M. L.; Asplund, P. T.; Miles, K. A.; Ben-Jebria, A.; Ultman, J. S. (1996a) Longitudinal distribution of O3
  8             absorption in the lung: gender differences and intersubject variability. J. Appl. Physiol. 81: 1651-1657.
  9       Bush, M. L.; Raybold, T.; Abeles, S.; Hu, S.-C; Ben-Jebria, A.; Ultman, J. S.  (1996b) Longitudinal distribution of
10             ozone absorption in the lung: simulation with a single-path model. Toxicol. Appl. Pharmacol. 140: 219-226.
11       Bush, M. L.; Zhang, W.; Ben-Jebria, A.; Ultman, J. S. (2001) Longitudinal distribution of ozone and chlorine in the
12             human respiratory tract: simulation of nasal and oral breathing with the single-path diffusion model. Toxicol.
13             Appl. Pharmacol. 173: 137-145.
14       Cohen-Hubal, E. A.; Kimbell, J. S.; Fedkiw, P. S. (1996) Incorporation of nasal-lining mass-transfer resistance into a
15             CFD model for prediction of ozone dosimetry in the upper respiratory tract. Inhalation Toxicol. 8: 831-857.
16       DeSesso, J. M. (1993) The relevance to humans of animal models for inhalation studies  of cancer in the nose and
17             upper airways. Qual. Assur. (San Diego) 2: 213-231.
18       Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
19             Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the
20             lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
21       Dormans, J. A. M. A.; VanBree, L.; Boere, A. J. F.; Marra, M.; Rombout, P. J. A. (1999) Interspecies differences in
22             time course of pulmonary toxicity following repeated exposure to ozone. Inhalation Toxicol. 11: 309-329.
23       Gerrity, T. R.; Biscardi, F.; Strong, A.; Garlington, A. R.; Brown, J. S.; Bromberg, P. A. (1995) Bronchoscopic
24             determination of ozone uptake in humans. J. Appl. Physiol. 79: 852-860.
25       Gunnison, A. F.; Hatch, G. E. (1999) O3-induced inflammation in prepregnant, pregnant, and lactating rats correlates
26             with O3 dose estimated by 18O. Am. J.  Physiol. 276: L332-L340.
27       Hatch, G. E.; Slade, R.; Stead, A. G.; Graham, J. A. (1986) Species comparison of acute inhalation toxicity of ozone
28             and phosgene.  J. Toxicol. Environ. Health 19: 43-53.
29       Hatch, G. E.; Slade, R.; Harris, L. P.; McDonnell, W. F.; Devlin, R. B.; Koren, H. S.; Costa, D. L.; McKee, J. (1994)
30             Ozone dose and effect in humans and rats: a comparison using oxygen-18 labeling and bronchoalveolar
31             lavage. Am. J.  Respir.  Crit. Care Med. 150: 676-683.
32       Hu, S.-C.; Ben-Jebria, A.; Ultman, J. S. (1994) Longitudinal distribution of ozone absorption in the lung: effects of
33             respiratory flow. J. Appl. Physiol. 77:  574-583.
34       Iwasaki, T.; Takahashi, M.; Saito, H.; Arito, H. (1998) Adaptation of extrapulmonary responses to ozone exposure in
35             conscious rats. Ind. Health 36: 57-60.
36       Kabel, J. R.; Ben-Jebria, A.; Ultman, J.  S. (1994) Longitudinal distribution of ozone absorption in the lung:
37             comparison of nasal and oral quiet breathing. J. Appl. Physiol. 77: 2584-2592.
38       Kari, F.; Hatch, G.; Slade, R.; Crissman, K.; Simeonova, P. P.; Luster, M. (1997) Dietary restriction mitigates
39             ozone-induced lung inflammation in rats: a role for endogenous antioxidants. Am. J. Respir. Cell Mol. Biol.
40             17:740-747.
41       Lee, C.; Watt, K. C.; Chang, A. M.; Plopper, C. G.; Buckpitt, A. R.; Pinkerton, K. E. (1998) Site-selective
42             differences in cytochrome P450 isoform activities: comparison of expression in rat and rhesus monkey lung
43             and induction in rats. Drug Metab. Dispos. 26: 396-400.
44       Miller, F. J.; Overton, J. H.; Gerrity, T.  R.; Graham, R. C. (1988) Interspecies dosimetry of reactive gases. In:
45             Mohr, U.; Dungworth, D.; McClellan, R.; Kimmerle, G.;  Stober, W.; Lewkowski, J., eds. Inhalation
46             toxicology: the design and interpretation of inhalation studies and their use in risk assessment. New York,
47             NY: Springer-Verlag; pp. 139-155.
48       Miller, F. J.; Overton, J. H.; Kimbell, J. S.; Russell, M. L. (1993) Regional respiratory tract absorption of inhaled
49             reactive gases. In: Gardner, D. E.; Crapo, J. D.; McClellan, R. O., eds. Toxicology of the lung. 2nd ed.
50             New York, NY: Raven Press; pp. 485-525. (Target organ toxicology series).
51       Morgan, K. T.; Monticello, T. M.; Patra, A. L.; Fleishman, A. (1989) Preparation of rat nasal airway casts and their
52             application to studies of nasal airflow. In: Crapo, J. D.; Smolko, E. D.; Miller, F.  J.; Graham, J. A.; Hayes,
53             A. W., eds. Extrapolation of dosimetric relationships for inhaled particles and gases. New York, NY:
54             Academic Press, Inc.; pp. 45-58.


         August 2005                                    4-21          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Mudway, I. S.; Kelly, F. J. (2004) An investigation of inhaled ozone dose and the magnitude of airway inflammation
  2             in healthy adults. Am. J. Respir. Crit. Care Med. 169: 1089-1095.
  3       Nodelman, V.; Ultman, J. S. (1999) Longitudinal distribution of chlorine absorption in human airways: a comparison
  4             to ozone absorption. J. Appl. Physiol. 87: 2073-2080.
  5       Overton, J. H.; Graham, R. C. (1995) Simulation of the uptake of a reactive gas in a rat respiratory tract model with
  6             an asymmetric tracheobronchial region patterned on complete conducting airway cast data. Comput. Biomed.
  7             Res. 28: 171-190.
  8       Overton, J. H.; Graham, R. C.; Menache, M. G.; Mercer, R. R.; Miller, F. J. (1996)  Influence of tracheobronchial
  9             region expansion and volume on reactive gas uptake  and interspecies dose extrapolations. Inhalation Toxicol.
10             8:723-745.
11       Plopper, C. G.; St. George, J.; Mariassy, A.; Nishio, S.; Heidsiek, J.; Weir, A.; Tyler, N.; Wilson, D.; Cranz, D.;
12             Hyde, D. (1989) Species differences in airway cell distribution and morphology. In: Crapo, J. D.; Smolko,
13             E. D.; Miller, F. J.; Graham, J. A.; Hayes, A. W., eds. Extrapolation of dosimetric relationships for inhaled
14             particles and gases. New York, NY: Academic Press, Inc.; pp.  19-34.
15       Plopper, C. G.; Hatch, G. E.; Wong, V.; Duan, X.; Weir, A. J.; Tarkington, B. K.; Devlin, R. B.;  Becker, S.;
16             Buckpitt, A. R. (1998) Relationship of inhaled ozone concentration to acute tracheobronchial epithelial injury,
17             site-specific ozone dose and glutathione depletion in rhesus monkeys. Am. J. Respir. Cell Mol. Biol.
18             19:387-399.
19       Postlethwait, E. M.;  Joad, J. P.; Hyde, D. M.; Schelegle, E.  S.; Brie, J. M.; Weir, A. J.; Putney, L. F.; Wong, V. J.;
20             Velsor, L. W.; Plopper, C. G. (2000) Three-dimensional mapping of ozone-induced acute  cytotoxicity in
21             tracheobronchial airways of isolated perfused rat lung. Am.  J. Respir. Cell Mol. Biol. 22:  191-199.
22       Rigas, M. L.; Ben-Jebria, A.; Ultman, J. S. (1997) Longitudinal distribution of ozone absorption  in the lung: effects
23             of nitrogen dioxide, sulfur dioxide, and ozone exposures. Arch. Environ. Health 52: 173-178.
24       Rigas, M. L.; Catlin, S. N.; Ben-Jebria, A.; Ultman, J. S. (2000) Ozone uptake in the intact human respiratory tract:
25             relationship between inhaled dose and actual dose. J. Appl. Physiol. 88: 2015-2022.
26       Santiago, L. Y.; Harm, M. C.; Ben-Jebria, A.; Ultman, J. S.  (2001) Ozone adsorption in the human nose during
27             unidirectional airflow. J. Appl. Physiol. 91: 725-732.
28       Sarangapani, R.;  Gentry, P. R.; Covington, T. R.; Teeguarden, J. G.; Clewell, H. J., III. (2003) Evaluation of the
29             potential impact of age- and gender-specific lung morphology and ventilation rate on the dosimetry of vapors.
30             Inhalation Toxicol. 15: 987-1016.
31       Sterner-Kock, A.; Kock, M.; Braun, R.; Hyde, D. M. (2000) Ozone-induced epithelial injury in the ferret is similar
32             to nonhuman primates. Am. J. Respir. Crit. Care Med. 162:  1152-1156.
33       Stone, K. C.; Mercer, R. R.; Gehr, P.; Stockstill, B.; Crapo,  J. D. (1992) Allometric relationships of cell numbers and
34             size in the mammalian lung. Am. J. Respir. Cell Mol. Biol. 6: 235-243.
35       U.S. Environmental  Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
36             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
37             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available online at:
38             www.epa.gov/ncea/ozone.htm.
39       Ultman, J. S.; Ben-Jebria, A.; Arnold, S. F. (2004) Uptake distribution of ozone in human lungs:  intersubject
40             variability in physiologic response. Boston, MA: Health Effects Institute. Available:
41             http://www.healtheffects.org/Pubs/Ultman.pdf [29  July, 2005].
42       Watkinson, W. P.; Campen, M. J.; Wichers, L. B.; Nolan, J. P.; Costa, D. L. (2003) Cardiac and thermoregulatory
43             responses to inhaled pollutants in healthy and compromised rodents:  modulation via interaction with
44             environmental factors. Environ. Res. 92: 35-47.
45       Weibel, E. R. (1980) Design and structure of the human lung. In: Fishman, A. P., ed. Pulmonary diseases and
46             disorders. New York, NY: McGraw-Hill; p. 231.
47
48
         August 2005                                    4-22         DRAFT-DO NOT QUOTE OR CITE

-------
 i           5.  TOXICOLOGICAL EFFECTS OF OZONE AND
 2             RELATED PHOTOCHEMICAL OXIDANTS IN
 3                LABORATORY ANIMALS AND IN VITRO
 4                                   TEST SYSTEMS
 7     5.1    INTRODUCTION
 8          A wide range of effects of ozone (O3) has been demonstrated in laboratory animals. The
 9     major research findings are that environmentally relevant levels of O3 cause lung inflammation;
10     decreases in host defenses against infectious lung disease; acute changes in lung function,
11     structure, and metabolism; chronic lung disease,  some elements of which are irreversible; and
12     systemic effects on target organs (e.g., brain, heart, liver, immune system) distant from the lung.
13     The research also has served to expand the understanding of mechanisms of O3 toxicity and the
14     relationships between concentration and duration of exposure.
15          The framework for presenting the health effects of O3 in animals begins with a presentation
16     of respiratory tract effects, followed by systemic effects, and then interactions of O3 with other
17     common co-occurring pollutants.  The information discussed in this chapter is founded on a very
18     wide body of literature on studies in laboratory animals and on in vitro test systems of animal
19     cell lines and organ systems that may mimic responses in intact animals.  The direct effects of O3
20     in humans are discussed in the following chapter (Chapter 6).
21          This chapter is  not intended to be a compendium of all that is known about O3; rather, it is
22     an update of the toxicology chapter from the last O3 criteria document (U.S. Environmental
23     Protection Agency, 1996), or 1996 O3 CD, and other reviews of the earlier published literature.
24     The historical O3 literature is very briefly  summarized in an opening paragraph of each section
25     or subsection. This paragraph is intended as a very concise overview of previous work, and the
26     reader is referred to the 1996 O3 CD for more detailed discussion of the literature prior to the
27     early 1990's. Each section then continues with brief discussions of the key new studies (or
28     somewhat older studies that were not included in the previous CD).  Longer discussions of new
29     studies are included where warranted. Sections are ended with comparisons of data from the
30     previous CD with new data and basic conclusions are drawn.  Summaries of new studies and
31     results are provided in tables in Annex AX5.

       August 2005                              5-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Except for nitrogen dioxide (NO2), the subject of another criteria document (U.S.
 2     Environmental Protection Agency, 1993), there is very little relevant information on other
 3     photochemical oxidants in the published literature.  What is known about the effects of these
 4     other oxidants is also summarized briefly in this chapter.
 5
 6
 7     5.2   RESPIRATORY TRACT EFFECTS OF OZONE
 8     5.2.1   Biochemical Effects
 9           Biochemically detected effects of O3 are integrally involved in effects on both structure
10     and function (respiratory and nonrespiratory) of the respiratory tract.  Changes in xenobiotic
11     metabolism, antioxidant metabolism and oxygen  consumption, lipids and arachidonic acid
12     metabolism, and collagen metabolism are all observed with O3 exposure, though the mechanisms
13     and associations are not fully understood.
14
15     5.2.1.1   Cellular Targets of O3 Interaction
16           Ozone has the potential to interact with a wide range of different cellular components that
17     include polyunsaturated fatty acids (PUFAs); some protein amino acid residues; and some
18     low-molecular-weight compounds that include glutathione (GSH), urate, vitamins C and E, and
19     free amino acids.  Early work demonstrated that O3 being a highly reactive compound, does not
20     penetrate much beyond the epithelial lining fluid (ELF). Reaction/diffusion analyses suggest
21     that O3, at environmentally-relevant concentrations, diffuses no more than 0.1 to 0.2 jim into the
22     ELF. Ozone-induced cell damage most likely results from its reactions with PUFAs to form
23     stable but less reactive ozonide, aldehyde, and hydroperoxide reaction products. These reaction
24     products (Crigee ozonides and hydroxyhydroperoxides) may act as signal transduction
25     molecules involved in signaling of cellular responses  such as inflammation, and thus mediate O3
26     toxicity. These reactions are summarized in Figure 5-1 and studies published since the 1996
27     AQCD are listed in Table AX5-1.
28           Frampton et al. (1999) demonstrated the ozonation of PUFA to form nonanal and hexanal
29     in rat BAL after exposures to 0.22 ppm O3 for 4 h with exercise.  Increases in nonanal were not
30     accompanied by significant changes in lung function, in epithelial permeability, or in airway
31     inflammation. Hexanal levels did not increase significantly and levels of both aldehydes

       August 2005                               5-2        DRAFT-DO  NOT QUOTE OR CITE

-------
            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.   Major secondary products of ozone interaction with epithelial lining fluid and
                    lung cells.
 1     returned to baseline by 18 h PE. Pryor et al. (1996) exposed rats to 0.5 to 10 ppm O3 both with
 2     and without 5% CO2 to measure the amount of aldehyde generated in BAL, and also the rate of
 3     disappearance of aldehydes from the ELF following the O3 exposure. Ozone exposure with CO2
 4     increased the tidal volume and the yield of aldehydes with a maximal aldehyde yield at 2.5 ppm
 5     for 1 h. Absolute yields were impossible to ascertain in this system because deposition of O3 is
 6     unknown and aldehyde recovery is not complete due to loss of aldehyde by volatization and by
 7     diffusion into underlying tissue. The data showed that at 0.5 ppm O3 with 5% CO2, levels of
 8     hexanal and nonanal increased at 30 minutes, decreased slightly from that level at 60 minutes,
 9     was maximal at 90 minutes and then dropped to 60 minutes levels at 120 minutes. Levels of
10     heptanal did not change appreciably during this time course. Levels of these aldehydes were
11     dependent on a dynamic relationship between their production and the disappearance from the
12     ELF.  The authors stated that O3 is the limiting reagent in this process because the amount of
13     PUFA far exceeds the amount of O3 on a molar basis. Because of the limitations of measuring
14     aldehydes in this study paradigm, it is not useful for quantitative dosimetry; however, the authors
15     suggest the study does serve to demonstrate the use of aldehydes as biomarkers of O3 exposure
16     since nonanol is produced in an O3-specific pathway.
17          Postlethwait et al. (1998) utilized three biologically relevant models (isolated epithelial
18     lining fluid, intact lung, and liposome suspensions) to determine the O3-induced production of
19     heptanal, nonanal and hexanal in an attempt to estimate formation of lipid-derived bioactive
       August 2005
                      DRAFT-DO NOT QUOTE OR CITE

-------
 1      compounds.  Exposures used were 0.25 to 1.0 ppm for 30 to 60 minutes. Data suggest that
 2      PUFAs directly react with O3 and the amount of bioactive lipids produced is inversely related to
 3      ascorbic acid availability.  The authors caution that there are limitations to the use of
 4      measurements of these reactions products in determining O3 dose-response relationships due to
 5      the heterogenous nature of O3 reactions in the epithelial lining fluid. Connor et al. (2004) have
 6      recently examined the reactive absorption of O3 (0.3 to 1.1 ppm for 1 to 2 h) within ELF using
 7      interfacial films composed of dipalmitoylglycero-3-phosphocholine (DPPC) and rat lung lavage
 8      fluid.  The films reduced O3 reactive absorption by antioxidants.  Further experiments using a
 9      human lung fibroblast cell line exposed to O3 demonstrated that ascorbic acid (AA) produced
10      cell injury that high levels of O3 and AA were needed to induce cell injury, and the DPPC films
11      reduced the amount of cell injury. From these data the authors suggest that O3 reactions with
12      ELF substrates cause cell injury that films of active, saturated phospholipids reduce the local
13      dose of O3-derived reaction products, and that these interfacial phospholipids modulate the
14      distribution of inhaled  O3 and the  extent of site-specific cell injury.
15           Recent studies have examined the formation of ozonation products such as
16      4-hydroxynonenal (FINE), a toxic aldehyde that reacts with cysteine, histamine, and lysine
17      amino acid residues and creates protein adducts. Hamilton et al. (1998) demonstrated (see
18      Chapter 6) using human AM exposed to 0.4 ppm O3 for 1 h that exposure caused apoptosis, an
19      increase in a 32-kDa protein adduct, and an increase in ferritin and a 72-kDa heat shock protein.
20      By exposing AM to FINE in vitro, all of these effects are replicated, which the authors interpret
21      to mean that creation of protein adducts and apoptotic cell death are cellular toxic effect of acute
22      O3 exposure and that it is mediated, at least in part by FINE.
23           These recent reports combined with observations reported in the previous O3 CD (US
24      Environmental Protection Agency, 1996) suggest that interactions of O3 with cellular
25      components and ELF generate toxic ozonation products and mediate toxic effects through these
26      products.
27
28      5.2.1.2  Monooxygenases
29           Both short- and long-term exposures to O3 have been shown to enhance lung xenobiotic
30      metabolism, possibly as a result of changes in the number and function of bronchiolar epithelial
31      Clara cells and alveolar epithelial  Type 2 cells.  Studies of the effects of O3 on lung

        August 2005                               5-4        DRAFT-DO NOT QUOTE OR CITE

-------
 1      monooxygenases are listed in Table AX5-2.  Early studies showed that exposure to O3 increased
 2      CYP 2B1 (the major CYP isoform in rat lung) content and activity in rat lung. Ozone exposures
 3      also caused hypertrophy and hyperplasia of CYP 2B 1-immunoreactive Clara cells. Comparisons
 4      of rat and rhesus monkey CYP isoforms demonstrated species-specific and region-specific (e.g.,
 5      trachea, parenchyma) differences in the activities of P450 isoforms (Lee et al., 1998)
 6          Watt et al. (1998) found that 1 ppm O3 in both acute (8h, 1 ppm) and chronic (90 days,
 7      1 ppm) exposures in rat increased CYP 2E1 in a region-specific manner. Paige et al. (2000a)
 8      showed that a long term exposure (0.8 ppm,  8h/day for 90 days) increased the activity of CYP
 9      2B in distal lung but not trachea or intrapulmonary airways.  Studies have focused on P450 gene
10      expression to examine possible genetic mechanisms that may explain differential O3-sensitivity
11      (Mango et al., 1998). Mice (129 strain) deficient in Clara cell secretory protein (CCSP-/-),
12      which are oxidant-sensitive, were exposed to 1 ppm O3 for 2 hours. The CCSP null mice
13      demonstrated increases in IL-6 and metallothionein (Mt) mRNA that preceded decreases in
14      Clara cell CYP 2F2 mRNA (normally expressed at high levels in mouse lung) levels. In 129
15      strain wildtype (WT) mice,  RNA levels changed similarly, to a lesser degree. These data
16      suggest a protective role against oxidant damage for CCSP, and further, that genetic
17      susceptibility to oxidant stress may be mediated, in part, by the gene coding for CCSP.
18
19      5.2.1.3  Antioxidants, Antioxidant Metabolism, and Mitochondrial Oxygen Consumption
20          Ozone also undergoes reactions with ascorbic acid (AA), reduced glutathione (GSH), and
21      uric acid (UA), all antioxidants present in ELF (see Figure 5-2, A). In vivo experiments have
22      shown that reactions with O3 occur preferentially with antioxidants compared to proteins and
23      lipids also present in ELF. This is a protective interaction, but even with environmentally
24      relevant exposures to O3, the reactivity of O3 is not quantitatively quenched. Antioxidants offer
25      some protection from O3 exposure but often  are not maintained at concentrations sufficient to
26      fully protect the lung. Thus, O3-induced cell injury occurs in both the lower and upper
27      respiratory tract. Early work has shown that acute (1  week) exposures to <1 ppm O3 increase
28      antioxidant metabolism, including levels of cytosolic  enzymes glucose-6-phosphate
29      dehydrogenase (G6PD), 6-phosphogluconate dehydrogenase (6PGD), glutathione reductase
30      (GR), and glutathione peroxidase (GSHPx).  Re-exposure after a recovery period causes
31      increases equivalent to first-time exposures,  thus previous exposure appears to not be protective.

        August 2005                              5-5        DRAFT-DO NOT QUOTE OR CITE

-------

          I
          ^
         Epi.  CCJL
         Inst.
       Endo.
 Ozone
     \
    t
                                                                                   Mucus
                       ||GSH||[~AA~| |  UA ||  ^os
                              Lipid
           CHO
 Secondary
  oxidants
*   I  X
                                            I//.C
=sc_i.
               7:/.r.
                                                                                        A.AI.
                                                                                          B.E.C.
       Figure 5-2. (Reprinted from Molecular Aspects of Medicine, LS. Mudway and F.J. Kelly,
                   Ozone and the Lung: a sensitive issue, page 36, (2000), with permission from
                   Elsevier).
 1          Increases in enzyme activity appear to increase as a function of age, suggesting that O3
 2     exposure can cause greater lung injury in the older animal. This has been attributed to
 3     differences in dose reaching lung target sites, differing base levels of antioxidants and
 4     antioxidant enzymes, and differences in cellular sensitivity. Species differences exist in
 5     antioxidant metabolism, with guinea pigs being very sensitive to O3 due to their diminished
 6     increases in antioxidants and antioxidant enzymes. Chronic exposures of rats to urban patterns
 7     of O3 (daily peaks of 0.25 ppm) caused increases in GSHPx and GR, but not superoxide
 8     dismutase (SOD).  The enzyme changes could be accounted for by differences in the steady-state
 9     cell population or in cellular antioxidant capacity. More recent studies examining antioxidants
10     and O3 exposure are listed in Table AX5-3.
11          Ozone induced both site- and cell-specific changes in copper-zinc (Cu-Zn) and manganese
12     (Mn) SOD in rats exposed to 1.0 ppm O3 for up to 3 months (Weller et al., 1997). Cu-Zn SOD
13     labeling was decreased in epithelial cells in airways and parenchyma.  Mn SOD labeling was
14     increased in both AM and epithelial type II cells  of the centriacinar region (CAR), which the
       August 2005
                       5-6
DRAFT-DO NOT QUOTE OR CITE

-------
 1      authors suggest may allow these cells to tolerate further O3 exposure. This work is in agreement
 2      with earlier work suggesting a role of SOD in protection of cells against oxidative stress.
 3           Freed et al. (1999) evaluated the role of antioxidants in O3-induced oxidant stress in dogs
 4      (exposed to 0.2 ppm in a 6 h exposure) by inhibiting the antioxidant transport using probenecid
 5      (an anion-transport inhibitor).  Blocking antioxidant transport caused heterogeneously
 6      distributed increases in peripheral airway resistance and reactivity, supporting the hypothesis
 7      that in the lung periphery, endogenous antioxidants moderate the effects of O3 and that this
 8      exposure is a subthreshold stimulus for producing effects on peripheral airway resistance and
 9      reactivity in dogs.  The authors further found that treatment with probenecid also inhibited O3-
10      induced neutrophilic inflammation, providing evidence for a dissociation between airway
11      function and inflammation. This suggests that O3-induced inflammation and airway
12      hyperreactivity (AHR) are independent phenomena operating through multiple mechanistic
13      pathways.
14           Mudway and Kelly (1998) modeled the interactions of O3 with ELF antioxidants using a
15      continually mixed, interfacial exposure set up with O3 concentrations of ranging from 0.1 to
16      1.5 ppm for durations ranging from 30 to  720 min.  Uric acid was ranked the most O3-reactive,
17      AA the second most reactive, and GSH the least reactive.  Thus, they concluded that GSH is not
18      an important substrate for O3, while UA appeared to be the most important reactive substrate,
19      which confers protection from O3 by  removing it from inhaled air and limiting the amount that
20      reaches the distal lung. By providing a substrate for O3 reactions in the ELF, UA effectively
21      reduces the diffusive resistance of O3 (see Bush et al., 2001) in the TB airways and thus may
22      serve to limit the amount of O3 reaching the distal lung.  The authors acknowledge limitations in
23      extrapolating these data to in vivo O3 exposures due to the absence of surfactant lipids and
24      airway mucus in the model.
25
26      5.2.1.4  Lipid Metabolism and Content of the Lung
27           One of the major postulated molecular mechanisms of action of O3 is peroxidation of
28      mono- and polyunsaturated fatty acids and unsaturated neutral  lipids in the lung. Because all of
29      these lipids appear both in cell membranes and as secretions in the ELF, it is difficult to ascertain
30      which lipid pool contributes to the formation of lipid ozonation products.  As mentioned, Os can
31      penetrate only  about 0.1 to 0.2 jim into the ELF, so it is unlikely that Os reacts directly with

        August 2005                               5-7         DRAFT-DO NOT QUOTE OR CITE

-------
 1      epithelial cell membranes, except in regions of distal lung where ELF is very thin or absent. The
 2      inflammatory cascade (shown in Figure 5-3) initiated by Ch generates a mix of secondary
 3      reactants (e.g., aldehydes) which then are likely to oxidize lipids and proteins in cell membranes.
 4           In both acute and short-term studies, a variety of lung lipid changes occur, including an
 5      increase in arachidonic acid.  Metabolism of arachidonic acid produces a variety of biologically
 6      active mediators that can, in turn, affect host defenses, lung function, the immune system, and
 7      other functions.  The protein A component of surfactant is also a primary target of O3 interaction.
 8      During the first few days of O3 exposure, the changes in lung lipid biosynthesis can be accounted
 9      for by the alveolar epithelial proliferative repair. With longer exposures (e.g., 0.12 ppm for
10      90 days) an increase in PUFAs and a decrease in cholesterol-esters are seen, indicative of
11      long-term alterations of surfactant lipid composition.
12           Several new studies listed in Table AX5-4 examined the effects of O3 exposure on
13      phospholipids in lung tissue.  Ozonation  of PUFAs has been shown to generate other aldehydes
14      such as nonanal and hexanal in rat (Pryor et al., 1996; Frampton et al., 1999). These aldehydes
15      are short-lived and found to not affect lung function (Frampton et al.,  1999). These observations
16      suggest that levels of these aldehydes are dependent on a  dynamic relationship between their
17      production and their disappearance from the ELF.
18           Pryor et al. (1995) proposed a cascade mechanism whereby ozonation products cause
19      activation of specific lipases,  which then trigger the activation of second messenger pathways
20      (e.g., phospholipase A2 or phospholipase C). This group (Kafoury et al.,  1999) showed that
21      exposure of cultured human bronchial epithelial cells to the lipid ozonation product 1 -palmitoyl-
22      2-(9-oxononanoyl)-sn-glycero-3-phosphocholine elicited  release of platelet-activating factor
23      (PAF) and prostaglandin E2, but not IL-6.  The lipid ozonation product 1-hydroxy-l-
24      hydroperoxynonane caused release of PAF and IL-6 in these cells, but not prostaglandin E2.
25      These results suggest to the authors that O3-induced production of lipid ozonation products
26      causes release of proinflammatory mediators that then generate an early inflammatory response.
27           Very new work (Ballinger et al., 2005) has shown that ozone-induced membrane oxidation
28      is augmented by antioxidants present in ELF.  They utilized a red cell membrane model exposed
29      to 0.8 ppm O3 for 30 min.  The monolayer of cells  was intermittently covered by an aqueous film
30      consisting of rat BALF or B ALF plus added antioxidants.  AA and GSH induced dose-dependent
        August 2005                               5-8        DRAFT-DO NOT QUOTE OR CITE

-------
Airway
                                     o.
Epithelial
Lining
Fluid
Epithelium
Interstitial
Space
Endothelium
Airway
Smooth
Muscle

H H
\ / »


R R' 1 ^""~
/ Superoxide
/ + H20

/^/°
RP

o ^*X
(^ ^ u r\ i
\ \ h-x
H H l\
Lipid Ozonation Products / 9
e.g., aldehydes, hydroxyhydroperoxides, H2O2 /
                                        Reactive Oxygen
                                          Intermediates
                      Alveolar
                     Macrophage
NF-KB-lKB
Complex
                                                              Altered
                                                            Spirometry
Increased Transcription
 flCAM-1   IL-2
 iVCAM    IL-
  GM-CSF  IL-6
  TNFa    etc.
  iNOS
                                                                            Lipase Activation

                                                                                   I
                                                                             Phospholipases

                                                                                   1
                                                                            ArachidonicAcid
                                                                                      \
        Recruitment and
         Activation of
       Inflammatory Cells
        e.g.,AMS, PMNs,
         blood monocytes
             mast cells
    Proteolytic
     Enzymes
   Serum Proteins
                                                                         Leukotrienes  Prostanoids
                                         Platelet
                                        Activating
                                         Factor
                                                          Tnrombolytic
                                                            Effects?
                                               Airway
                                           Hyperreactivity
                              Adapted from: Pryor et al. (1995); Krishna et al. (1998); Bhalla et al. (1999)
   Figure 5-3.  Mechanisms of ozone toxicity.
August 2005
             5-9
                                                 DRAFT-DO NOT QUOTE OR CITE

-------
 1      oxidative damage to the cell membrane proteins and lipids via secondary oxidant formation.
 2      The authors concluded that early in O3 exposure, ELF antioxidants are high enough to drive
 3      reactive absorption of O3 into the ELF and to concurrently quench secondary reaction products,
 4      thus limiting cell injury.  With continued exposure, antioxidants levels decrease such that
 5      unreacted O3 and cytotoxic products can diffuse to the cell membranes, causing injury.
 6      Limitations of this in vitro study are the possible differences in chemical species and
 7      mechanisms compared to in vivo systems.
 8           Uhlson et al. (2002) reacted O3 with calf lung surfactant which resulted in the production
 9      of l-palmitoyl-2-(9'-oxo-nonanoyl)-glycerophosphocholine (16:Oa/9-al-GPCho).  The biological
10      activity of this oxidized phospholipid included: (1) decreased macrophage viability,
11      (2) induction of apoptosis in pulmonary epithelial-like A549 cells, (3) and release of IL-8 from
12      A549 cells.  Exposure levels of 0.125 ppm O3 for 2-4 h in this system were capable of
13      generating biologically active phospholipids that were capable of mediating toxic effects of O3.
14           In addition to PUFA,  cholesterol, the most abundant neutral lipid present in ELF, is also a
15      target of O3.  Pulfer and Murphy (2004) demonstrated the ozonolysis of cholesterol in an in vitro
16      system using BALF isolated from rats that had been exposed to 2.0 ppm O3 for 4 h. Production
17      of 5-hydroperoxy-J8-homo-6-oxa-cholestan-3,7a-diol, 5p,6p-epoxycholesterol, and 3p-hydroxy-
18      5oxo-5,6-seco-cholestan-6-al was shown. Additionally, both 5p,6p-epoxycholesterol and its
19      most abundant metabolite, cholestan-6-oxo-3p,5a-diol, were demonstrated to be cytotoxic to
20      16-HBE cells and to inhibit cholesterol synthesis.  Studies (Pulfer et al., 2005) in C57BL/6J mice
21      exposed to 0.5, 1.0, 2.0 or 3.0 ppm O3 for 3 h demonstrated that these oxysterols were produced
22      in vivo also.  The authors suggest that this may be  an additional mechanism of O3 toxicity.
23      Though these oxysterol reaction products have not been fully characterized, they may be
24      involved in O3-induced inflammation by disrupting cellular membranes or altering signaling
25      between cells. Similar oxysterols have been implicated in the inflammatory cascade associated
26      with atherosclerosis.
27           Thus, new work has attempted to elucidate the mechanisms by which reactions of O3 with
28      lipids create phospholipids  that then mediate downstream toxic effects.  It is uncertain whether
29      these described changes in lipid content and/or metabolism lead to significant changes in surface
30      tension or compliance properties of the lung.
31

        August 2005                               5-10        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.1.5  Ozone Interactions with Proteins and Effects on Protein Synthesis
 2           Epithelial lining fluid contains proteins arising from airway secretions and from blood.
 3      Ozone can react with four amino acid residues (cysteine, histidine, methionine, and tryptophan)
 4      and can cause oxidation of functional groups on proteins, including aldehydes, alcohol, amines
 5      and sulfhydryls. A number of enzymes have been shown to be inhibited by O3 including
 6      cholinesterase, al-antiproteinase, and prostaglandin synthetase.  Additionally, O3 decreases the
 7      inhibitory activity of al-proteinase inhibitor, which is implicated in development of emphysema.
 8      Surfactant protein A (SP-A) is a target for O3 toxicity by modulation of SP-A self association,
 9      vesicle aggregation, phospholipid secretion, and stimulation of AM superoxide anion generation
10      (see Section 5.2.2.3). Further, O3 is thought to interfere in SP-A's homeostatic role in surfactant
11      release from alveolar Type 2 cell lamellar bodies and its subsequent uptake by Type 2 cells
12      and AMs.
13           Lung collagen, collagen synthesis, and prolyl hydroxylase activity associated with
14      fibrogenesis have been shown to increase in rodents with O3 exposure of >0.45 ppm.  Some
15      studies have shown that this increase persists after exposure stops and that there is an influence
16      of exposure pattern on the response.  The increased collagen has been correlated with structural
17      changes in the lung. Rats exposed to an urban pattern of O3 with daily peaks of 0.25 ppm for
18      38 weeks displayed extracellular matrix thickening.  Increased levels of collagen in CAR were
19      demonstrated  in female rats  exposed to 0.5 to 1.0 ppm O3 for 6 h/day for 20 months and in
20      monkeys exposed to 0.61 ppm for 1 year. Both increased age and health status (e.g.,
21      emphysemic)  were implicated in the increased collagen formation in response to O3 exposure.
22           A time-course study (van Bree et al., 2001 Table AX5-5) evaluating the lung injury and
23      changes in collagen content  in rats exposed acutely or subchronically to 0.4 ppm O3
24      demonstrated  CAR thickening of septa which progressed from 7 through 56 days of exposure.
25      Though collagen content decreased with PE recovery, the structural fibrotic changes in ductular
26      septa and respiratory bronchioles persisted, suggesting that subchronic O3 exposures in rats
27      creates a progression of structural lung injury that can evolve to a more chronic form, which
28      included fibrosis.  The biological relevance and adverse health effects of altered protein
29      synthesis and  collagen accumulation  are uncertain.
30
        August 2005                               5-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.1.6  Differential Gene Expression
 2           Gohil et al. (2003) examined differential gene expression in C57BL/6 mice exposed to
 3      1 ppm O3 for three consecutive nights for 8 hours (see Table AX5-6). Ozone exposure induced
 4      changes in expression of 260 genes ( 80% repressed and 20% induced). Differentially expressed
 5      genes included those involved in progression of the cell cycle such as S-adenosyl methionine
 6      decarboxylase 3 (SAMD), ribonucleotide reductase (RR), and clusterin. Increased transcription
 7      of these genes suggests O3-induced activation of the cell cycle with subsequent cellular
 8      proliferation.  This is in accord with the finding of increased epithelial proliferation with acute
 9      O3  exposure as discussed in studies in Sections 5.2.4.1 and 5.2.5.1. Several NF-KB-induced
10      genes were upregulated, included serum amyloid protein, topoisomerase Hoc, monocyte
11      chemoattractant protein, platelet-derived growth factor, and inhibitor of apoptosis.  Upregulation
12      of these genes suggests to the authors that they may account for O3-induced proliferation of
13      nonciliated cells and Clara cells. Downregulation of transcripts for isoforms of myosins and
14      actins were also observed, which may explain, in part, a mechanism of O3-induced vascular
15      permeability.  Several members of the CYP family were downregulated, including 2a4,and  2el,
16      and 2f2, as were aryl-hydrocarbon receptor and several glutathione transferases. Metallothionein
17      1 and 2 and lactotransferrin were upregulated, indicative of their function as antioxidants and
18      anti-inflammatory agents. Ozone-induced suppression of immune function is suggested by
19      downregulation of transcripts encoding major histocompatibility complex genes,
20      lymphocyte-specific proteins, and immunoglobulins. Section 5.2.2.3 discusses the effects of Os
21      exposure on the immune system.
22           Quinlan et al.  (1994) have reviewed the regulation of antioxidant enzymes in lung after
23      oxidant injury. A comparison of alterations in gene expression in rat following O3 or hyperoxia
24      exposure, both of which induce reactive oxygen species and injury to vascular endothelial cells
25      and cells of the alveoli, show that both ~1 ppm O3 and 85-95% O2 increase expression of
26      CuZnSOD, glutathione peroxidase, and catalase.  Studies in mice (Johnston et al, 1998) also
27      demonstrate that changes in gene expression indicative of inflammation and epithelial injury that
28      occur with hyperoxia in mice (95% O2) compare to  similar injury that occurs following O3
29      exposure.
30
        August 2005                               5-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.1.7  Summary and Conclusions - Biochemical Effects
 2           Ozone has been shown to interact with a wide range of different cellular components
 3      including PUFAs, amino acid residues, and some low-molecular-weight compounds (GSH,
 4      urate, vitamins C and E). As O3 does not penetrate much beyond the ELF, damage likely results
 5      from its PUFA ozonation products (mostly hydroxyhydroperoxides) involvement in signaling of
 6      cellular responses such as inflammation.  New work has shown that ozonation of PUFA also
 7      forms the aldehydes nonanal, heptanal, and hexanal, the production of which is dependent on
 8      AA availability. Saturated phospholipids are thought to reduce the local dose and limit site-
 9      specific cell injury from O3 exposure. Another ozonation product FINE creates protein adducts
10      that have been linked to apoptosis and heat shock proteins in vitro.
11           Both short- and long-term exposures to O3 have been shown to enhance lung xenobiotic
12      metabolism, possibly as a result of changes in the number and function of bronchiolar epithelial
13      Clara cells and alveolar epithelial Type 2 cells. This modulation is both species- and region-
14      specific and includes the isoforms CYP 2B1, CYP 2E1. CCSP is also involved in inflammatory
15      responses to O3 exposure.  Mice strains with differing sensitivities to O3 show that responses in
16      protein, LDH and inflammatory cell influx are due to CCSP levels and changes in lung epithelial
17      permeability.
18           Reactions of O3 with AA, GSH, and UA (all antioxidants present in ELF) are a protective
19      mechanism. But even with environmentally relevant exposures, the reactivity of O3 is not
20      quantitatively quenched and cell injury occurs in both the lower and upper respiratory tract.
21      Early work has shown that short-term exposures to <1 ppm O3 increase antioxidant metabolism.
22      Re-exposure after a recovery period causes increases equivalent to first-time exposures,
23      suggesting that previous exposure is not protective. Increases in enzyme activity appear to
24      increase as a function of age, suggesting that O3 exposure can cause greater lung injury in the
25      older animal. Long-term urban patterns of exposure to O3 (daily peaks of 0.25 ppm) caused
26      increases in GSHPx and GR, but not SOD. Recent work has suggested that endogenous
27      antioxidants moderate the effects of O3 and that this exposure is a subthreshold stimulus for
28      producing effects on peripheral airway resistance and reactivity, thus indicating a dissociation
29      between airway function and inflammation.
30           In both acute and short-term studies, a variety of lung lipid changes occur with O3
31      exposure, including an increase in AA. With longer exposures (e.g., 0.12 ppm for 90 days),

        August 2005                              5-13         DRAFT-DO NOT QUOTE OR CITE

-------
 1      an increase in PUFAs and a decrease in cholesterol-esters are seen, indicative of long-term
 2      alterations of surfactant lipid composition. Whether these changes in lipid content and/or
 3      metabolism lead to significant changes in surface tension or compliance properties of the lung
 4      remains unknown. New studies evaluating O3-induced alterations in lipid metabolism have not
 5      been completed.
 6           Collagen, a structural protein involved in fibrosis, increases with O3 exposure, and some
 7      studies have shown that this increase persists after exposure stops. Urban patterns of exposure
 8      (daily peaks of 0.25 ppm for 38 weeks) created extracellular matrix thickening. Increases in
 9      centriacinar collagen were demonstrated in female rats exposed to 0.5 to 1.0 ppm O3 for 6 h/day
10      for 20 months and in monkeys exposed to 0.61 ppm for 1 year. New work examining the time
11      course of lung injury and changes in collagen content in rats exposed acutely or subchronically
12      to 0.4  ppm O3 showed centriacinar thickening of septa. Collagen content decreased with PE
13      recovery but not the structural fibrotic changes in ductular septa and respiratory bronchioles,
14      which suggests that subchronic O3 exposures in rats creates a progression of structural lung
15      injury that can evolve to a more chronic form, which includes fibrosis.
16
17      5.2.2    Lung Host Defenses
18          Defense mechanisms, including the mucociliary clearance system, AMs, and humoral- and
19      cell-mediated immune system, exist in the lung to protect it from infectious and neoplastic
20      disease and inhaled particles.  Summaries of key new animal studies examining the effects of O3
21      on lung host defenses are presented in Table AX5-7 of Annex AX5.  Acute human exposures
22      to O3 result in similar effects on AMs (see Chapter 6).
23
24      5.2.2.1  Clearance
25          Early studies of the effect of O3 on the mucociliary escalator showed morphological
26      damage to ciliated epithelial cells of the tracheobronchial tree at doses of <1 ppm.
27      Functionally, O3 slowed particle clearance in rats at doses of 0.8 ppm for 4 h and in rabbits at
28      0.6 ppm for 2 h exposures. Acute exposures at 0.5 ppm O3 in sheep caused increased basal
29      secretion of glycoproteins, while longer exposures reduced tracheal glycoprotein secretions,
30      both of which can alter the effectiveness of the mucociliary escalator. Early postnatal exposures
31      of sheep to 1 ppm O3 caused retardation of normal morphologic development of the tracheal

        August 2005                               5-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1      epithelium, decreased epithelial mucosa density, decreased tracheal mucous velocity, and
 2      delayed development of carbohydrate composition. Conversely, alveolar clearance in rabbits
 3      after acute exposure (0.1 ppm, 2 h/day, for 1 to 4 days) is increased.  Longer exposures showed
 4      no effect and increased O3 (1.2 ppm) slowed clearance. This pattern of clearance occurs in rats
 5      also. A study using rat tracheal explants exposed to O3 for 10 min (Churg et al., 1996) showed
 6      that uptake of TiO2 and asbestos was enhanced at 0.01 and 0.1 ppm, respectively. The authors
 7      attribute the increased uptake as a direct effect of O3, suggesting mediation by H2O2 or hydroxyl
 8      radical. Studies of the clearance of the radiolabled chelate 99mTc diethylenetriamine pentaacetic
 9      acid (Tc-DTPA) have shown that clearance is significantly increased following a 3 h exposure
10      to 0.8 ppm O3 in SD rats (Pearson and Bhalla,1997).  Examination of regional clearance
11      of 99mTc-DTPA in dogs following a 6 h isolated sublobar exposure to 0.4 ppm O3 or air showed
12      that O3 decreased the clearance halftime by 50% at 1 day following exposure (Foster and Freed,
13      1999).  Clearance was still elevated at 7 d PE but had recovered by 14 d. So, a  single local
14      exposure to O3 increases transepithelial clearance but without any influence on  contralateral
15      segments, i.e., only for epithelia directly exposed to O3.
16          Alveolar clearance is slower than tracheobronchial clearance and involves particle
17      movement through interstitial pathways to the lymphatic system or movement of particle-laden
18      AMs to the bottom of the mucociliary  escalator. Exposures of rabbits to 0.1 ppm accelerated
19      clearance while 1.2 ppm slowed clearance. A chronic exposure has been shown to slow
20      clearance.  New evaluations of the  effects of O3 on alveolar clearance have not been performed.
21
22      5.2.2.2  Alveolar Macrophages
23          A primary function of AMs is to clear the lung of infectious and noninfectious particles by
24      phagocytosis, detoxification, and removal. Further, AMs secrete cellular mediators that recruit
25      and activate inflammatory cells in the lungs (see Figure 5-3).  Ozone has been shown to inhibit
26      phagocytosis at 0.1 ppm for 2 h in rabbits. This inhibition returns to control levels if exposures
27      are repeated for several days. The  production of superoxide anion radicals and  the activity AM
28      lysosomal  enzymes (both involved in bactericidal activity) are inhibited by 3  h exposures to
29      0.4 and 0.25 ppm O3 in rodents and rabbits, respectively.  Production of IFNy was  decreased in
30      rabbit AM by 1 ppm O3 for 3 h.
        August 2005                               5-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1           New studies have shown that O3 affects AM chemotaxis, cell adhesion, and surface
 2      expression of cell adhesion molecules (Bhalla,  1996). AM from SD rats exposed to 0.8 ppm O3
 3      for 3 h showed greater mobility and greater adhesion than air exposed controls. This increased
 4      mobility and adhesion were attenuated by CD16b and ICAM-1 antibodies, suggesting these
 5      adhesion molecules modulate O3-induced inflammation.  Antibodies to TNFa and ILla also
 6      mitigated AM adherence, suggesting further that the inflammatory response to O3 is mediated by
 7      these cytokines (Pearson and Bhalla, 1997).  Cohen et al. (1996) showed that 1 ppm O3 for 4 h
 8      reduces binding of INF y to AM in WEHI-3 cells, and additionally reduces phagocytic activity,
 9      production of reactive oxygen intermediates, and elevation of intracellular Ca++.
10           Cohen et al. (2001, 2002) exposed male F-344 rats to either 0.1 or 0.3 ppm O3 for 4 h/day,
11      5 days/week or either 1 or 3 weeks. In this study, superoxide anion production was increased at
12      1 week.  Hydrogen peroxide production was reduced at both exposure concentrations and
13      durations and was further reduced with INFy stimulation, suggesting that one effect of O3 is
14      compromised killing of bacteria by AM due to the reduction in hydrogen peroxide production.
15           Ozone treatment (2 ppm O3, 3 h in female SD rats) caused a time-dependent increase in
16      NO levels in both AM and type II epithelial cells that was correlated with increased expression
17      of iNOS mRNA and protein (Laskin et al., 1998). Inhibition of NF-KB, caused a dose-dependent
18      inhibition of NO and iNOS production.  Additionally, O3 caused a time-dependent increase
19      in NF-KB binding activity in the nucleus of both cell types. The authors hypothesize that O3
20      exposure causes the cytokines TNFa and IL-lpa to bind to surface receptors and initiate
21      intracellular signaling pathways in AM leading to activation of NF-KB, its entry into the nucleus,
22      and its binding to the regulatory sequences of genes such as iNOS to allow their transcription.
23      Additional studies (Laskin et al., 2002) using AM isolated from C57B16xl29 mice with a
24      targeted disruption of the gene for iNOS  showed no toxicity to 0.8 ppm O3 for 3 h, as measured
25      by BALF protein levels and nitrotyrosine staining of the lung. Additionally, mice
26      overexpressing human Cu, Zn superoxide dismutase (SOD) and mice with a targeted disruption
27      of p50 NF-KB were also resistant to O3 toxicity. WT mice exposed to O3 showed an increase in
28      expression of STAT-1, a protein that binds to the regulatory region of iNOS.  Taken together,
29      these results suggest to the authors that a number of proteins including NF-KB, phosphoinoside
30      3-kinase, and STAT-1 that bind to and regulate expression of iNOS are modulated by O3
31      exposure. The same iNOS knockout mice strain exposed to 0.8 ppm O3 for 3 h (Fakhrzadeh

        August 2005                              5-16        DRAFT-DO NOT QUOTE OR CITE

-------
 1      et al., 2002) showed no increase in AM superoxide anion and prostaglandin.  These data provide
 2      further evidence the NO and its reactive oxidative product peroxynitrite are important in O3-
 3      induced lung injury. Further discussions of the role of nitric oxide synthase/reactive nitrogen
 4      and cytokines/chemokines in O3-induced inflammation are provided in Section 5.2.3.
 5
 6      5.2.2.3  Immune System
 7           Other than by natural protection (e.g., opsonizing antibody, nonspecific phagocytosis by
 8      AM), the immune system defends the lung by mounting three major waves of response:  natural
 9      killer (NK) cells (nonspecific lymphocytes that kill viruses, bacteria, and tumor cells), followed
10      by cytotoxic T-lymphocytes (TCTL- lymphocytes that lyse specifically recognized microbial  and
11      tumor-cell targets), followed by antigen-specific antibodies. These T-cell types are involved
12      with other immunologically active cells (e.g., B-cells and AM), which in a complex manner,
13      interact in  immunological defense. To date, only a few of these mechanisms have been
14      investigated in the context of their role in O3 susceptibility. The effects of O3 on the immune
15      system are complex and depend on the exposure parameters and observation periods.  T-cell-
16      dependent functions appear to be more affected than B-cell-dependent functions. Generally,
17      there is an early immunosuppressive effect that can, with continued exposure, either return to
18      normal or actually enhance immunity. Changes in immune cell population occur with O3
19      exposure including T:B-cell ratios in the MLN. Natural killer (NK) cell activity increases with
20      1 week exposures of 0.2 to 0.4 ppm O3 but decreases with exposures to 0.82 ppm.  Ozone
21      exposure has also shown to be responsible  for enhancement of allergic sensitization at levels of
22      0.5 to 0.8 ppm for 3 days.  Studies of the effects of O3 on the immune system are summarized in
23      Table AX5-7.
24           Garssen et al. (1997) have studied the effects of O3 on non-IgE-mediated pulmonary
25      hyper-immune reactions induced by picryl  chloride (PCI). BALB/c mice sensitized with PCI,
26      both actively and passively (by adoptive transfer of lymphoid cells from pre-sensitized mice),
27      were then challenged with picryl sulfonic acid (PSA). The mice were exposed to!2 h of 0.4, 0.8,
28      or 1.6 mg/m3 O3 during one night, at 4 days or 7 days after skin sensitization (which was either
29      just before or just after PSA challenge, i.e., during the induction or effector phase).
30      Nonsensitized mice showed no changes in  tracheal reactivity to carbacol with O3 exposure.
31      Sensitized mice were hyperreactive to carbachol 48 h after PSA challenge, whereas sensitized

        August 2005                              5-17         DRAFT-DO NOT QUOTE OR CITE

-------
 1      mice exposed to all concentrations of O3 showed no significant tracheal hyperreactivity to
 2      carbachol. The sensitized mice also demonstrated a suppressed inflammatory reaction (PMN)
 3      with 1.6 mg O3 exposure. Ozone exposure following PSA challenge also caused a suppression
 4      of tracheal hyperresponsiveness.  In a separate experiment wherein mice were exposed to O3
 5      before sensitization and then lymphoid cells from these mice were injected into nonexposed
 6      mice, the  recipients also demonstrated an inhibition of the induction of hyperreactivity. These
 7      results are opposite to the effect on type I (IgE-mediated) allergic reactions, which the authors
 8      suggest is due to activation of Th-2 cell-dependent reactions that are possibly potentiated by O3
 9      or to a direct effect by O3 on Th-1 cells or other cells that are crucial for the tracheal
10      hyperreactivity and inflammation seen in this mouse model.
11           Kleeberger et al. (2000, 200la) have demonstrated a potential interaction between the
12      innate and acquired immune system with O3 exposure.  Using O3-susceptible (C57BL/6J)
13      and O3-resistant (C3H/HeJ) mice, they identified a candidate gene on chromosome 4, Toll-like
14      receptor 4 (Tlr4). Ozone exposure (0.3 ppm for 24 to 72 h)  of C3H/HeJ and C3H/HeOuJ mice,
15      the latter differing from the O3-resistant strain by a polymorphism in the coding region of Tlr4,
16      demonstrated greater protein concentrations in the OuJ strain. The two strains exhibited
17      differential expression of Tlr4 mRNA with O3 exposure.  Thus, a  quantitative trait locus on
18      chromosome 4 appears to be responsible for a significant portion  of the genetic variance in
19      O3-induced lung hyperpermeability. In these mouse strains  lavageable protein concentration was
20      lowered by inhibition of inducible nitric oxide synthase (iNOS) and by targeted disruption of
21      Nos2. Comparisons of C3H/HeJ and C3H/HeOuJ O3 exposures demonstrated reduced Nos2 and
22      Tlr4 mRNA levels in the O3-resistant C3H/HeJ mice. These data are consistent with the
23      hypothesis that O3 -induced lung hyperpermeability is mediated by iNOS. These studies suggest
24      a role for  TLR4 in the host response to O3 similar to the role it has demonstrated in
25      lipopolysaccharide (LPS) sensitivity (Schwartz 2002; Wells et al. 2003).  TLR4 signaling is
26      thought to be critical to linking the innate and acquired immune system through antigen
27      presenting cells and Thl/Th2 differentiation.
28           Ozone exposure has been shown to affect Ig responses both in vitro and in mice. Becker
29      et al. (1991) demonstrated changes in IgG production in cultured human lymphocytes with O3
30      exposures of 1.0, 0.5, and 0.1 ppm for 2 h.  Subsequent to O3 exposure, cells were stimulated
31      with pokeweed mitogen (PWM, a T-cell-dependent stimulus) or Staphylococcus aureus Cowan 1

        August 2005                              5-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1      strain (SAC, a T-cell-independent stimulus). Both B and T cells were affected by O3.
 2      T cells also demonstrated an increase in IL-6 and a decrease in IL-2, which suggested to the
 3      authors that O3 may have direct effects on IgG producing cells and concurrently an effect that
 4      is mediated by altered production of T cell immunoregulatory molecules. Responses to
 5      repeated O3 (0.08 - 0.25 ppm) and OVA (1%) exposures were compared in "IgE-high responder"
 6      (BALB/c) and "IgE-low responder" (C57BL/6) mice (Neuhaus-Steinmetz et al., 2000). Ozone
 7      appeared to shift the immune response toward a Th2-like pattern in the two mouse strains with
 8      differing potentials for developing allergic reactions.
 9          Another study (Depuydt et  al., 2002) demonstrated that O3 (0.1 ppm for 2 h) increases
10      allergen-induced airway inflammation in previously sensitized mice but has no effect on the
11      sensitization process itself.  This  study uses OVA-pulsed dendritic cells instead of systemic
12      adjuvant, which the authors consider a more relevant model of sensitization as it clearly
13      separates the immune response from the challenge and does not obscure regulatory processes as
14      does i.p. injections of OVA.  They further suggest that dendritic cells, the principal antigen-
15      presenting cells in the airway, are an important component of O3-induced eosinophilic airway
16      inflammation.
17          Surfactant protein A and D  (SP-A and SP-D) were shown to create  an inflammatory
18      feedback loop with perturbations in lung immune defenses (reviewed in Hawgood and Poulain,
19      2001). Earlier studies suggested  that SP-A is a target for O3 toxicity by causing inhibition of
20      SP-A self-association and SP-A-mediated lipid vesicle aggregation. Further, O3 reduced the
21      ability of SP-A to inhibit phospholipid secretion by alveolar type II cells  and reduced the
22      capacity of SP-A to induce superoxide anion production and enhance phagocytosis of herpes
23      simplex virus. Bridges et al. (2000) reported that both SP-A and  SP-D directly protect surfactant
24      phospholipids and macrophages from oxidative damage by blocking accumulation of TEARS
25      and conjugated dienes.
26          Eight human variants of SP-A in CHO cells exposed to O3 (Ippm for 4 h) showed
27      decreased ability to stimulate cytokine (TNF-a and IL-8) production in THP-1 cells, a
28      macrophage-like cell line (Wang  et al., 2002).  Each variant had a unique time- and
29      dose-dependent pattern of stimulation of cytokine production with O3 exposure which the
30      authors attribute to possible differences in susceptibility to O3 oxidation.  Targeted disruption of
31      mouse SP-A and SP-D (Hawgood et al, 2002)  caused increases in BAL phospholipid,

        August 2005                              5-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1      macrophage, and protein through 24 weeks of age. Further, the deficient mice developed patchy
 2      lung inflammation and air space enlargement consistent with emphysema.  Future experiments
 3      using these null mice will help to establish the role of SP-A and SP-D in pulmonary host defense
 4      to O3 exposure.
 5
 6      5.2.2.4  Interactions with Infectious Microorganisms
 7           Ozone-induced dysfunction of host defense systems results in enhanced susceptibility to
 8      bacterial lung infections. Acute exposures of 0.08 ppm (3 h) O3 can overcome the ability of
 9      mice to resist infection (by decreasing lung bactericidal activity) with Streptococcal bacteria,
10      resulting in mortality. Changes in antibacterial defenses are dependent on exposure regimens,
11      species and strain of test animal, species of bacteria, and age of animal, with young mice more
12      susceptible to the effects of O3. The effect of O3 exposure on antibacterial host defenses appears
13      to be concentration- and time-dependent. Early studies using the mouse "infectivity model,"
14      consisting of exposure to clean air or O3 followed by exposure to an aerosolized microorganism,
15      showed that the difference in mortality between O3-exposed groups and controls is
16      concentration-related. Chronic exposures (weeks, months) of 0.1 ppm do not cause greater
17      effects on infectivity than short exposures, due to defense parameters becoming reestablished
18      with prolonged exposures.
19          More recent studies of O3-induced modulation of cell-mediated immune responses showed
20      effects on the onset and persistence of infection.  Cohen et al. (2001, 2002) exposed male F-344
21      rats subchronically to either 0.1 or 0.3 ppm O3 for 4 h/day 15 days/week, for 1 or 3 weeks.
22      Subsequent exposure with viable Listeria monocytogenes demonstrated no observed effect on
23      cumulative mortality but did show a concentration-related effect on morbidity onset and
24      persistence.  These data suggest that O3 may cause a possible imbalance between Th-1 and Th-2
25      cells, which can subsequently lead to suppression of the resistance to intracellular pathogens.
26          Effects of O3 on viral infections are dependent on the temporal relationship between O3
27      exposure and viral infection.  Only high concentrations (1.0 ppm O3, 3 h/day, 5 days, mice)
28      increased viral-induced mortality. No detrimental effects were seen with a 120-day exposure to
29      0.5 ppm O3 on acute lung injury from influenza virus administered immediately before O3
30      exposure started. But there were O3-enhanced postifluenzal alveolitis and lung parenchymal
31      changes.  As O3 does not affect lung influenza viral liters, it apparently does not impact antiviral

        August 2005                               5-20        DRAFT-DO NOT QUOTE OR CITE

-------
 1      clearance mechanisms. In general, the evidence suggests that O3 can enhance both bacterial and
 2      viral lung infections, but the key mechanisms have not yet been identified. New studies on the
 3      interactions of O3 and viral infections have not been published.
 4
 5      5.2.2.5  Summary and Conclusions - Lung Host Defenses
 6          New data on lung host defenses support earlier work which suggests that mucociliary
 7      clearance is affected in most test species at just under 1 ppm, with lower levels (~0.1 ppm)
 8      increasing clearance and somewhat higher levels decreasing clearance. These data also propose
 9      mechanisms whereby O3 affects clearance, which include uptake being a direct effect of O3, but
10      modulated by ROS and hydroxyl radicals.
11          Alveolar macrophage function is disrupted by O3 as shown by a number of studies
12      demonstrating inhibition of phagocytosis at concentrations ranging from 0.1 to 1.2 ppm. This
13      inhibition returns to control levels if exposures are repeated for several days. Two new studies
14      corroborate earlier findings of increases in AM number in that same exposure range. In this
15      environmentally relevant exposure range, new studies support older findings of decreased
16      resistance to microbial pathogens as shown by the endpoints examining superoxide radical
17      formation, altered chemotaxis/motility, decreased INFy levels, decreased lysosomal activity,
18      increased PGE levels, and increased NO mRNA and protein.
19          New research evaluating the effects of O3 on immune function advances previous work that
20      has shown that exposures can enhance or suppress immune responsiveness depending on the
21      species studied, concentration of O3, route of exposure of allergen, and timing of exposure.
22      Continuous exposure to O3 impairs immune responses for the first several days of exposure,
23      followed by an adaptation to O3 that allows a return of normal immune responses.  Most species
24      show little effect of O3 exposures prior to immunization, but a suppression of responses to
25      antigen in O3 exposures post-immunization.  The use of mouse strains with genetically
26      determined sensitivity or resistance to O3 indicated a possible interaction between the innate and
27      acquired immune system, and further, that O3 may shift the immune response towards a Th-2-
28      like pattern. Work has also focused the deleterious effects of O3 exposure on SP-A and SP-D
29      and their immunomodulatory function in protecting against oxidative stress.
30           Several new studies evaluating the effects of O3 exposures on infectious microorganisms
31      are in concurrence with previous studies which showed, in general, increased mortality and

        August 2005                              5-21        DRAFT-DO NOT QUOTE OR CITE

-------
 1      morbidity, decreased clearance, increased bacterial growth, and increased severity of infection at
 2      exposure levels of 0.1 to 1 ppm O3 for 1 week.
 3
 4      5.2.3   Inflammation and Lung Permeability Changes
 5           The normal lung has an effective barrier function that controls bidirectional flow of fluids
 6      and cells between the air and blood compartments. Ozone disrupts this function, resulting in two
 7      well-characterized effects of O3 exposure, lung inflammation and increased permeability, which
 8      are distinct events controlled by independent mechanisms. Ozone initiates inflammation of lung
 9      tissue by reactions with antioxidants and lipids in ELF (discussed in 5.2.1, see Figure 5-2).
10      Secondary reaction products generated in this process then cause changes in cell membranes,
11      disruption of the lung barrier leading to leakage of serum proteins, influx of polymorphonuclear
12      leukocytes (PMNs), release of bioactive mediators, and movement of compounds from the
13      airspaces into the blood. This increased permeability allows accumulation of co-occurring
14      pollutants into the lung tissue.  The framework for presenting this stereotypical response to Os
15      consists of discussions covering: 1) the time course of these changes; 2) concentration x time
16      (C x T) relationships; 3) susceptibility factors; 4) mediators of inflammation;  and 5) nitric oxide
17      and reactive nitrogen.
18           Rats appear to be more resistant to Os-induced inflammation than humans (see Chapter 4).
19      With comparable exposure protocols, both species have similar observed inflammatory and
20      permeability changes, i.e., controlled human exposure studies discussed in Chapter 6 indicate
21      that the majority of acute responses in humans are similar to those observed in animals.
22           Ozone also increases the permeability from the air to the blood compartment. Ozone
23      (0.8 ppm; 2 h) caused a 2-fold increase of the transport of labeled DTPA from the rat tracheal
24      lumen to the blood.  This coincided with a 2-fold increase in the number of endocytic vesicles in
25      epithelial cells that contained intraluminally instilled HRP as a tracer. These studies also suggest
26      an uneven disruption of tight junctions and alternate transport through endocytotic mechanisms.
27      In studies aimed at detecting the effects of O3 exposure on regional permeability, O3 increased
28      the transmucosal transport of DTPA and BSA more in the trachea and bronchoalveolar zone than
29      in the nose. These changes in barrier integrity may allow increased entry of antigens and other
30      bioactive compounds (e.g., bronchoconstrictors) into lung tissue. Data from analyses at regular
        August 2005                              5-22        DRAFT-DO NOT QUOTE OR CITE

-------
 1      intervals PE indicate that maximal increases in BALF protein, albumin and number of PMNs
 2      occur 8 to 18 h (depending on the study) after an acute exposure ceases.
 3          Increases in permeability and inflammation have been observed at levels as low as
 4      0.1 ppm O3 for 2 h/day for 6 days in rabbit and 0.12 ppm in mice (24 h exposure) and rats
 5      (6 h exposure). After acute exposures, the influence of the time of exposure increases as the
 6      concentration of O3 increases.  The exact role of inflammation in causation of lung disease is not
 7      known, nor is the relationship between inflammation and changes in lung function. Table
 8      AX5-8 in Annex AX5 summarizes new key studies describing the potential for O3 exposure
 9      effects on lung permeability and inflammation.
10
11      5.2.3.1   Time Course of Inflammation and Lung Permeability Changes
12          The maximal increase in BALF protein, albumin, and PMN occurs in most species 8  to
13      18 h after the cessation of acute exposures of 0.5 to 1.0 ppm.  A study  of O VA-sensitized male
14      Dunkin-Hartley guinea pigs exposed to 1.0 ppm O3 for 3 h showed that levels of PMN
15      significantly increased at 3 h PE, but BAL protein levels did not, suggesting a lack of correlation
16      between the two endpoints (Sun et al., 1997).  Increased PMN without a concordant increase in
17      BAL protein levels were found when the guinea pigs were exposed to  1.0 ppm O3 for 1 h and
18      evaluated 24-h PE.  The first group also had an increase in AHR, but not the second group,
19      which suggests a dissociation between PMN levels and AHR.
20          Earlier work demonstrated that O3 exposures of 0.8 to 1 ppm in rat and guinea pig
21      transiently increase the permeability from the air to the blood compartment. This permeability is
22      greatest in trachea and bronchoalveolar zone, and may allow increased entry of antigens and
23      other bioactive compounds (e.g., bronchoconstrictors)  into lung tissues. The time course of the
24      influx of PMNs into the lung and the BALF fluid levels of macrophage inflammatory protein-2
25      (MIP-2) were found to be roughly  similar to that for proteins (Bhalla and Gupta, 2000).
26      Adherence of neutrophils to pulmonary vascular endothelium is maximal within 2 h after
27      exposure and returns to control levels by 12 h PE (Lavnikova et al., 1998).  Cheek et al. (1995)
28      cultured monolayers of rat alveolar type II cells and exposed them to 0.1  or 0.5 ppm O3 for 0.5 h
29      to evaluate the effects of O3 on permeability. Permeability increased dose-dependently and the
30      higher exposures elicited greater numbers of injured epithelial cells. Exposure to 0.1 ppm  O3
31      was thought to expedite the restoration of epithelial barrier functions, while in higher exposures,

        August 2005                              5-23        DRAFT-DO NOT QUOTE OR CITE

-------
 1      neutrophils exacerbated the O3-induced injury.  Vesely et al. (1999a) have demonstrated that
 2      neutrophils contribute to the repair process in O3-injured airway epithelium and they may play a
 3      role in removal of O3-injured cells.
 4           Exposures of 3 to 7 days have been found to cause increases in BALF protein and PMNs
 5      that typically peak after a few days (depending upon species tested and exposures) and return
 6      towards control even with continuing exposure. Van Bree et al. (2002) observed lower
 7      BALF levels of protein, fibronectin, IL-6 and inflammatory cells in rats exposed for 5 days to
 8      0.4 ppm O3 than in rats exposed for 1 day, suggesting adaptation to O3 exposure.  Postexposure
 9      challenge with single O3 exposures at different time points showed recovery of susceptibility
10      to O3. McKinney et al. (1998) observed differences in IL-6 levels due to repetitive exposures
11      and demonstrated a role of IL-6 in the adaptive  response induced by repeated O3 exposures of
12      0.5 ppm for 4 h.
13
14      5.2.3.2  Concentration and Time of Exposure
15           The relative influence of concentration and duration of exposure (i.e., C x T) has been
16      investigated extensively in rats,  using BALF protein as an endpoint. Earlier work utilizing
17      concentrations of 0.1 to 2 ppm O3 and durations of 1 to 8  h has shown that the interaction
18      between C and T is  complex. At these levels of exposure, concentration generally dominates the
19      response.  C x T studies using the endpoints of changes in lung protein or cell type showed that
20      acute damage is a function of cumulative dose.  The impact of T is C-dependent (at higher Cs,
21      the impact of T is greater); at the lowest C and T values, this dependence appears to be lost. The
22      controlled human exposure data described in Chapter 6 concur with most animal data, showing
23      that concentration of O3 is the most important factor determining O3 responses, and that duration
24      of exposure and ventilation rate are secondary factors.
25           New studies evaluating C  x T relationships in animal models have not been completed.
26      However, a full understanding of C x T relationships in ambient exposures must include the
27      recognition that 'real world' exposures are cyclic in nature, due to the daily and seasonal
28      variations in O3 levels. The concentration of O3, the duration of the exposure, and duration
29      between exposures are all  relevant to the type and level of Cyinduced injury.
30
        August 2005                              5-24        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.3.3  Susceptibility Factors
 2           Factors that have been studied for potential impact on the effects of O3 exposure include
 3      age, gender, nutritional status, exposure to co-pollutants, exercise, and genetic variability. A full
 4      characterization of the effects of age on O3 responses has not been completed. Data available
 5      indicate that effects of age on O3 responses are endpoint-dependent, with young mice, rats, and
 6      rabbits having greater prostaglandin levels with exposure and senescent rats having greater IL-6
 7      and N-acetyly-p-D-glucosaminidase levels with exposure.
 8           A study (Johnston et al., 2000a) compared gene expression of chemokines and cytokine in
 9      newborn and 8-week-old C57B1/6J mice exposed to 1.0 or 2.5 ppm for 4, 20, or 24 h. The
10      newborn mice displayed increased levels of Mt mRNA only, while the 8-week-old mice had
11      increases in MIP-1 a, MIP-2, IL-6, and Mt mRNA. Comparisons were made with mice of the
12      same age groups with exposures to endotoxin (10 ng/mouse for 10 min). Both age groups
13      displayed similar cytokine/chemokine profiles with endotoxin exposure. This suggested to the
14      authors that the responses to endotoxin, which does not cause epithelial  injury, and the responses
15      to O3, which does, demonstrate that differences in inflammatory control between newborn and
16      adult mice is secondary to epithelial injury.
17           Pregnancy and lactation increased the susceptibility of rats to acute O3, but no clear  effects
18      of gender have been identified.  The effects of vitamin C deficiency on O3 responses are unclear.
19      Ascorbate-deficient guinea pigs exposed to O3 demonstrated only minimal effects on injury and
20      inflammation (Kodavanti et al.,  1995). Utilizing a diet-restricted (20% of the freely-fed diet) rat
21      model, Elsayed (2001) demonstrated higher survivability on exposure to higher O3 (0.8 ppm
22      continuously for 3 days) compared to freely-fed rats.  Pre-exposure to sidestream cigarette
23      smoke had been found to cause increased lung injury (Yu et al., 2002).  In vitro  studies on the
24      macrophages from smoke + O3- exposed animals responded by a greater release of TNF-a
25      following LPS stimulation when compared to macrophages exposed to air, smoke or O3
26      (0.5 ppm, 24 h) alone.
27           Lines of evidence illustrate that genetic background is an extremely important determinant
28      of susceptibility to O3. Earlier studies using inflammation-prone (susceptible) C57BL/6J (B6)
29      and inflammation-resistant C3H/HeJ (C3) mouse strains and high doses of O3 (2 ppm for 3 h)
30      identified Inf-2 as a locus controlling susceptibility. Further studies in these two strains of mice
31      using more relevant exposures (0.3 ppm for 72 h) identified that the acute and subacute

        August 2005                              5-25        DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposures are controlled by two distinct genes, referred to as Inf-l and Inf-2, respectively
 2      (Tankersley and Kleeberger,  1994). Exposures to 0.3 ppm O3 for 48 or 72 h, when repeated
 3      fourteen days after the initial exposures, caused a smaller increase in BALF protein and number
 4      of macrophages, lymphocytes and epithelial cells in both strains, but PMN number was greater
 5      in both strains compared to initial exposure (Paquette et al., 1994).  Kleeberger et al. (1997) also
 6      identified another potential susceptibility gene, tumor necrosis factor (Tnf, which codes for the
 7      pro-inflammatory cytokine TNF-a ) on a qualitative trait locus on mouse chromosome 17.
 8      By neutralizing the function of TNF-a with a specific antibody,  they were able to confer
 9      protection against O3 (0.3 ppm, 48 h) injury in susceptible mice. The group then demonstrated a
10      role for TNF receptor 1 and 2 (TNFR1 and TNFR2, respectively) signaling in subacute (0.3 ppm
11      for 48 h) O3-induced pulmonary epithelial injury and inflammation (Cho et al., 2001).  TNFR1
12      and TNFR2 knockouts were less sensitive to subacute O3 exposure than WT C57BL/6J mice.
13           An integrated and more comprehensive effort to identify the genetic basis for the
14      susceptibility to O3-induced lung injury was reported by Savov et al. (2004). In this report, acute
15      lung injury to high dose of O3 (2 ppm for 3 h) was assessed and  integrated with physiological,
16      biochemical, and genetic observations using 9 inbred mouse strains. This work indicated the
17      presence of genetic loci on chromosomes 1, 7, and 15 associated with phenotypic
18      characteristics for resistance to acute O3-induced lung injury.  They identified C3H/HeJ and
19      A/J as consistently O3-resistant, C57BL/6J and 129/SvIm as consistently O3-vulnerable, and
20      CAST/Ei, BTBR, DBA/2J, FVB/NJ, and BALB/cJ as intermediate in response to O3.
21           Ozone-induced changes in CCSP (called CC16 by this group) expression were evaluated in
22      five inbred mouse strains:  C57BL/6J and CBA both considered sensitive to acute O3-induced
23      inflammation, C3H/HeJ and AKR/J both considered resistant, and SJL/J considered intermediate
24      (Broeckaert et al., 2003).  Two exposures paradigms were used,  1.8 ppm O3 for 3 h or
25      0.11 ppm O3, 24/h day for up to 3 days, and BALF and serum was assayed immediately after
26      exposure or at 6 h PE. Both exposure levels caused a transient increase in CC16 in serum that
27      correlated with BALF changes in protein, LDH, and inflammatory cells.  There was an inverse
28      relationship between preexposure levels of CC16 in BALF and epithelial damage based on
29      serum CC16 levels and BALF markers of inflammation.  There was also an inverse relationship
30      between preexposure levels of albumin in BALF and lung epithelium damage. Based on these
31      results, the authors conclude that a major determinant of susceptibility to O3 is basal lung

        August 2005                              5-26        DRAFT-DO NOT QUOTE OR CITE

-------
 1      epithelial permeability. As all of the mouse strains had similar levels of preexposure CC16
 2      mRNA, they explored the possible role of CC16 isozymes in differences among strains.  The
 3      CC16 monomer a 7kD protein exists in two isoforms with differing pi values, CC16a (4.9) and
 4      CC16b (5.2). To evaluate the role of CC16 isoform profiles in permeability differences between
 5      C57BL/6J and C3H/HeJ, this group evaluated the CC16 protein profiles in BALF of the strains
 6      before and after O3 exposure following two-dimensional protein electrophoresis analysis.
 7      C57BL/6J mice had lower levels of CC16a (the more acidic form) than C3H/HeJ. But both the
 8      strains had similar levels of CC16b.  Based on these observations, Broeckaert et al (2003)
 9      conclude that greater epithelial permeability observed in C57BL/6J may be due to difference in
10      the expression of CC16a and possibly other antioxidant/inflammatory proteins.
11           Wattiez et al. (2003)  examined BALF protein from C57BL/6J (O3-sensitive) and
12      C3FI/HeJ (O3-resistant) mice exposed to filtered air using a two-dimensional polyacrylamide
13      gel approach to analyze the protein profiles. C3H/HeJ mice expressed 1.3 times more Clara cell
14      protein!6 (CC16) than C57BL/6J mice and, further, expressed more of the acidic  isoform of
15      CC16. Strain-specific differential expression of isoforms of the antioxidant protein 2 (AOP2),
16      the isoelectric point 5.7 isoform in C3H/HeJ and isoelectric point 6.0 isoform in C57BL/6J were
17      observed. These data suggest a protective role for CCSP against oxidative damage, and further,
18      that genetic  susceptibility to oxidant stress may be moderated, in part, by the gene coding for
19      CCSP. Taken together, these mouse studies of genetic susceptibility are useful for
20      understanding underlying mechanisms leading to O3-induced effects. However, at this point,
21      corresponding human polymorphisms have not yet been identified which associate with differing
22      human sensitivities to O3.
23
24      5.2.3.4  Mediators of Inflammatory Response and Injury
25           Ozone reacts with lipids in the ELF or epithelial cell membranes, creating ozonation
26      products which then stimulate airway epithelial cells, AMs, and PMNs to release a host of
27      pro-inflammatory mediators including cytokines, chemokines, reactive oxygen species,
28      eicosanoids, and platelet activating factor (see Figure 5-3). While neutrophils in the lung
29      characterize an inflammatory response to O3, the release of chemotactic mediators by
30      inflammatory cells indicates their state of activation and their role in continued inflammation
31      and injury.  At O3 exposures  of > 1 ppm, these mediators recruit PMN, and increase expression of

        August 2005                              5-27        DRAFT-DO NOT QUOTE OR CITE

-------
 1     MIP-2 mRNA or BALF levels of MIP-2 (Driscoll et al., 1993; Haddad et al., 1995; Bhalla and
 2     Gupta, 2000). The increased mRNA expression was associated with an increased neutrophilia in
 3     the lung. Zhao et al. (1998) showed that 0.6 ppm O3 exposure for 2 h in mice and rats causes an
 4     increase in monocyte chemotactic protein-1 (MCP-1).
 5           Fibronectin, an extracellular matrix glycoprotein, is thought to have a role in lung
 6     inflammation and inflammatory disorders, and has shown to be  increased with exposure to
 7     1 ppm O3 for 14 days.  Gupta et al. (1998) observed an increase in both fibronectin protein and
 8     mRNA expression in the lung of rats exposed to 0.8 ppm O3 for 3 h. A mechanistic role of
 9     fibronectin in O3-induced inflammation and injury was suggested on the basis of comparability
10     of temporal changes in BALF protein, fibronectin and alkaline phosphatase activity with
11     exposures of 1 ppm for 3 h (Bhalla et al.,  1999).  Studies have reported an effect of O3 on other
12     cytokines and inflammatory mediators. An increase occurred for cytokine-induced neutrophil
13     chemoattractant (CINC) and NF-KB expression in vivo (Koto et al., 1997), for IL-8 in vivo and
14     in vitro (Chang et al., 1998), TNFa, fibronectin, IL-1 and CINC release by macrophages ex vivo
15     (Pendino et al., 1994; Ishii et al., 1997), and NF-KB and TNFa (Nichols et al., 2001; see 6.9.2).
16     An increase in lung CINC mRNA occurred within 2 h after the end of a 3 h exposure of rats to
17     1 ppm O3.  The CINC mRNA expression  was associated with neutrophilia at 24 h PE. Exposure
18     of guinea pig AMs recovered in BALF and exposed in vitro to 0.4 ppm O3 for 1 h produced a
19     significant increase  in IL-6 and TNFa (Arsalane et al., 1995). An exposure of human AMs to an
20     identical O3 concentration increased TNFa, IL-lp, IL-6 and IL-8 and their mRNAs. Ozone
21     exposure (0.3 to 2.5 ppm, 1-48 h) of mice caused an increase in IL-6, MlP-la, MIP-2,
22     eotaxin and Mt abundance (Johnston et al., 1999a). The IL-6 and Mt increase was enhanced in
23     mice deficient in CCSP, suggesting a protective role of Clara cells and their secretions (Mango
24     et al., 1998).  CCSP deficiency, also increased sensitivity of mice to O3, as determined by an
25     increase in abundance of MlP-la and MIP-2 following a 4 h exposure to 1.0 ppm O3
26     (Johnston et al., 1999b).
27           Mast cells, which are located below the  epithelium, release proinflammatory mediators and
28     have been shown to contribute to O3-induced  epithelial damage.  Greater increases in lavageable
29     macrophages, epithelial cells and PMNs were observed in mast  cell-sufficient mice than in mast
30     cell-deficient mice exposed to 0.26 ppm O3 for 8 h per day, 5 days per week (Kleeberger et al.,
31     200Ib). Increases in inflammatory cells were also observed in mast cell-deficient mice repleted

       August 2005                               5-28        DRAFT-DO NOT QUOTE OR CITE

-------
 1      of mast cells, however O3-induced permeability changes were similar between genotypic groups
 2      exposed to 0.26 ppm. When a RBL-2H3 mast cell line was exposed to 0.1 to 1.0 ppm O3 for 1 h,
 3      spontaneous release of serotonin and modest generation of prostaglandin D2 occurred only under
 4      conditions that caused cytotoxicity (Peden and Dailey, 1995). Additionally, O3 inhibited IgE-
 5      and A23187-induced degranulation. Mast cells recovered from O3-exposed peripheral airways
 6      of ascaris sensitive dogs released significantly less histamine and PGD2 following in vitro
 1      challenge with ascaris antigen or calcium ionophore (Spannhake, 1996).  Ozone (0.4 ppm,
 8      5 weeks) exposure also promoted eosinophil recruitment in the nose and airways in response to
 9      instillation of OVA or OVA-pulsed dendritic cells and aggravated allergy like symptoms in
10      guinea pigs (lijima et al., 2001).
11           The role of PMNs and cellular mediators in lung injury and epithelial permeability has
12      been investigated using antibodies and inhibitors of known  specificity to block inflammatory cell
13      functions and cytokine activity. Treatment of rats with cyclophosphamide prior to O3 exposure
14      (0.8 ppm, 48 h) resulted in a decreased recovery of PMNs in the BALF and attenuated
15      permeability induced by O3 (Bassett et al., 2001).
16           Pretreatment of animals with antiserum against rat neutrophils abrogated PMN
17      accumulation in the lung, but did not alter permeability increase produced by O3.  Studies
18      utilizing antibodies to selected pro- or anti-inflammatory cytokines suggest a role of TNFa,
19      IL-10, and IL-lp in O3-induced changes in permeability, inflammation and cytokine release
20      (Ishii et al., 1997; Reinhart et al., 1999; Bhalla et al., 2002)  in exposures of ~ 1 ppm for 3-6 h.
21      An attenuation of O3-induced increases in permeability and inflammation was also observed in
22      mice treated, either before or after exposure, with UK-74505, a platelet-activating factor (PAF)
23      receptor antagonist (Longphre et al., 1999).  These results were interpreted to indicate that
24      O3-induced epithelial and inflammatory changes are mediated in part by activation of PAF
25      receptors.
26           Ozone exposure stimulates macrophage motility towards a chemotactic gradient, and
27      macrophages isolated from rats exposed to 0.8 ppm O3 for 3 h adhered to epithelial cells
28      (ARL-14) in culture to a greater extent than macrophages from air-exposed controls (Bhalla,
29      1996). Both macrophage motility and chemotaxis were attenuated by antibodies to cell adhesion
30      molecules CD-I Ib and ICAM-1, suggesting a role for cell adhesion molecules in O3-induced
31      cellular interactions. This may also explain the increased tissue localization and reduced

        August 2005                              5-29         DRAFT-DO NOT QUOTE OR CITE

-------
 1      recovery of macrophages in BALF (Pearson and Bhalla, 1997) following O3 exposure (0.8 ppm,
 2      3 h).  Studies investigating the mechanisms of PMN recruitment in the lung have explored the
 3      role of cell adhesion molecules that mediate PMN-endothelial interactions. An exposure of
 4      female rats to O3 (1 ppm, 2 h) had an attenuating effect on CD-18 expression on AMs and
 5      vascular PMNs, but the expression of CD62L, a member of selection family,  on vascular PMNs
 6      was not affected (Hoffer et al., 1999). In monkeys, O3-induced  (0.8 ppm, 8 h) inflammation was
 7      blocked by treatment with a monoclonal antibody to CD 18, suggesting dependence of PMN
 8      recruitment on this adhesion molecule (Hyde et al., 1999).  Treatment of monkeys with CD18
 9      antibody also reduced tracheal expression of the P6 integrin (Miller et al., 2001) suggesting that
10      lung epithelial cell expression of this adhesion molecule is  associated with sites of neutrophil
11      recruitment. A single 3 h exposure of rats to 1 ppm O3 caused an elevation in concentration of
12      ICAM-1, but not CD-18, in the BALF (Bhalla and Gupta, 2000).  Takahashi et al. (1995a) found
13      an increase in tissue expression of ICAM-1 in mice exposed to 2 ppm O3 for 3 h, noting a
14      temporal correlation of inflammatory  activity and ICAM-1  expression which varied in different
15      regions of the lung. A comparable pattern of time-related changes in total protein, fibronectin
16      and alkaline phosphatase activity in the BALF of rats exposed to 0.8 ppm O3  for 3 h was also
17      noted by Bhalla et al. (1999).  Together, these studies support the role of extracellular matrix
18      protein and cell adhesion molecules in the induction of lung inflammation and injury.
19
20      5.2.3.5  The Role of Nitric Oxide Synthase and Reactive Nitrogen in Inflammation
21          Nitric oxide (NO) is a messenger molecule involved in many biological processes,
22      including inflammation (see Figure 5-3). Cells in the respiratory tract (including mast cells,
23      neutrophils, epithelial cells, neurons, and macrophages) produce three differing forms of nitric
24      oxide synthase (NOS), the enzyme that catalyzes the formation  of NO. NOS-l(neuronal) and
25      NOS-3 (endothelial) are constitutively expressed, whereas NOS-2 (also referred to as iNOS) is
26      inducible, commonly by pro-inflammatory cytokines. Macrophages isolated  from O3-exposed
27      (0.8 ppm for 3 h) mice produced increased amounts of NO, superoxide anion, and PGE2,
28      but production of these mediators by macrophages from NOS knockout mice was not
29      elevated (Fakhrzadeh et al., 2002).  Additionally, mice deficient in NOS or mice treated
30      with N -monomethyl-L-arginine, an inhibitor of total NOS, were protected from O3-induced
31      permeability, inflammation, and injury, suggesting a role of NO in the production of O3 effects

        August 2005                              5-30        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (Kleeberger et al., 2001a; Fakhrzadeh et al., 2002).  These results contrast with a study showing
 2      that O3 exposure (of 1 ppm for 8 h/night for 3 nights) produced greater injury, as determined by
 3      measurement of MTP-2, matrix metalloproteinases, total protein,  cell content and tyrosine
 4      nitration of whole lung protein, in iNOS knockout mice than in wild type mice (Kenyon et al.,
 5      2002). This group suggests that protein nitration is related to inflammation and is not dependent
 6      on iNOS-derived NO.  They point out the possible experimental differences, such as O3
 7      concentration, for inconsistency between their results and those of Kleeberger et al. (2001a).
 8           Rats pretreated with ebselen, a potent anti-inflammatory, immunomodulator, and
 9      NO/peroxynitrite scavenger, then exposed to 2 ppm O3 for 4 h had decreased numbers of
10      neutrophils, lowered albumin levels, and inhibited nitration of tyrosine residues in BALF 18 h
11      PE,  though macrophage iNOS expression was not changed (Ishii et al., 2000a). These results
12      suggest an iNOS-independent mechanism for O3-induced inflammation.  Jang et al. (2002)
13      showed dose-dependent increases in nitrate (indicative of in vivo NO generation) with O3
14      exposure (0.12, 0.5,  1, or 2 ppm for 3 h).  Functional studies of enhanced pause (Penh)
15      demonstrated increases with O3 exposure which were also dose-dependent.  Western blot
16      analysis  of lung tissue showed increases in NOS-1, but not in NOS -3 or iNOS isoforms. These
17      results suggest that in mice NOS-1 may induce airway responsiveness by a neutrophilic airway
18      inflammation. The literature regarding the effects of O3 exposure on NOS activity is complex
19      and  conflicting.  Similarly, the issue of protein nitration as it relates to cell injury due  to O3
20      exposure is somewhat controversial.
21
22      5.2.3.6  Summary and Conclusions - Inflammation and Permeability Changes
23           Figure 5-3 depicts many of the inflammatory and permeability changes that occur with O3
24      exposure. Additionally, the figure demonstrates links between inflammatory/permeability
25      responses and altered spirometric responses (discussed in Section 5.2.5), ciliary motility
26      (discussed in Section 5.2.2.1), airway hyperreactivity (discussed  in Section 5.2.5.3), and possible
27      thrombolytic effects (Section 5.3.3). Airway mucosa in the normal lung serves as an effective
28      barrier that controls bidirectional flow of fluids and cells between the air and blood
29      compartments. Ozone disrupts this function, resulting in a cascade of effects which includes an
30      increase in serum proteins, bioactive mediators,  and PMNs in the interstitium and air spaces of
31      the lung.  Damaged epithelial cells release cytokines, which function to recruit and activate AMs

        August 2005                              5-31        DRAFT-DO NOT QUOTE OR CITE

-------
 1      and PMNs. PMN recruitment into the lung is maximal at several hours PE. PMN recruitment is
 2      followed by blood monocytes which enter the lung and enlarge to become AMs.  The AMs
 3      persist for days to weeks, phagocytizing injured cells. Activated PMNs and AMs continue the
 4      cascade of effects by further releasing inflammatory mediators, which serve to amplify the initial
 5      effects of O3.  Generally, the initiation of inflammation is an important component of the defense
 6      process; however, its persistence and/or repeated occurrence can result in adverse health effects.
 7      Activation of this inflammatory cascade takes several hours.  Chemical mediators released early
 8      in the cascade contribute to effects on pulmonary function. Events later in the cascade, by
 9      which time O3-induced alterations pulmonary function have attenuated, are related to sustained
10      inflammation. Further, mechanistic separation of inflammation, permeability, and airway
11      hyperreactivity (AHR) is suggested by the temporal disparities of their increases.
12      The O3-induced disruption of the tight junctions between epithelial cells also increases
13      the permeability between the air and blood compartments.  This disruption, occurring with
14      exposures of 0.8 ppm for 2 h, is greater in the trachea and bronchoalveolar zone than in the
15      nose and allows entry of particles, including bioactive compounds, into the lung tissue.
16           For environmentally-relevant exposures to O3, concentration of exposure dominates the
17      response.  Studies evaluating C x  T relationships have not been published recently. Other
18      factors that have been studied for  potential impact on the effects of O3 include age, gender,
19      nutritional status, genetic variability, exercise and exposure to co-pollutants.  The effects of age
20      on lung inflammation are not well known.  After an acute exposure to 0.8 or  1 ppm, young mice,
21      rats, and rabbits had greater changes in prostaglandins in BALF, but there were no age-
22      dependent effects on BALF protein or cell number. Comparisons of male and female animals,
23      and vitamin C or ascorbate deficiency did not reveal significant differences in the effects of O3,
24      but exercise during exposure increased susceptibility.
25           Important  new work has revealed that susceptibility to O3 is, in part, genetically
26      determined. Mouse strains with differing sensitivities to O3 have identified genes on separate
27      loci controlling various aspects of inflammation, providing additional  evidence for the
28      mechanistic separation of responses to O3. The research is summarized in Figure 5-4.
29      Kleeberger's group has identified Inf-1, which modulates acute inflammatory responses; Inf-2
30      which modulates responses to subacute exposures; and TNF-a and TNF receptors, which are
31      involved in inflammatory responses. Other research groups have identified loci linked to other

        August 2005                               5-32        DRAFT-DO NOT QUOTE OR CITE

-------
Mouse Chromosomes

1 CZ

4CZ

7CZ

11 ( 	

15 CZ

17 r. 	

19 (__
rAOP2 ^_^__- , , u ,!«,,,,
oxidative stress /^ ^ 24 h inflammation
(Wattiezetal.,2003) / (Savovet al., 2004)
<==, ;;,,,;,,l N
•I «» ) ,— — ^-Toll-like receptor-4
__, 	 _ 	 	 	 	 	 — 	 — (Kleeberger et al.. 2000)
:_ )
_, 	 — 	 	 	 	 > 6h inflammation
6 h protein
: i ) (Savov etai., 2004)

jjg 	 }
^— __ _
> PMN influx
~\ (Kieebeiyei el al., 1997)
	 ~~ 	 — 	 ->• 24 h
• "} (Savov et ai,
V__

_: 2) (Kleeberger et al. , 1 997)
                   CC16 •
                   oxidant stress
                   (Mango et al., 1998)
                   epithelial permeability
                   (Broeckaert et al., 2003)
       Figure 5-4.  Mouse chromosomes on which genes or gene loci have been identified that
                    modulate responses to O3.
 1     endpoints.  This line of research provides a ground work for understanding the underlying
 2     mechanisms of O3-induced injury, and can shed light on genes responsible for human
 3     susceptibility to O3.
 4          Recent studies have placed a major focus on mediators released from inflammatory cells to
 5     understand the mechanisms of O3-induced inflammation and injury.  Cytokines and chemokines
 6     have been shown to be released as a result of stimulation or injury of macrophages, epithelial
 7     cells and PMNs.  Exposure of guinea pig AMs recovered in BALF and exposed in vitro to
 8     0.4 ppm O3 produced a significant increase in IL-6 and TNFa. An exposure of human AMs
 9     to an identical O3 concentration increased TNFa, IL-lp, IL-6 and IL-8.  The expression
10     of MIP-2 mRNA or BALF levels of MIP-2 increased in mice and rats exposed to O3
11     concentrations > 1 ppm. An increase after O3 exposure has also been reported for other cytokines
12     and inflammatory mediators,  including CINC and fibronectin. The CINC mRNA expression

       August 2005                               5-33        DRAFT-DO NOT QUOTE OR CITE

-------
 1      was associated with neutrophilia at 24 hrs PE. Ozone exposure of mice also caused an increase
 2      in IL-6, MlP-la and eotaxin in mice. Further understanding of the role of mediators has
 3      come from studies utilizing antibodies and inhibitors of known specificity.  In these studies
 4      treatment of rats with an anti IL-6 receptor antibody prior to a nighttime exposure to O3
 5      abolished O3-induced cellular adaptive response following a subsequent exposure.  Studies
 6      utilizing antibodies to selected pro- or anti-inflammatory cytokines suggest a role of TNFa,
 7      interleukin-10 (IL-10) and IL-lp in O3-induced changes in permeability, inflammation and
 8      cytokine release.
 9           Studies investigating the mechanisms of PMN recruitment in the lung have explored the
10      role of cell adhesion molecules that mediate PMN-endothelial cell interactions.  An increase in
11      tissue expression of ICAM-1 occurred in mice exposed to 0.8 ppm O3. A comparable pattern of
12      time-related changes in total protein, fibronectin and alkaline phosphatase activity in the BALF
13      was observed in rats exposed to 1 ppm O3. In monkeys, the O3-induced inflammation was
14      blocked by treatment with a monocolonal antibody to CD 18, suggesting dependence of PMN
15      recruitment on this adhesion molecule.  Together, these studies support the role  of extracellular
16      matrix protein and cell adhesion molecules in lung inflammation and injury.
17           Ozone exposure also affects macrophage functions, and consequently their role in lung
18      inflammation. Macrophages isolated from O3-exposed mice produced increased amounts of
19      NO, superoxide anion and PGE2, but production of these mediators by macrophages from
20      NOS knockout mice was not elevated. Additionally, mice deficient in NOS or mice treated
21      with NG-monomethyl-L-arginine, an inhibitor of total NOS, were protected from O3-induced
22      permeability, inflammation and injury.  These findings suggest a role  of NO in the production
23      of O3 effects.
24
25      5.2.4   Morphological Effects
26           Most mammalian species show generally similar morphological responses to <1 ppm O3
27      which differ only by region, cell type, exposure parameters, and length of time between exposure
28      and examination. Constant low exposures to O3 create an early bronchoalveolar exudation,
29      which declines with continued exposure and drops in the PE period. Epithelial hyperplasia also
30      starts early, increases in magnitude for several weeks, plateaus with continuing exposure, and
31      declines slowly during PE.  Interstitial fibrosis has a later onset,  continues to increase throughout

        August 2005                              5-34        DRAFT-DO NOT QUOTE OR CITE

-------
 1      the exposure, and can continue to increase after the exposure ends. Nonhuman primates respond
 2      more than rats at this concentration, due to differences in antioxidants, the CAR (predicted to
 3      receive the highest dose of O3), the presence of respiratory bronchioles, acinar volume, and
 4      differences in the nasal cavity's ability to "scrub" the O3.  Ciliated epithelial cells of the airway,
 5      Type 1 epithelial cells of the gas-exchange region, and ciliated cells in the nasal cavity are the
 6      cells most affected by O3. Ozone-damaged ciliated cells are replaced by nonciliated cells  (which
 7      are unable to provide clearance function) and Type 1  cells are replaced by Type 2 cells, which
 8      are thicker and produce more lipids.  Inflammation also occurs, especially in the CAR, wherein
 9      the tissue is thickened as collagen accumulates. At exposures of 0.25 ppm O3 (8 h/day, 18 mo)
10      in monkeys,  the distal airway is remodeled as bronchiolar epithelium replaces the cells present in
11      alveolar ducts.  In both rodents and monkeys, it appears that the natural seasonal patterns of O3
12      exposure alters morphology more than continuous exposures, thus long-term animal studies with
13      uninterrupted exposures may underestimate morphological effects.
14
15      5.2.4.1  Acute and Subchronic Exposure Effects
16           Morphological effects of key acute and subchronic exposure studies are summarized in
17      Table AX5-9. Harkema et al.  (1997a) reviewed toxicological studies of the nasal epithelial
18      response to short-term O3.  New  information regarding the effects of O3 in this region include
19      demonstrations that the topical anti-inflammatory corticosteriod fluticasone propionate prevents
20      inflammation and mucous cell metaplasia in rats after cumulative O3 exposure (0.5 ppm O3,
21      8 h/day, for 3 or 5 days) (Hotchkiss et al., 1998). Exposure to bacterial endotoxin, a common
22      ambient air toxicant, can potentiate mucous cell metaplasia in the nasal transitional epithelium of
23      rats caused by a previous 3 day 0.5 ppm O3 exposure  (Fanucchi et al., 1998). Male F344/N Hsd
24      rats were intranasally instilled with endotoxin after exposure to filtered air (FA) or 0.5 ppm O3,
25      (8 h/d for 3 d).  Mucous cell metaplasia was not found in the air/endotoxin group, but was found
26      in the O3/saline group and was most severe in the O3/endotoxin group. A similar synergistic
27      effect was demonstrated by Wagner et al. (2001a,b) with exposure of Fischer rats to O3 for 8 h
28      per day for 3 days and endotoxin. Ozone alone created epithelial lesions in the nasal transitional
29      epithelium, while endotoxin alone caused lesions in the epithelium of the nose and conducting
30      airways.  The enhanced O3-induced mucous cell metaplasia was related to neutrophilic
31      inflammation.

        August 2005                               5-35        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Pre-metaplastic responses, such as mucin mRNA upregulation, neutrophilic inflammation,
 2      and epithelial proliferation, were shown to be responsible for O3-induced mucous cell metaplasia
 3      in the transitional epithelium of rats (Cho et al., 1999a, 2000).  Male F344/N rats exposed to O3
 4      (0.5 ppm, 8 h/d for 1, 2, or 3 d) demonstrated a rapid increase in an airway-specific mucin gene
 5      mRNA after exposure to O3, both before and during the onset of mucous cell metaplasia.
 6      Neutrophilic inflammation coincided with epithelial DNA synthesis and upregulation of
 7      rMuc-5 AC, but was resolved before the development of epithelial metaplasia.  The mucous cell
 8      metaplasia was neutrophil-dependent, whereas O3-induced epithelial cell proliferation and mucin
 9      gene upregulation were neutrophil-independent.
10           Dormans et al.  (1999) compared the extent and time course of fibrotic changes in mice,
11      rats, and guinea pigs exposed to 0.2 and 0.4 ppm O3 for 3, 7, 28, and 56 days.  They found a
12      concentration-related centriacinar inflammation in all three species, with a maximum after
13      3 days of exposure and total recovery within 3 days after exposure. Repair of O3-induced
14      damage by removal of injured epithelial cells is enhanced by the influx of neutrophils (Hyde
15      et al., 1999; Veseley et al., 1999b; Miller et al., 2001; see Section  5.2.3). A study examining the
16      kinetics of early cellular responses to O3 utilized bromodeoxyuridine to label s-phase cells
17      (Hotchkiss et al., 1997). Labeling indices for rat nasal transitional epithelial cell DNA were
18      greatest 20 to 24 h after O3 (0.5 ppm for 8 h) exposure, and remained greater than control at
19      36hPE.
20           Very few published  studies have explicitly explored susceptibility factors such as species,
21      gender, age, antioxidant defense, acute  and chronic airway disease, and exercise.  Most typical
22      laboratory species studied have qualitatively similar effects associated with  O3 exposure.
23      Dormans et al. (1999) compared morphological, histological, and biochemical effects in the rat,
24      mouse, and guinea pig following O3 exposure and recovery in clean air. Wistar RIV:Tox male
25      rats, NIH male mice, and Hartley Crl:(HA)BR male guinea pigs were continuously exposed to
26      FA, 0.2, or 0.4 ppm for  3, 7, 28, and 56 days. Recovery from 28 days of exposure was studied at
27      intervals of 3, 7, and 28 days PE. The mouse was the most sensitive as shown by a concentration
28      and exposure-time dependent persistence of bronchiolar epithelial hypertrophy, elevated lung
29      enzymes, and slow recovery from exposure.  Exposure to the high dose for 56 d in both rats and
30      guinea pigs caused increased amounts of collagen in ductal septa and large lamellar bodies  in
        August 2005                               5-36        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Type II cells.  The inflammatory response was greater in the guinea pig. Overall, the authors
 2      rated mice as most susceptible, followed by guinea pigs and rats.
 3           Ferrets, monkeys, and rats were exposed to O3 (1.0 ppm, 8 h) to compare airway effects
 4      Sterner-Kock et al. (2000). The ferrets and monkeys had similar epithelial necrosis and
 5      inflammation that was more severe than that found in rats.  Because ferrets have a similar
 6      pulmonary structure as humans (e.g., well-developed respiratory bronchioles and submucosal
 7      glands), the authors concluded that the ferret would be a better model than rodents
 8      for O3-induced airway effects. Age susceptibility is dependent on the endpoint examined
 9      (see Chapter 4 for discussions of age-related differences in O3 dosimetry).  One study (Dormans
10      et al., 1996) demonstrated that O3-induced centriacinar lesions are larger in younger rats than in
11      older rats with exposures to 0.4 ppm for 1 to 7 days.
12           New studies have examined O3-induced  morphological effects in compromised laboratory
13      animals. Rats with endotoxin-induced rhinitis were more susceptible to mucous cell metaplasia
14      in the nasal transitional epithelium caused by a 3 day exposure to 0.5 ppm O3 (Cho et al.,1999b).
15      Wagner et al. (2002) reported a similar O3-induced enhancement of inflammatory and epithelial
16      responses associated with allergic rhinitis. Brown Norway rats were exposed to 0.5 ppm O3,
17      8 h/day for 1 day or 3 consecutive days and then immediately challenged intranasally with either
18      saline or ovalbumin (OVA).  Multiple exposures to O3 caused greater increases in
19      mucosubstances produced in the nose by allergen challenge.
20           Recent research has focused on the concept of O3 susceptible and nonsusceptible sites
21      within the respiratory tract, including in situ antioxidant status and metabolic activity. Plopper
22      et al. (1998) examined whether the variability of acute epithelial injury to short-term O3
23      exposure within the tracheobronchial tree is related to local tissue doses of O3 or to local
24      concentrations of reduced glutathione (GSH).  Adult male rhesus monkeys exposed to O3 (0.4 or
25      1.0 ppm for 2 h) demonstrated significant cellular injury at all sites, but the most damage, along
26      with increased inflammatory cells, occurred in the proximal respiratory bronchiole.
27      A significant reduction in GSH was found in the proximal bronchus at 0.4 ppm O3 and in the
28      respiratory bronchiole at 1.0 ppm O3. A significant decrease in the percent of macrophages,
29      along with significant increases in the percent of neutrophils and eosinophils, and a doubling of
30      total lavage protein, were found after exposure to 1.0 ppm O3 only. The authors concluded that
        August 2005                               5-37        DRAFT-DO NOT QUOTE OR CITE

-------
 1      the variability of local O3 dose in the respiratory tract was related to inhaled O3 concentration
 2      and was closely associated with local GSH depletion and with the degree of epithelial injury.
 3           Plopper and colleagues (e.g., Watt et al.,1998; Paige et al., 2000a) explored the site-
 4      specific relationship between epithelial effects of O3 exposure and the metabolism of
 5      bioactivated compounds within the respiratory tract of rats.  The distribution of CYP2E1-
 6      dependent activity, measured with a selective substrate (p-nitrocatechol), was found to be
 7      highest in the distal bronchioles and minor daughter airways, and lower in the lobar bronchi and
 8      major daughter airways. Short-term  O3 exposure (1 ppm for 8 h) increased CYP2E1  activity in
 9      the lobar bronchi/major daughter airways only; however, long-term O3 exposure (1 ppm for
10      90 days) decreased CYP2E1 activity in the major and minor airways, further complicating the
11      interpretation of O3 effects based on concentration and duration of exposure and recovery.  Rats
12      treated i.p. with 1-nitronaphthalene, a pulmonary toxicant requiring metabolic activation, and
13      exposed to 0.8  ppm O3, 8h/day for 90 days showed greater histopathologic and morphometric
14      effects in the CAR of the lung (Paige et al., 2000b). Despite reported tolerance to oxidant stress
15      after long-term O3 exposure, there was increased severity of ciliated cell toxicity.
16
17      5.2.4.2  Summary of Acute and Subchronic Morphological Effects
18           Short-term exposures to O3 cause similar alterations in lung structure in a variety of
19      laboratory animal species at concentrations of 0.15 ppm  in rats and lower concentrations in
20      primates. Cells in the CAR are the primary targets of O3, but ciliated epithelial cells in the nasal
21      cavity and airways and Type 1 epithelial cells in the gas  exchange region are also targets. New
22      work has shown that a topical anti-inflammatory corticosteroid can prevent these effects in nasal
23      epithelia, while exposure to bacterial endotoxin can potentiate the effects.  Ozone-induced
24      fibrotic changes in the CAR are maximal at 3 d of exposure and recover 3 d PE with exposures
25      of 0.2 ppm in rodents. New studies of susceptibility factors demonstrated that ferrets and
26      monkeys have  similar inflammatory and necrotic responses  to 1 ppm O3, which differs from
27      lesser injury seen in rats. Rats with induced allergic rhinitis are more susceptible to 0.5 ppm
28      than are controls. Important new work has demonstrated variability of local O3 dose and
29      subsequent injury in the RT due to depletion of GSH.  The proximal respiratory bronchiole
30      receives the most acute epithelial injury from exposures  < 1  ppm, while metabolic  effects were
31      greatest in the distal bronchioles and minor daughter airways.

        August 2005                               5-38        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.4.3  Subchronic and Chronic Exposure Effects
 2           Summaries of new studies of morphological effects of subchronic and chronic exposures
 3      are listed in Table AX5-10 in Annex AX5. In general, as the duration of exposure lengthens,
 4      there is not a concomitant linear increase in the intensity of effect of a given endpoint.  Rather,
 5      as exposure proceeds past 1 week to 1 year, Type 1 cell necrosis and inflammatory responses
 6      generally decrease to near control values, and hyperplastic and fibrotic changes remain elevated.
 7      After long-term exposure ended, some indicies of fibrosis persisted and in some cases became
 8      more severe during PE periods in clean air.
 9           Effects of O3 on the upper respiratory tract of F344 rats exposed to O3 (0.12, 0.5, or
10      1.0 ppm for 20 months) included marked mucous cell metaplasia in the rats exposed to 0.5 and
11      1.0 pm O3, but not at 0.12 ppm O3 (Harkema et al., 1997a) . In a follow-up study, hyperplasia
12      was found  in the nasal epithelium of rats exposed to 0.25 and 0.5 ppm, 8h/day, 7 days/week, for
13      13 weeks (Harkema et al., 1999). The mucous cell metaplasia, and associated intraepithelial
14      mucosubstances, induced by 0.5 ppm O3 persisted for 13 weeks after exposure.  An acute (8 h)
15      exposure to 0.5 ppm O3 13 weeks after the chronic exposure induced an additional increase of
16      mucosubstances in the nasal epithelium  of rats but not in rats chronically exposed to 0 or
17      0.25 ppm O3.  The persistent nature of the O3-induced mucous cell metaplasia in rats reported in
18      this study suggests that O3 exposure may have the potential to induce similar long-lasting
19      alterations  in the airways of humans.
20           No significant changes in nasal tissue were seen in rats continuously exposed for 49 days
21      to the ambient air of Mexico City, Mexico (Moss et al., 2001). A rat study using 6-month
22      exposures to ambient air of Sao Paulo, with a disparate pollutant composition than that of
23      Mexico City, demonstrated development of secretory hyperplasia in rats (Lemos et al., 1994).
24      However, without information on differences in ambient pollution composition in the two cities,
25      the studies cannot be compared.  Because of the persistent nature of these changes in the
26      controlled  studies with rats, and the fact that the upper airways of humans are probably more
27      sensitive, like the monkey, the authors suggested that long-term exposure to ambient levels of O3
28      could induce significant nasal epithelial lesions that may compromise the upper respiratory tract
29      defense mechanisms of exposed human  populations.
30           Rats  exposed to 0.5 ppm O3 for 1 month exhibited Bcl-2 in protein extracts of nasal
31      epithelium (Tesfaigzi et al., 1998).  Further, after 3 and 6 months of exposure, the number of

        August 2005                               5-39        DRAFT-DO NOT QUOTE OR CITE

-------
 1      metaplastic mucous cells in the transitional epithelium was indirectly related to the percentage of
 2      cells that were Bcl-2 positive .  Cells from rats exposed to FA did not express any Bcl-2. This
 3      study suggests that apoptosis regulators like Bcl-2 may play a role in the development and
 4      resolution of mucous cell metaplasia in the nasal airway.
 5           A spectrum of lesions was reported (Herbert et al., 1996) in the nasal cavity and
 6      centriacinar lung of male and female mice exposed to 0.5 or 1.0 ppm of O3 for 2 years, which
 7      persisted with continued exposure for 30 months. These lesions included bone loss in the
 8      maxilloturbinates, mucosal inflammation, mucous cell metaplasia in the nasal transitional
 9      epithelium and increased interstitial and epithelial thickening in the proximal alveolar region.
10      In the CAR, there were increased numbers of nonciliated cells.  However, changes in other
11      endpoints including lung function and lung biochemistry were not evident.  The investigators'
12      interpretation of the entire study is that rodents exposed to the two higher O3 concentrations had
13      some structural hallmarks of chronic airway disease in humans.
14           A chronic study using a simulated, seasonal O3-exposure pattern was reported by Plopper
15      and colleagues (Evans et al., 2003; Schelegle et al., 2003a; Chen et al., 2003; Plopper and
16      Fanucchi, 2000).  Infant rhesus monkeys (30 days old) were exposed to FA, house dust mite
17      allergen aerosol (HDMA), or O3 + HDMA. The 0.5 ppm O3 exposures were 8 h/day for 5 days,
18      every 14 days for a total of 11 O3 episodes. Half of the monkeys were sensitized to house dust
19      mite allergen (Dermatophagoides farinae) at  14 and 28 days of age. The sensitized monkeys
20      were exposed to HDMA for  2 h/day on Days  3-5 of the FA or O3 exposures. The lungs were
21      removed during the last FA exposure and the  right and left cranial and right middle lobes were
22      separately inflation fixed. Microdisection and morphometric analyses were performed on the
23      conducting airways to the level of the most proximal respiratory bronchiole. Repeated
24      exposures to O3 or O3 + HDMA over a 6-month period resulted in an atypical development of
25      the basement membrane zone of airways in nonsensitized developing monkeys.  Remodeling in
26      the distal conducting airways was found in the sensitized monkeys as a result of the damage and
27      repair processes occurring with repeated exposure (Evans et al., 2003; Schelegle  et al., 2003a).
28      Lung function changes in these monkeys (Schelegle et al., 2003b), and associated adaptation of
29      the respiratory motor responses (Chen et al., 2003), are described in Section 5.2.5.2.
30      Collectively, these findings provide a pathophysiologic basis for changes in airway function
        August 2005                              5-40        DRAFT-DO NOT QUOTE OR CITE

-------
 1      described in children growing up in polluted metropolitan areas (e.g., Tager, 1999)
 2      (see Chapter 7).
 3           Necropsy of the left caudal lobe of these infant monkeys showed accumulation of
 4      eosinophils and mucous cells within the combined epithelial and interstitial compartments in the
 5      conducting airways and in the terminal/respiratory bronchioles (Schelegle et al., 2003a) . House
 6      dust mite sensitization and HDMA challenge alone, or combined with O3 exposure, resulted in
 7      significantly greater eosinophil accumulation in the conducting airways when compared to FA
 8      and O3 only exposures.  A significant accumulation of eosinophils was found in the
 9      terminal/respiratory bronchioles of the sensitized monkeys challenged with HDMA when
10      compared to monkeys exposed to FA, O3, and HDMA + O3. The mean mass of mucous cells
11      increased in the fifth generation conducting airways of sensitized animals challenged with
12      HDMA alone and when combined with O3 exposure, and in the terminal bronchioles of
13      sensitized animals exposed to HDMA + O3. The tracheal basement membrane of HDMA-
14      sensitized monkeys exposed to HDMA or to HDMA + O3 was significantly increased over
15      controls; however, there were no significant changes in the airway diameter of proximal and
16      mid-level airways. Exposures of sensitized young monkeys to HDMA alone, or to O3 alone,
17      resulted in eosinophilia of the mid-level conducting airways and the terminal/respiratory
18      bronchioles, but without alterations in airway structure or function.  The authors interpreted
19      these findings to indicate that the combination of cyclic O3 exposure and HDMA challenge in
20      HDMA-sensitized infant monkeys act synergistically to produce an allergic-reactive airway
21      phenotype characterized by significant eosinophilia of midlevel conducting airways,
22      transmigration of eosinophils into the lumen, and an altered structural development of
23      conducting airways that is associated with increased airway resistance and nonspecific airway
24      reactivity (see Section 5.2.5).
25           Examination of development of the tracheal basement membrane zone (BMZ) in these
26      monkeys (Evans et al., 2003) showed that with exposures to either O3 or HDMA + O3 BMZ
27      development was affected.  Abnormalities in the BMZ included: (1) irregular and thin collagen
28      throughout the BMZ; (2) perlecan depeleted or severely reduced; (3) FGFR-1 immunoreactivity
29      reduced; (4) FGF-2 immunoreactivity absent in perlecan-deficient BMZ, but present in the
30      lateral intercellular space (LIS), in basal cells, and in attenuated fibroblasts; (5) syndecan-4
31      immunoreactivity increased in basal cells.  The authors interpret these data as suggesting that O3

        August 2005                              5-41        DRAFT-DO NOT QUOTE OR CITE

-------
 1      targets cells are associated with synthesis of epithelial BMZ perlecan. The absence of FGF-2,
 2      normally stored in the BMZ, could affect downstream signaling in airway epithelium and could
 3      be responsible for the abnormal development of the airway seen in this study, and thus be an
 4      important mechanism modulating O3-induced injury. Midlevel bronchi and bronchioles from
 5      these monkeys (Larson et al., 2004) demonstrated decrements in the density of epithelial nerves
 6      in the axial path between the sixth and seventh airway generations in exposures to O3.
 7      Combined O3 + HDMA exposures exacerbated this reduction. They attribute this loss of nerve
 8      plexuses to neural regression or stunted nerve development, the latter corroborated by the Evans
 9      et al. (2003) finding of decreased growth factors following O3 exposure.  Additionally, they
10      found streaks or clusters of cells immunoreactive for protein gene product 9.5 (PGP 9.5,  a pan-
11      neuronal marker) and negative for calcitonin gene-related peptide. The functional significance
12      of this is unknown but suggests to the authors a possible injury-repair process induced by O3.
13           Remodeling of the distal  airways and CAR is one of the most disturbing aspects of the
14      morphological changes occurring after subchronic and chronic exposure to O3.  Recently,
15      bronchiolization was reported in rats exposed to 0.4 ppm O3 for only 56 days (van Bree et al.,
16      2001). They also found collagen formation progressively increased with increasing O3 exposure
17      and persisted into PE recovery. In addition to centriacinar remodeling, Pinkerton et al. (1998)
18      reported thickening of tracheal, bronchial, and bronchiolar epithelium after 3 or 20 months
19      exposure to 1 ppm O3, but not to 0.12 ppm.  Although some older literature had reported that
20      chronic exposures to < 1.0 ppm O3 cause emphysema, none of the more recent literature supports
21      this hypothesis.
22
23      5.2.4.4   Summary and Conclusions - Subchronic and Chronic Morphological Effects
24           The progression of effects during and after a chronic exposure at a range of 0.5 to 1.0 ppm
25      is complex, with inflammation peaking over the first few days of exposure, then dropping, then
26      plateauing, and finally, largely disappearing. Epithelial hyperplasia follows a somewhat similar
27      pattern.  Effects of 0.5 ppm O3 for 20 months on the nasal mucosa include atrophy of nasal
28      turbinates and mucous cell metaplasia, which persisted long after the exposure ceased. Fibrotic
29      changes  in the tissue increase very slowly over months of exposure, and,  after exposure ceases,
30      the changes sometimes persist  or increase.  The pattern of exposure in this same concentration
31      range determines effects, with  18 mo of daily exposure causing less morphologic damage than

        August 2005                              5-42       DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposures on alternating months. This is important, given that environmental O3 exposure is
 2      typically seasonal. Plopper and colleagues' long term study of infant rhesus monkeys exposed to
 3      simulated, seasonal O3 (0.5 ppm 8 h/day for 5 days, every 14 days for 11 episodes)
 4      demonstrated: (1) remodeling in the distal airways; (2) abnormalities in tracheal basement
 5      membrane; (3) eosinophil accumulation in conducting airways; and (4) decrements in airway
 6      innervation. These findings advance earlier information regarding possible injury-repair
 7      processes occurring with seasonal O3 exposures.
 8
 9      5.2.5  Effects on Pulmonary Function
10      5.2.5.1  Acute and Subchronic Exposure Effects on Pulmonary Function
11           Numerous pulmonary function studies of the effects of acute O3 exposure (defined here
12      as < 1 week of exposure) in several animal species have been conducted and generally show
13      responses similar to those of humans (e.g., increased breathing frequency, decreased tidal
14      volume, increased resistance, decreased forced vital capacity (FVC) and changes in the
15      expiratory flow-volume curve).  These effects are seen at 0.25 to 0.4 ppm O3 for several h in a
16      number of species. At concentrations of > 1 ppm, breathing mechanics (compliance and
17      resistance) are affected.  The breathing pattern returns to normal after O3 exposure. In rats
18      exposed to 0.35 to 1 ppm O3 for 2 h/day for 5 days, there was a pattern of attenuation of
19      pulmonary function responses similar to that observed in humans.  Concurrently, there was no
20      attenuation of biochemical indicators of lung injury or of morphological changes.
21           Work demonstrating attenuation of pulmonary functions (see Table AX5-11) was
22      completed by Wiester et al. (1996) who exposed male Fischer 344 rats to 0.5 ppm O3 for either
23      6 or 23 h/day over 5 days.  Ozone-induced changes in lung volume were attenuated during the
24      5 exposure days and returned to control levels after 7 days recovery. The responses to repeated
25      O3 exposure in rats were exacerbated by reduced ambient temperature, presumably as a result of
26      increased metabolic activity.
27           Researchers have utilized inbred mouse strains with varying ventilatory responses to O3 to
28      attempt to model susceptible populations.  As differences were seen in inflammatory responses
29      to acute O3 exposures in C57BL/6J and C3H/HeJ mice, comparisons were made of their
30      ventilatory responses also (Tankersley et al., 1993). Following an exposure of 2 ppm O3 for 3 h,
        August 2005                              5-43        DRAFT-DO NOT QUOTE OR CITE

-------
 1      breathing frequency (f), tidal volume (VT), and minute ventilation were measured 1 and 24 h in
 2      both normocapnia (or air at -0% CO2) and hypercapnia (5 or 8% CO2). They demonstrated that
 3      acute O3 exposures caused altered hypercapnic ventilatory control, which varied between strains.
 4      This suggested to the authors that O3-induced alterations in ventilation are determined, at least in
 5      part, by genetic factors. A caveat regarding studies such as this using high exposure
 6      concentrations is that events observed at high concentrations may differ from those observed at
 7      near-ambient O3 levels.
 8           Paquette et al. (1994) measured ventilatory responses in C57BL/6J and C3H/HeJ mice
 9      given repeated acute exposures of 0.3  ppm for 48 and 72 h.  The two strains had differing
10      responses to both normocapnia and hypercapnia. Normocapnic VE was greater following
11      subacute O3 exposure in C57BL/6J mice than in C3H/HeJ mice, due to increased/and
12      reduced VT, respectively. This suggests that the increased VT in C57BL/6J mice may contribute
13      to the increased susceptibility to lung injury due to a greater dose of O3 reaching the lower lung.
14      Hypercapnic ventilatory responses following subacute O3 exposures demonstrated reduced VE
15      (due to decreased VT) in C57BL/6J only.  Evaluations of O3 dosimetry were performed in these
16      two strains using 18O3-labeled ozone (2 ppm for 2-3 h) (Slade et al., 1997).  Immediately after
17      exposures of 2 ppm 18O3 for 2-3 h, C3H/HeJ mice had 46% less 18O in lungs and 61% less in
18      trachea, than C57BL/6J. Additionally, C3H/HeJ mice had a greater body temperature decrease
19      following O3 exposure  than C57BL/6J mice, suggesting that the differences in susceptibility
20      to O3 are due to differences the ability to decrease body temperature and, consequently decrease
21      the dose of O3 to the lung.
22           Tracheal transepithelial potential has also been shown to differ in eight mouse strains 6 h
23      after exposure to 2 ppm O3 for 3 h (Takahashi et  al., 1995b). AKR/J, C3H/HeJ, and CBA/J were
24      identified as resistant strains and 129/J, A/J, C57BL/6J, C3HeB/FeJ and SJL/J were identified as
25      susceptible strains.  The authors noted that strains' responses to this parameter did not show
26      concordance with inflammatory responses, suggesting to the authors that the two phenotypes are
27      not controlled by the same genetic factors.
28           Savov et al. (2004) characterized ventilatory responses in nine mouse strains exposed to O3
29      (2.0 ppm O3 for 3 h). C57BL/6J was hyporeactive to MCh prior to O3, but was very responsive
30      to MCh following O3.  Conversely, C3H/HeJ had an intermediate baseline Penh and a  small
        August 2005                              5-44        DRAFT-DO NOT QUOTE OR CITE

-------
 1      response to MCh following O3 exposure.  This study corroborates the evidence of no consistent
 2      relationship between respiratory Penh and inflammation.
 3
 4      5.2.5.2  Summary and Conclusions - Acute and Subchronic Effects on
 5              Pulmonary Function
 6           Early work has demonstrated that during acute exposure of -0.2 ppm O3 in rats, the most
 7      commonly observed alterations are increased frequency of breathing and decreased tidal volume
 8      (i.e., rapid, shallow breathing).  Exposures of-1.0 ppm O3 affect breathing mechanics
 9      (compliance and resistance). Additionally, decreased lung volumes are observed in rats with
10      acute exposures at levels of 0.5 ppm.  New work utilizing inbred mouse strains with varying
11      ventilatory responses to O3 has suggested that: (1) control of the ventilatory response is
12      determined, at least in part, by genetic factors; (2) increased VT in some strains may contribute
13      to lung injury due to a greater dose of O3 reaching the lower lung; (3) some strains' ability to
14      reduce body temperature may account for their decreased O3-induce lung injury; and (4) tracheal
15      transepithelial potential is determined, in part, by genetic factors. Importantly, the genetic loci
16      that appear to be modulating various aspects of pulmonary responses to O3 differ from each
17      other and from loci controlling inflammatory responses.
18           Exposures of 2 h/day for 5 days create a pattern of attenuation of pulmonary function in
19      both rats and humans without concurrent attenuation of lung injury and morphological changes,
20      indicating that the attenuation did not result in protection against all the effects of O3.
21      Chronic O3 exposure studies evaluating pulmonary function are not available.  Earlier work
22      has demonstrated that repeated daily exposure of rats to an episodic profile of O3 caused small,
23      but significant decrements in lung function that were consistent with early indicators of focal
24      fibrogenesis in the proximal alveolar region, without overt fibrosis.
25
26      5.2.5.3  Ozone Effects on Airway Responsiveness
27           Effects of O3 on airway reactivity  have been observed in a variety of species at an exposure
28      range of 0.5 to  1 ppm.  Many of the new studies on pulmonary function in laboratory animals
29      allow a better prediction of the effects of O3 exposure on the exacerbation of asthma symptoms
30      and the risk of developing asthma in humans. However, it is necessary to understand the factors
31      that determine airway responsiveness across different mammalian species as discussed  in
32      Chapter 4.

        August 2005                              5-45         DRAFT-DO NOT QUOTE OR CITE

-------
 1           Traditional studies of airway responsiveness require sedation in both infants and laboratory
 2      animals. Laboratory animal studies employ intravenous agonist challenges as well as inhalation
 3      challenges, though inhaled agonist challenges are preferred in humans. Exercise testing is
 4      not possible with sedation unless exercise is "simulated" by increasing ventilation using
 5      elevated F;CO2; and the need for artificial ventilation in laboratory animal studies may cause
 6      breathing patterns that affect O3 deposition. load et al. (2000) reported that when 1 ppm O3 for
 7      90 min is administered to isolated rat lung at either 2.4 mL/40 bpm or 1.2 mL/80 bpm, the more
 8      rapid breathing pattern elicits less epithelial cell injury than the slower breathing pattern.
 9      Though this study design does not really model rapid shallow breathing elicited in the intact
10      animal, it shows greater reduction in injury in the proximal axial airway compared to its adjacent
11      airway branch and terminal bronchiole. The rapid, shallow breathing pattern protects the large
12      conducting airways of rats, but causes a more even distribution of epithelial cell injury to the
13      terminal bronchioles (Schelegle et al., 2001).  Postlethwait et al. (2000) demonstrated that the
14      conducting airways are the primary site of acute cytotoxicity  from O3 exposure. Three-
15      dimensional mapping of the airway tree in SD rat isolated lung exposed to 0, 0.25, 0.5, or
16      1.0 ppm O3 showed a concentration-dependent increase in injured cells.  Injury was evident in
17      proximal and  distal conduction airways, lowest in terminal bronchioles, and highest in the small
18      side branches downstream of bifurcations. These exposure levels did not concurrently elicit
19      changes in LDH activity or total protein in BALF, suggesting that the mapping technique is a
20      more sensitive measure of injury and is useful in dosimetry studies.
21           Whole-body plethysmography of unanesthetized, unrestrained rodents has been used to
22      indirectly measure pulmonary resistance (Shore et al., 2002; Goldsmith et al., 2002; Jang et al.,
23      2002).  However, these indices of inspiratory/expiratory pressure  differences, including
24      enhanced pause (Penh) may be less sensitive than direct measurements of lung airflow resistance
25      (Murphy, 2002). Changes in airway structure caused by viral infections also must be considered
26      when evaluating laboratory animal studies.  Animals with acute viral illness have morphological
27      evidence of inflammatory cell infiltration, bronchiolar wall edema, epithelial hyperplasia, and
28      increased airway mucous plugs that can cause airway narrowing, air trapping, and serious
29      functional changes in the lung (Folkerts et al., 1998).
30           Exercise-induced bronchoconstriction in humans appears to be mediated by changes in the
31      tonicity of the airway lining fluid (Anderson and Daviskas, 2000). Brannan et al. (1998) suggest

        August 2005                               5-46        DRAFT-DO NOT QUOTE  OR  CITE

-------
 1      that a test in laboratory animals based on the inhalation of mannitol aerosol (hyperosmolar)
 2      might be feasible and provide information similar to that from exercise challenges in cooperative
 3      children and adults. Unfortunately, there have been few reports of mannitol or adenosine
 4      monophosphate challenges in laboratory animals; most studies have utilized histamine,
 5      methacholine, acetylcholine, or carbachol to determine outcome. In active humans with asthma,
 6      adenosine monophosphate challenges appear to better reflect ongoing airway inflammation than
 7      histamine or methacholine challenges (Polosa and Holgate,  1997; Avital et al, 1995a,b), and
 8      might be useful in identifying mechanisms of asthma in laboratory animals and their
 9      responsiveness to environmental pollutants.
10           The increased responsiveness to bronchoconstrictor challenge in asthma is thought to result
11      from a combination of structural  and physiological factors that include increased inner-wall
12      thickness, increased smooth-muscle responsiveness, and mucus  secretion. These factors also
13      are likely to determine a level of innate airway responsiveness that is genetically influenced.
14      Chapter 6 (Section 6.8) discusses cellular and biochemical changes that have been identified
15      in human asthmatics.  These studies suggest that the mechanisms involved in AHR are
16      multifactoral, with general agreement that there is an inconsistent relationship between AHR
17      and markers of inflammation.
18           A large data base of laboratory animal research has been collected on the role of O3 in
19      producing an increase in AHR (see Table AX5-12). Exposure levels (> 1 ppm for >30 min)
20      in many of these studies are not environmentally relevant, but information may be obtained
21      regarding the mechanisms of action of O3 concerning: O3 concentration  and peak response time,
22      inhaled versus intravenous challenge with nonspecific bronchoconstrictors, neurogenic
23      mediation, neutrophilic inflammation, and  interactions with specific biological agents (e.g.,
24      antigens and viruses). However, as with other toxicants, high-dose and  low-dose mechanisms
25      may differ, so interpretation of results must take this into consideration.
26           Many species of laboratory animals have been used to study the effects of O3  on airway
27      bronchoconstriction. Ozone-induced AHR in guinea pigs has been used to model human
28      bronchospasm  (van Hoof et al., 1996; 1997a,b; Matsubara et al., 1997a,b;  Sun and Chung,  1997;
29      Aizawa et al., 1999a,b; Tsai et al., 1998; Nakano et al., 2000). Because these studies were done
30      at 2 to 3 ppm O3, these results are not directly relevant for extrapolation to potential airway
31      responses in humans exposed to ambient levels of O3. Humans with reactive airway disease

        August 2005                              5-47        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (e.g., asthma) appear to be sensitive to ambient levels of O3 (see Chapters 6 and 7) and the
 2      current understanding is that O3 exacerbates airway responsiveness to specific allergens,
 3      presumably by nonspecifically increasing AHR.
 4           Shore et al. (2000, 2002) have shown that O3-induced AHR is reduced in immature rats and
 5      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
 6      3 ppm for 3 h at age 2, 4, 8, or 12 weeks had similar concentration-related decreases in VE except
 7      at the youngest ages. This smaller decrement in VE suggested a delivered dose that was much
 8      greater in the younger animals. This group (Shore et al., 2003) has also recently shown that
 9      obese mice have greater ventilatory responses to O3. Exposures of 2.0 ppm O3 for 3 h to lean,
10      WT C57BL/6J and ob/ob mice (mice with a genetic defect in the coding for leptin, the satiety
11      hormone) showed that the ob/ob mice had enhanced AHR and inflammation compared to  the
12      WT mice.  These data correlate with epidemiological data showing increased incidence of
13      asthma in overweight children.
14           Increased AHR to various nonspecific bronchoconstrictive agents  (e.g., ACh,
15      methacholine, histamine, carbachol) given by inhalation or intravenous routes has been
16      previously shown in laboratory animals exposed to O3 concentrations < 1.0 ppm. Dye et al.
17      (1999) showed hyperresponsiveness to methacholine in rats 2 h after exposure to 2 ppm O3 for
18      2 h. AHR can be induced by specific antigens as well as O3. The  most commonly used
19      laboratory  animal model is the OVA sensitized guinea pig. Animals sensitized with OVA have
20      been shown to have similar responses to nonspecific bronchoconstrictors as control animals.
21           OVA-sensitized guinea pigs (Sun et al., 1997) and mice (Yamauchi et al., 2002) were used
22      to determine the enhancement of antigen-induced bronchoconstriction by acute, high-level O3
23      (1.0 ppm O3 for 1 h). Male Dunkin-Hartley guinea pigs were sensitized by i.p. injection of OVA
24      and exposed to O3 alone, OVA aerosol, or O3 + OVA.  Ozone exposure  alone increased
25      bronchial responsiveness to ACh at 3 h, but not 24 h, while OVA alone had no effect. Combined
26      exposure to O3 and OVA (1 ppm for 1 h, then 3 min OVA) increased bronchial responsiveness to
27      ACh 3 h after O3 exposure. At 24 h following O3 exposure, AHR  increased when OVA
28      challenge was performed at 21 h, suggesting that O3 pre-exposure can potentiate OVA-induced
29      AHR. Neutrophil counts in the BALF increased at 3 and 24 h after O3 exposure alone but were
30      not further increased when O3 exposure was combined with OVA  airway challenge; however
        August 2005                              5-48        DRAFT-DO NOT QUOTE OR CITE

-------
 1      protein content of the BALF did increase at 3 and 24 h in the O3 and OVA groups.  Thus, this
 2      study also indicates that high-ambient O3 exposure can augment antigen (OVA)-induced AHR in
 3      guinea pigs.
 4           Yamauchi et al. (2002) sensitized male C57BL/6 mice by i.p. injection of OVA and then
 5      exposed them to O3. The sensitized mice had AHR to methacholine. Ozone exposure caused
 6      significant decreases in dynamic lung compliance, minute ventilation, and PaO2 in OVA-
 7      sensitized mice, but not in controls. A marker of inflammation (soluble intercellular adhesion
 8      molecule-1 [sICAM-1]) was elevated in the BAL fluid of OVA-sensitized mice, but sICAM-1
 9      levels were not significantly changed by O3 exposure, indicating that the O3-induced AHR to
10      methacholine was not caused by O3-induced inflammation.
11           Ozone-induced AHR may be temporally associated with inflammatory cells stimulated by
12      cytokines (Koto et al., 1997), mast cells (Igarashi et al., 1998; Noviski et al., 1999), or by oxygen
13      radicals (Takahashi et al.,  1993).  One study, however, has shown that inflammation is not a
14      prerequisite of AHR (Koto et al., 1997), and it has been suggested that O3-induced AHR may be
15      epithelium dependent (McGraw et al., 2000). For example, neonatal rats pretreated with
16      capsaicin, which will permanently destroy C-fibers and prevent O3-induced (1 ppm, 8 h) release
17      of neuropeptides (Vesely et al., 1999a), and then exposed to O3 when adults, showed a marked
18      increase in airway responsiveness to inhaled aerosolized methacholine (Jimba et al., 1995).
19      Takebayashi et al. (1998) has shown that depletion of tachykinins by capsaicin treatment, or by a
20      specific tachykinin receptor antagonist, can block the induction of AHR by O3. The seemingly
21      disparate  responses in laboratory animals may be due to species- or strain-specific differences in
22      inherent reactivity to bronchoconstrictors, or to inherent differences in susceptibility to O3-
23      induced inflammation (Zhang et al., 1995; Depuydt et al., 1999; Dye et al., 1999).
24           Studies that may be potentially relevant to ambient levels of O3 were conducted in vivo, in
25      an isolated perfused lung model, and in ex vivo lung segments using multihour and repeated
26      multihour exposures with ambient levels of O3. A study on the relationship between O3-induced
27      AHR and tracheal epithelial function was conducted in New Zealand white rabbits by Freed
28      et al. (1996). Rabbits exposed to O3 (0.2 ppm for 7 h) demonstrated significantly decreased
29      tracheal transepithelial potential difference but no changes in lung resistance. Changes in the
30      compartmentalized lung resistance, measured in response to ACh challenge before and after
31      bilateral vagotomy, were not significantly different in air-exposed rabbits; however, bilateral

        August 2005                              5-49        DRAFT-DO NOT QUOTE OR CITE

-------
 1      vagotomy did enhance peripheral lung reactivity in O3-exposed rabbits. The ACh-induced
 2      a 140% increase in lung resistance with O3 exposure was two times higher than with air
 3      exposure, indicating that ambient-level O3 exposure affects tracheal epithelial function in rabbits
 4      and increases central airway reactivity, possibly through vagally-mediated mechanisms.
 5           Pulmonary mechanics and hemodynamics were studied in the New Zealand white rabbit
 6      isolated perfused lung model that allowed partitioning of the total pressure gradient into arterial,
 7      pre- and post-capillary, and venous components (Delaunois et al., 1998). Exposures to O3
 8      (0.4 ppm for 4 h) were followed by evaluation of airway responsiveness to ACh, substance P
 9      (SP), or histamine immediately or 48 h later.  Ozone inhibited pulmonary mechanical reactivity
10      to all three bronchoconstrictors that persisted for 48 h and modified vasoreactivity of the
11      vascular bed, but only at 48 h PE.  Arterial segmental pressure, normally insensitive to ACh and
12      SP, was significantly elevated by O3; precapillary segmental pressure decreased in response to
13      Ach, suggesting that O3 can induce direct vascular constriction, but the vascular responses are
14      variable and depend on the agonist used and on the species studied.
15           Airway responsiveness to the same three compounds was evaluated by Segura et al. (1997)
16      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
17      ACh or histamine, except at the highest concentration (1.2 ppm O3) for histamine. However, O3
18      did cause AHR to  SP at >0.3 ppm, suggesting that O3 destroys neutral endopeptidases
19      (responsible for SP inactivation) in airway epithelial cells.  Vargas et al. (1998), in a follow-up
20      study, demonstrated that guinea pigs chronically exposed to 0.3 ppm O3 for 4 h/day became
21      adapted to SP-induced AHR. Ozone caused increased sensitivity to SP after 1,3,6, 12, and
22      24 days of exposure that was associated with airway inflammation; however, after 48 days of
23      exposure, the increased sensitivity to SP was lost.
24           This  study is in accordance with  Szarek et al. (1995)  who demonstrated that AHR
25      associated with acute O3 exposures does not persist during long-term exposure to near-ambient-
26      levels of O3 (< 1 ppm).  Fischer 344 rats, exposed to 0.0, 0.12, 0.5, or 1.0 ppm O3, 6 h/day,
27      5 days/week for 20 months, demonstrated significantly reduced responses to bethanechol, ACh,
28      and electrical field stimulation in eighth generation airway segments. This suggests that some
29      adaptation had taken place during long-term exposure, possibly increased inner wall thickness.
30           It is well known that the changes in breathing pattern and lung function caused by O3 are
31      attenuated with repeated daily exposures for at least 3 to 5  days. But guinea pigs exposed to

        August 2005                              5-50        DRAFT-DO NOT QUOTE OR CITE

-------
 1      0.5 ppm O3, 8 h/day for 7 days showed enhancement of responsiveness of rapidly adapting
 2      airway receptors (load et al., 1998).  Repeated exposure increased receptor activity to SP,
 3      methacholine, and hyperinflation; there were no significant effects on baseline or SP- and
 4      methacholine-induced changes in lung compliance and resistance, suggesting that the
 5      responsiveness  of rapidly adapting receptors was enhanced.
 6           Male and female Hartley guinea pigs exposed to O3 (0.1 and 0.3 ppm, 4 h/day, 4 days/week
 7      for 24 weeks) were evaluated for airway responsiveness following ACh or OVA inhalation
 8      challenges (Schlesinger et al., 2002a,b). Ozone exposure did not cause AHR in nonsensitized
 9      animals but did exacerbate AHR to both ACh and OVA in sensitized animals that persisted for
10      4 weeks after exposure. The effects of O3 on airway responsiveness were gender independent
11      and were concentration-related for the ACh challenges.
12           Schelegle et al. (2003a) evaluated airway responsiveness in infant rhesus monkeys exposed
13      to a 5 day O3 episode repeated every 14 days over a 6-month period.  Half of the monkeys were
14      sensitized to house dust mite allergen (HDMA; Dermatophagoides farinae) at 14 and 28 days of
15      age before exposure to a total of 11 episodes of O3 (0.5 ppm, 8 h/day for 5 days followed
16      by 9 days of FA), HDMA, or O3 + HDMA. Baseline Raw was significantly elevated after
17      10 exposure episodes in the HDMA + O3 group compared to the FA, HDMA, and O3 exposure
18      groups. Aerosol challenge with HDMA at the end of the 10th episode did not significantly
19      affect Raw, VT, fB, or SaO2. Aerosol challenge with histamine was not significantly different after
20      6 episodes; however, the EC 150 Raw for the HDMA + O3 group was significantly reduced after
21      10 episodes when compared to the FA, HDMA, and  O3 exposure groups, indicating the
22      development of AHR in this group sometime between episodes 6 and 10. The results are
23      consistent with altered structural development of the conducting airways.
24           During repeated episodic exposures to O3, respiratory responses are first altered to a rapid,
25      shallow breathing pattern, which has long been considered protective, especially to the deep
26      lung. This dogma has been discounted recently as discussed above (Schelegle et al.,  2001).
27      Alfaro et al. (2004) examined the site-specific  deposition of 18O (1 ppm 2 h) at breathing
28      frequencies of 80, 120, 160, or 200 breaths/minute (bpm). At all frequencies, parenchymal areas
29      had a lower content of 18O than trachea and bronchi.  As breathing frequency increased from 80
30      to 160 bpm, the deposition showed a reduction in midlevel trachea and  an increase in both
31      mainstream bronchi.  At this frequency there was also an increase in deposition in parenchyma

        August 2005                              5-51        DRAFT-DO NOT QUOTE OR CITE

-------
 1      supplied by short (cranial) airway paths, consistent with results seen by Schelegle et al., (2001).
 2      At 200 bpm 18O deposition in trachea increased, concurrent with increases in right cranial and
 3      caudal bronchi regions. Right cranial parenchymal content decreased at 200 bpm, whereas right
 4      caudal parenchymal levels did not change at any breathing frequency.  The authors list some
 5      limitations of this study, such as the possible effect on regional distribution of ventilation by use
 6      of the negative-pressure ventilator, the effect of paralysis on airway geometry, and possible
 7      translocation of 18O during the 2 h exposure period.  These two studies provide evidence
 8      that O3-induced rapid, shallow breathing creates a more evenly distributed injury pattern,
 9      with possibly greater protection from focal injury to the large conducting airways including the
10      trachea and the left mainstem bronchus.
11           Another study of the adaptive phenomena in SD rats used an exposure paradigm consisting
12      of 5 days of daily 8 h 1 ppm O3 exposures followed by 9 days of recovery in FA (Schelegle
13      et al.,  2003b).  This O3/FA pattern was repeated for 4 cycles and demonstrated that the O3-
14      induced rapid shallow breathing pattern was followed by adaptation that occurred with each
15      cycle. However, the release of SP from the trachea , the neutrophil content, and cell
16      proliferation became attenuated after the first cycle, suggesting a disconnect from the rapid
17      shallow breathing response. Hypercellularity of the CAR epithelium and thickening of the CAR
18      interstitium, not linked to changes in cell proliferation, were also found. The authors suggest
19      mechanism(s) of injury from repeated O3 exposures consisting of diminished neutrophilic
20      inflammation/and or release of mitogenic neuropeptides, depressed cell proliferative response,
21      and cumulative distal airway lesion.
22           Following the initial response of a rapid, shallow breathing pattern, animals eventually
23      adapt  with continued episodic exposure despite the continued presence of epithelial damage,
24      altered structural development, and inflammation of the airways.  Chen et al. (2003) used a
25      subset of the monkeys from the Schlegele et al. (2003a) study to demonstrate that attenuation
26      of O3-induced rapid shallow breathing and lung function changes typically  seen with repeated O3
27      exposure may be caused by the adaptation of the respiratory motor responses.  This episodic O3
28      exposure appeared to create neuroplasticity of the nucleus tractus solitarius (NTS; a region of the
29      brainstem which controls  respiration), including increased nonspecific excitability of the NTS
30      neurons, an increased input resistance,  and an increased spiking response to intracellular
31      inj ections of depolarizing current.

        August 2005                               5-52        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.5.4  Summary and Conclusions - Effects on Airway Responsiveness
 2           Ozone-induced AHR has been reported in a number of laboratory species at an exposure
 3      range of 0.5 to 1.0 ppm and in human asthmatics at ambient levels. In asthmatics, O3 is thought
 4      to exacerbate AHR to specific allergens by nonspecifically increasing AHR. New studies have
 5      demonstrated that AHR in asthmatics is due in part to chronic inflammation and airway
 6      remodeling. Animal studies have shown that O3 exposure can augment OVA-induced AHR.
 7      Importantly, there is a temporal relationship between inflammatory cell influx and O3-induced
 8      AHR, but inflammation is not a prerequisite of AHR.  Repeated O3 exposures enhance AHR,
 9      possibly by modulating rapidly adapting airway receptors or by altering the structure of
10      conducting airways.
11           Currently reported investigations on AHR with repeated O3 exposure to nonsensitized
12      laboratory animals have shown equivocal results, especially at the most relevant ambient O3
13      concentrations of <0.3  ppm. The few available studies in sensitized laboratory animals are
14      consistent with the O3-induced exacerbation of AHR reported in atopic humans with asthma (see
15      Chapter 6) but the results are difficult to extrapolate because of interindividual and interspecies
16      differences in responsiveness to bronchoprovocation and possible adaptation of airway
17      responsiveness with long-term, repeated O3 exposures. Therefore, further studies in laboratory
18      animals are needed to investigate responses to the different challenges in relation to
19      measurements of airway inflammation and the other physiological and structural factors  known
20      to contribute to airway responsiveness in human subjects.
21           Important new information indicates that rapid shallow breathing in response to O3 causes
22      a more evenly distributed injury pattern rather than protects from injury.  New insights into the
23      mechanisms of O3-induced AHR suggest that: (1) exercise-induced bronchoconstriction may be
24      mediated by changes in tonicity of the bronchial smooth muscles; (2) vagally-mediated
25      mechanisms may affect tracheal epithelial function and increase central airway reactivity;
26      (3) O3 may induce direct vascular constriction; (4) O3 may destroy neural endopeptidases in
27      airway epithelial  cells,  thus preventing the inactivation of SP; and (5) repeated O3 exposures may
28      diminish neutrophilic inflammation, depress cell proliferation, and cause cumulative distal
29      airway lesions.
30
        August 2005                              5-53        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.6   Genotoxicity Potential of Ozone
 2           There has been an historical interest in the ability of ground-level pollution to cause cancer,
 3      especially lung cancer.  This interest has been amplified in recent years by results of an
 4      epidemiologic study that suggest association of increased risks of incident lung cancer with
 5      elevated long-term ambient concentrations of O3, PM10, and SO2 in nonsmoking California males
 6      (Beeson et al., 1998; Abbey et al., 1999).  However, another larger, nationwide American Cancer
 7      Society study (Pope et al., 2002) showed no significant effect of O3 on mortality risk, but
 8      positive associations between warm season (July-September) O3 concentrations and
 9      cardiopulmonary mortality. Studies of children and young adults of southwest metropolitan
10      Mexico City, repeatedly exposed to high levels of O3, PM, NOX, aldehydes, metals, and other
11      components in a complex ambient mixture, also report DNA damage in blood leukocytes and
12      nasal epithelial cells (Valverde et al., 1997; Calderon-Garciduefias et al., 1999) and abnormal
13      nasal biopsies (Calderon-Garciduefias et al., 2001a). (See Chapter 6 for a discussion of the
14      human studies.)
15           A number of experimental studies have been done to explore the mutagenic/carcinogenic
16      potential of O3. In vitro studies are difficult to interpret due to very high exposure levels and
17      culture systems that allowed the potential formation of artifacts. Some recently published in
18      vivo exposure studies (see Table AX5-13) found increased DNA strand breaks in respiratory
19      cells from guinea pigs (Ferng et al., 1997) and mice (Bornholdt et al.,  2002) but, again, only on
20      exposure to high doses of O3 (1 ppm for 72 h and 1 or 2 ppm for 90 min, respectively).
21           Exposing the A/J mouse strain (known to have a high incidence  of spontaneous pulmonary
22      adenomas) to 0.12,  0.50, and 1.0 ppm O3 for 6 h/day, 5 days/week for up to 9 months, Witschi
23      et al.  (1999) did not find O3 exposure-related differences in lung tumor multiplicity or incidence.
24      Similarly, in  a subchronic  exposure study (B6C3FJ mice to 0.5 ppm O3 for 6 h/day, 5 days/week
25      for 12 weeks) Kim et  al. (2001) did not find statistically significant increases in the incidence of
26      lung tumors.  Significant differences in mean body weight as well as mean absolute and relative
27      weights of several organs (e.g., liver, spleen, kidney, testes, and ovary) were observed  between
28      O3-exposed and air-exposed mice. Histopathologic examination of major organs revealed
29      oviductal carcinomas  in 3/10 O3-exposed female mice.
30
        August 2005                              5-54        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.6.1   Summary and Conclusions - Genotoxicity Potential of Ozone
 2           The weight of evidence from new experimental studies does not appear to support
 3      ambient O3 as a pulmonary carcinogen in laboratory animal models. These new data are in
 4      agreement with a study of carcinogenicity of O3 from the NTP study (National Toxicology
 5      Program,  1994; Boorman et al., 1994), which was negative in male and female rats, ambiguous
 6      in male mice, and positive only in female mice at high concentrations of O3 (i.e., 1.0 ppm).
 7      As none of the new experimental studies of genotoxicity provided lifetime exposure durations
 8      such as those used in NTP cancer studies, the observation of no effects must be tempered by
 9      consideration of the limited duration of the exposure. Overall, then, the new animal studies are
10      inconclusive as are the epidemiologic studies discussed in Chapter 7, which may be due to
11      significant species differences in this health endpoint. Also, O3 could act as a co-carcinogen
12      functioning to stimulate hyperplasia. In epidemiology studies, exposures typically consist of
13      mixtures of co-pollutants, some of which are known carcinogens (see Section 5.4.3).
14
15
16      5.3    SYSTEMIC EFFECTS OF OZONE EXPOSURE
17           Ozone indirectly affects organs beyond the respiratory system due to O3 reaction products
18      entering the bloodstream and being transported to target sites.  Extra-pulmonary effects could
19      also be  due to the exposure-related production of mediators, metabolic products and cell
20      trafficking.  Although systemic effects are of interest and indicate a very broad array
21      of O3 effects, they  are of limited influence and difficult to interpret. By protecting from
22      respiratory tract effects, these systemic effects will likely be protected against also. Systemic
23      effects are only summarized briefly here and in Table AX5-14.
24
25      5.3.1  Neurobehavioral Effects
26           Animal behavior, both motor activity and operant behavior, has been shown to be
27      suppressed by acute O3 exposures (3 to 6 h) of 0.12 ppm. There is a dose dependent decrease in
28      activity with increasing exposure levels. Additionally, these lowered activity levels tend to
29      attenuate with longer exposure periods. New studies in adult laboratory animals confirm that
30      environmentally relevant O3 concentrations from 0.2 to  1.0 ppm can decrease motor activity and
31      affect short- and long-term memory, as tested by passive avoidance conditioning in 4 h

        August 2005                              5-55        DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposures in rats (Rivas-Arancibia et al., 1998; Avila-Costa et al., 1999; Dorado-Martinez et al.,
 2      2001), or water-maze learning tasks in mice following a 30-day exposure (Sorace et al., 2001).
 3      The effects have been attributed to reactive oxygen/nitrogen species and/or ozonation products.
 4      The memory deficits could be blocked by administration of vitamin E (Guerrero et al, 1999) or
 5      taurine (Rivas-Arancibia et al., 2000). Increased freezing and decreased exploratory behaviors
 6      were accompanied by decreased serotonin levels and increased levels of NO, glutamate,
 7      dopamine and striatal lipoperoxidation in rats exposed to 1  ppm of O3 for 4 h (Rivas-Arancibia
 8      et al., 2003). The O3-exposed animals also demonstrated neuronal cytoplasm and dendrite
 9      vacuolation and dilation of rough endoplasmic reticulium cisterns, which the authors interpret  as
10      a neurodegenerative process resulting from the oxidative stress of acute O3 exposure.  Nifio-
11      Cabrera et al. (2002) demonstrated that a 0.7 ppm O3 exposure for 4 h can induce ultrastructural
12      alterations in the hippocampus and prefrontal cortex in aged rats.  These are areas of the brain
13      where degenerative age-related changes in learning and memory functions have been reported
14      (Bimonte et al., 2003).
15           Paz (1997) reviewed a series of studies that demonstrated significant alterations of
16      electroencephalographic (EEG) patterns during sleep in animals acutely  exposed to O3 (0.35 to
17      1.0 ppm). Rats and cats both showed loss of paradoxical sleep time after 2 to 8 h of O3 exposure
18      (Paz and Bazan-Perkins,  1992; Paz and Huitron-Resendiz,  1996).  Increased total wakefulness,
19      alterations in circadian rhythm, and a permanent 50% loss  of paradoxical sleep time were shown
20      in rat pups born to dams exposed to 1.0 ppm O3 during gestation (Haro and Paz, 1993). Effects
21      on sleep patterns were associated with alterations in brain neurotransmitter levels (Huitron-
22      Resendiz et al., 1994; Gonzalez-Pifia and Paz, 1997) thought to be caused by O3 reaction
23      products or prostaglandins (Koyama and Hayaishi,  1994).  The permanent effects in pups caused
24      by high O3 exposure during gestation were attributed to the diminished antioxidant capability of
25      fetal tissue (Gunther et al.,  1993).
26           High, nonambient levels of O3 (e.g., >1.0 ppm) affect visual and olfactory neural pathways
27      in the rat. For  example, Custodio-Ramierez and Paz (1997) reported a significant delay in visual
28      evoked potentials recorded in the visual cortex and  the lateral geniculate nucleus of male Wistar
29      rats acutely exposed to high levels of O3 (1.5, and 3.0  ppm  for 4 h).  Colin-Barenque et al.
30      (1999), using the same strain, reported cytological and ultrastructural changes in the granule
31      layer of the olfactory bulb after a 4-h exposure to 1  to 1.5 ppm O3.  Although these neural effects

        August 2005                               5-56         DRAFT-DO NOT QUOTE OR CITE

-------
 1      are thought to be caused by O3 reaction products, especially free radicals, the studies do not add
 2      much to an understanding of the underlying mechanisms.
 3
 4      5.3.2   Neuroendocrine Effects
 5          Early studies suggested an interaction of O3 with the pituitary-thyroid-adrenal axis because
 6      thyroidectomy, hypophysectomy, and adrenalectomy protected against the lethal effects of high
 7      concentrations of O3 Concentrations of 0.7 to 1.0 ppm O3 for a 1 day exposure in male rats
 8      caused changes in the parathyroid; thymic atrophy; decreased serum levels of thyroid stimulating
 9      hormone, triiodothyronine (T3), thyroxine (T4), free T4, and protein binding; and increased
10      prolactin.  In more recent studies, increased toxicity to O3 was reported in hyperthyroid rats by
11      Huffman et al. (2001) and T3 supplementation was shown to increase metabolic rate and
12      pulmonary injury in the lungs of O3-treated animals (Sen et al., 1993).
13          The mechanisms by which O3 affects neuroendocrine function are not well understood.
14      Cottet-Emard et al. (1997) examined catecholamine activity in rat sympathetic efferents and
15      brain areas of prime importance to adaptation to environmental stressors. Exposures of
16      0.5 ppm O3 for 5 days caused inhibition of norepinephrine turnover in heart (-48% of the
17      control level) but not in lungs and failed to modify the tyrosine hydroxylase activity in superior
18      cervical ganglia and the catecholamine content in the adrenal glands. In the CNS, O3 inhibited
19      tyrosine hydroxylase activity in noradrenergic brainstem cell groups and decreased
20      catecholamine turnover in the cortex (-49%) and striatum (-18%) but not in the hypothalamus.
21      This suggests that high ambient levels of O3  can produce marked neural disturbances in
22      structures involved in the integration of chemosensory inputs, arousal, and motor control, effects
23      that may be responsible for some of the behavioral effects seen with O3 exposure.
24
25      5.3.3   Cardiovascular Effects
26          Studies of the effects on hematological parameters and blood chemistry in rats have shown
27      that erythrocytes are a target of O3.  Exposures to 1.0 ppm O3 for 3 h have been found to
28      decrease heart rate (HR), mean arterial pressure (MAP), and core temperature (Tco) and to induce
29      arrhythmias with some  exposures in rats. These effects are more pronounced in adult and awake
30      rats than in younger or  sleeping animals.  Exposures of 0.2 ppm for 48 h have been shown to
        August 2005                              5-57        DRAFT-DO NOT QUOTE OR CITE

-------
 1      cause bradycardia, while exposures of 0.1 ppm for 3 days have been shown to cause
 2      bradyarrhythmia in these animals.
 3           A more recent study of rats exposed to FA for 6 h, followed 2 days later by a 5 h exposure
 4      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
 5      to 0.5 ppm O3 used the head-out plethysmograph for continuous measurements (Arito et al.,
 6      1997). Each of the O3 exposures was preceded by a 1 h exposure to FA.  Transient rapid shallow
 7      breathing with slightly increased HR appeared 1-2 min after the start of O3 exposures and was
 8      attributed to an olfactory response. Persistent rapid shallow breathing with a progressive
 9      decrease in FIR occurred with a latent period  of 12 h. During the last 90-min of exposure,
10      averaged values for relative minute ventilation tended to decrease with the increase in O3
11      concentration for young (4-6 mo) but not old (20-22 mo) rats.
12           Studies utilizing radiotelemetry transmitters in unanesthetized and unrestrained rats,
13      Watkinson et al. (1995; 2001) and Highfill and Watkinson (1996) demonstrated that when HR
14      was reduced during a 5 day 0.5 ppm O3 exposure, the Tco and activity levels also decreased. The
15      decreases in Tco and blood pressure reported by in these studies and by Arito et al. (1997)
16      suggest that the changes in ventilation and FIR are mediated through physiological and
17      behavioral defense mechanisms in an attempt to minimize the irritant effects of O3 inhalation.
18      Decreased activity was previously reported in laboratory animals during exposure to O3
19      (see above).
20           Similar cardiovascular and thermoregulatory responses in rats to O3 were reported by
21      Iwasaki et al. (1998).  Repeated exposure to 0.1, 0.3, and 0.5 ppm O3 8 h/day for 4 consecutive
22      days caused disruption of circadian rhythms of HR and Tco on the first and second exposure days
23      that was concentration-dependent. The decreased HR and Tco recovered to control values on the
24      third and fourth days of O3 exposure.
25           The thermoregulatory response to O3 was further  characterized by Watkinson et al. (2003).
26      Male Fischer-344 rats were exposed to 0.0 ppm for 24 h/day (air), 0.5 ppm for 6 h/day
27      (intermittent) or 0.5 ppm for 23 h/day (continuous) at 3 temperatures, 10 °C (cold), 22 °C
28      (room), or 34 °C (warm).  Another protocol examined the effects of O3 exposure (0.5 ppm) and
29      exercise described as  rest, moderate, heavy or CO2-stimulated ventilation. Both intermittent and
30      continuous O3 exposure caused decreases in HR and Tco and increases in BALF inflammatory
31      markers.  Exercise in  FA caused increases in  HR and Tco while exercise in O3 caused decreases

        August 2005                              5-58       DRAFT-DO NOT QUOTE OR CITE

-------
 1      in those parameters.  Carbon dioxide and O3 induced the greatest deficits in HR and Tco.  Several
 2      factors were suggested that may modulate the hypothermic response, including dose, animal
 3      mass, and environmental stress.
 4          Laboratory animals exposed to relatively high O3 concentrations (>0.5 ppm) demonstrate
 5      tissue edema in the heart and lungs. This may be due to increased circulating levels of atrial
 6      natriuretic factor (ANF), which is known to mediate capillary permeability, vasodilation, and
 7      blood pressure (Daly et al., 2002). Increased levels of ANF were reported in the heart, lungs,
 8      and circulation of rats exposed to 0.5 ppm O3 for 8 h (Vesely et al., 1994a,b,c).
 9          Earlier work demonstrated O3-induced release of functionally active PAF from rodent
10      epithelial cells and the presence of PAF receptors on AMs.  New work examining lipid
11      metabolism (Section 5.2.1.4) and mediators of inflammatory response and injury
12      (Section 5.2.3.4) confirm these earlier studies that PAF (Kafoury et al., 1999) and PAF
13      receptors (Longphre et al., 1999) are involved in responses to O3. In addition to the role of PAF
14      in pulmonary inflammation and hyperpermeability, this potent inflammatory mediator may have
15      clotting and thrombolytic effects, though this has not been demonstrated experimentally (see
16      Figure 5-2). This cardiovascular effect may explain, in part, epidemiologic findings of heart
17      attack and stroke (see Chapter 7).  The findings of Pulfer and Murphy (2004); Pulfer et al.,
18      (2005); Section 5.2.1.4), describing the in vitro  and in vivo production of two biologically active
19      oxysterols, are also suggestive of a mechanism whereby O3  exposure may be implicated in the
20      increased risk of cardiopulmonary disease.
21
22      5.3.4    Reproductive and Developmental Effects
23          Early studies of pre- and postnatal exposure to O3 were performed at relatively high
24      concentrations.  Teratogenic effects were not observed with intermittent exposures of 0.44 to
25      1.97 ppm O3 during any part of gestation.  Continuous exposure during mid-gestation increased
26      the resorption of embryos while exposures during late gestation delayed some behavioral
27      developments (e.g., righting, eye opening). There were no effects on neonatal mortality up to
28      1.5 ppm O3, whereas some transient effects on weight gain were observed at exposures of
29      0.6 ppm O3.
30          More recent  studies tend to confirm previous conclusions that prenatal exposures to O3
31      concentrations <1.0 ppm do not cause major or  widespread somatic or neurobehavioral effects in

        August 2005                              5-59        DRAFT-DO NOT QUOTE OR CITE

-------
 1     the offspring of laboratory animals. These studies generally add some weight toward a negative
 2     interpretation of the importance of contributions of low, ambient O3 to lower birth weights and
 3     gross development defects reported in neonates born to women exposed to typical ambient
 4     pollution (e.g., Renner, 2002; Chen et al., 2002; Ritz and Yu, 1999).  Some postnatal O3
 5     exposure studies continue to find a few, subtle or borderline somatic and behavioral deficits that
 6     will require further research to better assess potential risk to developing humans.
 7           Studies of somatic and neurobehavioral development in female  CD-I mice exposed during
 8     pregnancy (days 7 to 17) to O3 (0, 0.4, 0.8, or 1.2 ppm) failed to show any O3 effects on
 9     reproductive or behavioral performance (Bignami et al., 1994). The study did find significant
10     decreases in body weight gain and delayed eye opening in pups in the 1.2 ppm exposure group.
11     The lack of effect on behavioral performance contrasts with earlier findings, which may be due
12     to the use of different species, differing exposure durations, cross-fostering used in the latter
13     study, different species, and exposure durations during pregnancy.  A second study using CD-I
14     mice exposed in utero from conception through day 17 of pregnancy to 0, 0.2, 0.4, and
15     0.6 ppm O3 found no significant deficits in reproductive performance, postnatal somatic and
16     neurobehavioral development, or adult motor activity (Petruzzi et al.,  1995).  A third study by
17     the same group (Petruzzi et al., 1999), using O3 exposures (0.3, 0.6, or 0.9 ppm) which continued
18     postnatally until weaning, showed subtle changes in handedness and morphine reactivity.
19     Exposures to 0.6 ppm O3 caused a reduced preference for the right paw in adulthood. Exposures
20     to 0.9 ppm O3 altered hot plate avoidance after i.p. treatment with morphine in adulthood.
21           CD-I mice exposed to 0.6 ppm O3 from birth through weaning demonstrated no
22     impairment of navigational performance during acquisition and only subtle changes during
23     reversal (DeH'Omo et al., 1995a). Additionally, there were no O3-induced effects on
24     reproductive performance, but offspring showed a significant reduction in body weight.  Effects
25     on neurobehavioral  development  with this exposure were minor, with some attenuation of
26     activity responses and impairment of passive avoidance acquisition (Dell'Omo et al. (1995b).
27     The offspring of CD-I mice continuously exposed from 30 days prior to the formation of
28     breeding pairs until  PND 17 to 0.0, 0.3, or 0.6 ppm O3 showed only small and selective effects
29     on somatic and sensorimotor development (Sorace et al., 2001).
30          Morphological changes were found in the anterior cerebellar lobe of rat pups born to dams
31     exposed during the entire gestation period to very high (1.0 ppm) O3 concentrations for 12 h/day.

       August 2005                              5-60        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (Rivas-Manzano and Paz, 1999). Additionally, the dams displayed significantly fewer
 2      implantations, increased rate of reabsorptions, a high incidence of spontaneous abortion, and
 3      offspring with low birth weight, as noted by previous investigators.
 4
 5      5.3.5   Effects on the Liver, Spleen, and Thymus
 6           Early investigations of the effects of O3 on liver centered on xenobiotic metabolism, and
 7      the prolongation of sleeping time, which was observed at 0.1 ppm O3. In some species, only
 8      adults and especially females were affected. In rats, high (1.0 to 2.0 ppm for 3 h) acute O3
 9      exposures caused increased production of NO by hepatocytes and enhanced protein synthesis
10      (Laskin et al., 1994; 1996).  The O3-associated effects shown in the liver are thought to be
11      mediated by inflammatory cytokines or other cytotoxic mediators released by activated
12      macrophages in the lungs (Vincent et al., 1996; Laskin et al., 1998; Laskin and Laskin, 2001).
13      Except for the earlier work on xenobiotic metabolism, the responses occurred only after very
14      high acute O3 exposures.
15           Examinations of the effects of O3 on spleen and thymus have shown that O3 primarily
16      affects T-cell mediated systemic immunity. As with the O3-associated effects shown in the liver,
17      most of the statistically significant changes occurred after acute exposures to very high O3
18      concentrations and relate to systemic oxidative stress.  Using more relevant ambient urban O3
19      exposure patterns, effects were not found on systemic immune function of rats.
20
21      5.3.6   Effects on Cutaneous and Ocular Tissues
22           Ozone exposure not only affects various organ systems, when inhaled, but also has direct
23      effects on the exposed skin and eyes.  The outermost layer of the skin (stratum corneum; SC)
24      may be oxidized, which can lead to compromise of the skin barrier and an epidermal
25      proinflammatory response ( Weber et al., 2001; Thiele, 2001). These effects are found only at
26      very high concentrations (>l-5 ppm) and have not been shown at more relevant ambient levels
27      of exposure.  The skin possesses a well-developed defense system against oxidative stress,
28      utilizing nonenzymatic (e.g., vitamin C and E, glutathione, uric acid,  a-tocopherol) and
29      enzymatic (e.g., superoxide dismutase, catalase, glutathione reductase and peroxidase)
30      antioxidants (Cross et al., 1998).  Ocular tissues have similar antioxidant protective function as
31      the skin but are not as well studied (Mucke, 1996; Rose et al., 1998).  Effects of ground-level

        August 2005                              5-61        DRAFT-DO NOT QUOTE OR CITE

-------
 1      smog on the eyes have been reported but generally are attributed to related photochemical
 2      oxidants like peroxyacetyl nitrate (Vyskocil et al., 1998) or possibly to atmospheric O3
 3      precursors or reaction products like aldehydes. As in other tissues, O3 may have disparate
 4      high-dose and low-dose mechanisms of effect on skin and eyes, so results must be interpreted
 5      in this light.
 6          Hairless mice (SKH-1) exposed to O3 (0.8 to 10 ppm for 2 h) were used to demonstrate that
 7      O3 depletes the low molecular weight antioxidants (e.g., a-tocopherol, vitamin C, glutathione,
 8      uric acid) in the SC at > 1.0 ppm and causes increased MDA at >5 ppm (Weber et al, 1999, 2000,
 9      2001). Valacchi et al. (2000) demonstrated that preexposure to 0.5 O3 for 2 h followed by low-
10      dose ultraviolet (UV) radiation (0.33 MED) caused depletion of a-tocopherol. This suggests that
11      combined low doses  of UV radiation and near-ambient levels of O3 may cause oxidative stress
12      on the SC. Prolonged exposure to 0.8 ppm O2 for 6 h also induces cellular stress responses that
13      included the formation of HNE protein adducts, HSP27, and heme-oxygenase-1 in the deeper
14      cellular layers of the  skin that continued for up to  18 h after O3 exposure, followed by repair
15      processes (Valacchi et al., 2003).
16          The importance of O3 and UV-induced cellular protein oxidation found in murine skin
17      models to possibly similar environmentally-induced changes in human SC keratins was
18      identified by Thiele et al.  (1998, 1999) and Thiele (2001). Using the presence of carbonyl
19      groups in proteins as a marker of reactive oxygen mediated protein oxidation, they reported
20      higher carbonyl levels in the upper SC from the tanned skin of humans and in the skin of healthy
21      human volunteers exposed to model chemical oxidants (e.g., hypochlorite, benzoyl peroxide)
22      that were inversely correlated with vitamin E levels. The environmentally-induced oxidative
23      damage identified in  human SC represents an early pathophysiological stage in the development
24      of barrier disruption and inflammation, and possibly has implications for the process of
25      desquamation.  The relevance of potentiation of environmental oxidative stress by O3 exposure
26      of human skin needs  further study.
27
28      5.3.7    Summary and Conclusions - Systemic Effects of Ozone
29          Neurobehavioral effects of O3 at concentrations of 0.2 to 1.0 ppm include decreased motor
30      activity,  short- and long-term memory deficits, increased freezing behavior, and decreased
31      exploratory behaviors. These effects have been associated with reactive oxygen/nitrogen

        August 2005                              5-62        DRAFT-DO NOT QUOTE OR CITE

-------
 1      species, ozonation products, altered neurotransmitter levels, morphological changes in several
 2      brain regions, and altered EEG patterns during sleep.  Neuroendocrine effects of O3 include
 3      morphological and hormonal changes in the pituitary-thyroid-adrenal axis at concentrations
 4      of-0.75 ppm and alterations of visual and olfactory neural pathways at concentrations >1 ppm.
 5      Mechanisms underlying these effects are not understood at this time.  Cardiovascular effects
 6      of O3 at concentrations of 0.3 to 0.5 ppm include decreased HR, Tco, and BP, which have been
 7      termed a hypothermic response. Concentrations of O3 >0.5 ppm cause tissue edema (possibly
 8      mediated by ANF).
 9           Prenatal exposures to O3 concentrations <1.0 ppm did not cause noticeable somatic or
10      neurobehavioral effects in offspring, while concentrations of 1.0 to 1.5 ppm caused varying
11      effects on neonatal mortality.  Some studies have shown an effect of O3 on liver xenobiotic
12      enzymes at concentrations as low as 0.1  ppm, while other studies have shown no alterations in
13      metabolic enzymes at even 1 ppm, with the effects appearing to be highly-species specific.
14      Effects on spleen and thymus appear to only occur at high O3 concentrations (>1.0 ppm), while
15      relevant ambient, urban exposures have no effect on systemic immune function in rats.  Effects
16      of O3 on cutaneous and ocular tissue are only seen at high, nonrelevant concentrations.
17
18
19      5.4   INTERACTIONS OF OZONE WITH OTHER CO-OCCURRING
20            POLLUTANTS
21           Ozone is part of a complex mixture of air pollutants with a composition and pattern that
22      varies geographically and temporally (by hour of the day, day of the week, and season).  Health
23      effects caused by the complex mixture are undoubtedly different (either subtly or significantly)
24      from the additive effects of a few of the hundreds of compounds present. The only disciplinary
25      approach that can evaluate a "real-world" complex mixture is epidemiology (Chapter 7).
26      However, because of the difficulty in evaluation of causative factors and quantitative
27      relationships in epidemiology studies, it is useful to consider animal toxicological studies of
28      mixtures.  Such studies can be divided into three categories: (1) ambient air mixtures,
29      (2) laboratory-generated complex mixtures (e.g., gasoline combustion mixtures having
30      ultraviolet-irradiation, other reaction mixtures with O3 and several other components), and
31      (3) binary mixtures.  In most cases, experimental designs in the first two classes did not have

        August 2005                              5-63        DRAFT-DO NOT  QUOTE OR CITE

-------
 1      an O3-only group, making it difficult to impossible to discern the influence of O3.  The more
 2      recent mixture studies that are discussed here typically have been with NO2, sulfuric acid
 3      (H2SO4), or ammonium sulfate ([NH4]2SO4).
 4           Interpreting the mixture studies in terms of real-world risk is difficult because laboratory
 5      exposure patterns do not always represent real-world exposure patterns. For example, in the real
 6      world, nitrogen dioxide (NO2) often peaks before O3 peaks, with a mixture occurring between
 7      the peaks, but most laboratory exposures used mixtures only.  Also, most studies of O3 and NO2
 8      mixtures used ambient levels  of O3 and levels of NO2 high above ambient. As shall be seen, all
 9      interaction possibilities have occurred, depending upon the composition of the mixture, the
10      endpoint examined, and the exposure regimen.  In some cases, no interaction was found. Most
11      often, additivity (the effects of the mixture are equal to the sum of the effects of the individual
12      components) or synergism (the effects of the mixture are greater than the sum of the effects of
13      the individual components) was observed. Antagonism (the effects of the mixture are less than
14      the sum of the individual components) was rarely found.
15
16      5.4.1  Ozone and  Nitrogen Oxides
17           The most commonly studied copollutant in binary mixtures with O3 is NO2.  Both early
18      work and more recent studies indicate that, although interaction  may occur between these two
19      pollutants, in general, O3 often masked the effects of the NO2 or accounted for most of the
20      response, due to the greater toxicity of O3. Very generally, additivity occurred after acute
21      exposure and synergism occurred with prolonged exposure. Interpreting the mixture studies is
22      challenging because laboratory exposure patterns rarely simulate real-world exposure patterns.
23      In the case of NO2 and O3, NO2 typically peaks  before O3, with a mixture occurring between the
24      peaks, but most laboratory exposures used mixtures only.  Also, most studies of O3 and NO2
25      mixtures used ambient levels  of O3 and levels of NO2 high above ambient. Table AX5-15 lists
26      more recent studies evaluating coexposures to NO2 and O3.
27           Chronic exposures of rats to O3 (0.8 ppm) and NO2 (14.4 ppm) for 6 h/day caused
28      development of respiratory insufficiency and severe weight loss. Half of these animals died after
29      55 to 78 days of exposure due to severe fibrosis (Farman et al., 1997). Increased total lung
30      collagen and elastin were observed, with loss of mature collagen, suggesting breakdown and
31      remodeling of the lung parenchyma. Morphological examination following these coexposures

        August 2005                              5-64        DRAFT-DO NOT QUOTE OR CITE

-------
 1      demonstrates a sequence of events starting with increasing inflammatory and mild fibrotic
 2      changes for the first 3 weeks of exposure stabilized or even reduced changes after 4 to 6 weeks,
 3      and severe increases over 7 to 9 weeks of exposure (Farman et al., 1999).  This suggests that
 4      repair processes occurring during the middle 4 to 6 weeks of exposure become overwhelmed,
 5      leading to progressive fibrosis after 7 to 8 weeks of exposure. When the coexposure was
 6      extended for 90 days, lesions were noted far into the acinus, but the extent of tissue involvement
 7      was the same after 7, 78, and 90 days of exposure. At the end of exposure, high levels of
 8      procollagen types I and III mRNA were observed within central acini in the lungs from the
 9      combined exposure  group but not in lungs from the rats exposed to O3 or NO2 alone.
10           Sprague-Dawley rats exposed to 0.3 ppm O3 and the combined exposure of O3 and
11      1.2 ppm NO2 for 3 d demonstrated significant DNA single-strand breaks in AMs (Bermudez
12      et al., 1999).  No changes were caused by NO2-only exposure.  The same exposures stimulated
13      the activity of polyADPR synthetase, suggesting a response to lung cellular DNA repair caused
14      by oxidant-induced lung injury (Bermudez, 2001). The laboratory animal model of progressive
15      pulmonary fibrosis,  utilizing long-term, combined O3  (0.4 to 0.8 ppm) and high-level NO2 (7 to
16      14 ppm) exposure, causes an initial acute pulmonary inflammation, followed by adaptation and
17      repair, and eventually causing pulmonary fibrosis after 6 to 13 weeks of exposure (Ishii et al.,
18      2000a; Weller et al., 2000). Unfortunately, this model is not very useful for understanding
19      potential interactive effects of ambient concentrations of O3 and NO2.
20
21      5.4.2  Ozone and Other Copollutants
22      Ozone and Formaldehyde
23           Early studies with combined exposures to O3 and formaldehyde (HCHO) found evidence
24      of both synergistic and non-interactive effects. Newer work listed in Table AX5-16 includes
25      studies of biochemical and histopathological endpoints in rats exposed to 0.4 ppm O3 and
26      3.6 ppm HCHO, alone and combined, for 8 h/day for 3 days (Cassee and Feron, 1994).  They
27      demonstrated no interactive effects in the nasal respiratory epithelium, despite the high levels of
28      HCHO when compared to typical ambient levels of 1  to 10 ppb (e.g., Rehle et al., 2001). Mautz
29      (2003) studied changes in breathing pattern and epithelial cell proliferation using exposures of
30      0.6 ppm O3 and 10 ppm HCHO alone  and in combination for 3 h with exercise at two times
31      resting ventilation. Even with exercise, HCHO does not substantially penetrate to the lower

        August 2005                               5-65       DRAFT-DO NOT QUOTE OR CITE

-------
 1      respiratory tract to interact with O3 and does not alter breathing patterns to modify local O3 dose.
 2      Parenchymal injury was, therefore, due to O3 alone. In the nasal transitional epithelium and in
 3      the trachea, however, combined exposure produced additive effects due to the increased volume
 4      of toxicants during exercise. No other combined pollutant studies have been published in the
 5      peer-reviewed literature, although two studies compared the respiratory effects of O3 to HCHO.
 6      Nielsen et al., (1999) compared upper airway sensory irritation caused by HCHO concentrations
 7      up to 4 ppm to the lower airway irritation caused by O3.  Using BALB/c mice, they continuously
 8      measured fB, VT, expiratory flow, T;, Te, and respiratory patterns during acute, 30-min exposures.
 9      They reported a no effect level of 0.3 ppm for HCHO and 1.0 ppm for O3.
10           Thus, O3 and HCHO do not appear to have additive effects, except during exercise, and
11      that is due to increased volume of gas reaching the tissue. Any possible synergism occurs in the
12      nasal epithelium.  HCHO exerts its effects primarily in the upper respiratory tract, whereas the
13      primary site of acute cell injury from O3 occurs in the conducting airways. EPA is currently
14      completing a toxicological and epidemiological review and risk characterization for
15      formaldehyde.
16
17      Ozone and Tobacco Smoke
18           Early studies of combined exposures of O3 (1 ppm) and tobacco smoke demonstrated
19      altered airway responsiveness to inhaled bronchoconstrictor challenge and tracheal vascular
20      permeability in guinea pigs.  Table AX5-17 lists studies  completed since the 1996 AQCD
21      evaluating coexposures of tobacco smoke and O3.
22           Wu et al. (1997) reported that inhalation of cigarette smoke evokes a transient
23      bronchoconstrictive effect in anesthetized guinea pigs. Total pulmonary resistance (RL) and
24      dynamic lung compliance (Cdyn) were compared before and after acute exposure to 1.5 ppm O3
25      for 1 h. Cigarette smoke alone (7 ml) at a low concentration (33%) induced a mild and
26      reproducible bronchoconstriction that slowly developed  and reached its peak after a delay
27      of >1 min. After O3  exposure, the same cigarette smoke inhalation challenge evoked an intense
28      bronchoconstriction that occurred more rapidly, reaching its peak within 20 s, and was sustained
29      for >2 min. Pretreatment with selective antagonists of neurokinin type 1 and 2 receptors
30      completely blocked the enhanced airway responsiveness suggesting that O3 exposure induced
        August 2005                              5-66        DRAFT-DO NOT QUOTE OR CITE

-------
 1     AHR to inhaled cigarette smoke, which resulted primarily from the bronchoconstrictive effect of
 2     endogenous tachykinins.
 3           The above studies were conducted with undiluted tobacco smoke and high O3
 4     concentrations.  To determine the effects of aged and diluted sidestream cigarette smoke (ADSS)
 5     as a surrogate of environmental tobacco smoke (ETS) on O3-induced lung injury, Yu et al.
 6     (2002) exposed male B6C3F1 mice to (1) FA, (2) ADSS, (3) O3, or (4) ADSS followed by O3
 7     (ADSS/O3). Exposure to 30 mg/m3 ADSS, 6 h/day for 3 days, followed by exposure to
 8     0.5 ppm O3 for 24 h was associated with a significant increase in the number of cells recovered
 9     by BAL compared with exposure to ADSS alone or O3 alone.  Neutrophils, lymphocytes, and
10     total protein levels in BAL were increased following the combined exposure when compared
11     with all other groups. Within the CAR, the percentage of proliferating cells was unchanged from
12     control following exposure to ADSS alone but was significantly elevated following exposure
13     to O3 and further augmented in a statistically significant manner in mice exposed to ADSS/O3.
14     Following exposure to O3 alone or ADSS/O3, the ability of AMs to release IL-6 under LPS
15     stimulation was significantly decreased, while exposure to ADSS alone or ADSS/O3 caused a
16     significantly increased release of TNFa from AMs under LPS stimulation. These data suggest
17     that ADSS  exposure enhances the sensitivity of animals to O3-induced lung injury.
18           Acute exposure to ETS also may make a healthy person more susceptible to sequential O3
19     exposure by affecting lung barrier function or the underlying epithelium.  Toxicological studies
20     with components of ETS (e.g., nicotine receptor agonists, acrolein, and oxidants) have shown
21     that the vagal bronchopulmonary C-fibers are stimulated by acute exposures that initiate both
22     central and local responses (Bonham et al., 2001; Mutoh et al., 2000).  The central responses
23     (e.g., tachypnea, cough, bronchoconstriction, increased mucous secretion) are more protective of
24     the lungs; however,  local responses may include increased sensitization of the C-fibers to other
25     irritants, including O3.  Active tobacco smokers should not be similarly affected because they
26     already have significant chronic airway inflammation and increased mucus production.  In fact,
27     chronic smokers appear to have diminished lung function responses to O3  (see  Chapter 6).
28
29     5.4.3  Complex (Multicomponent) Mixtures Containing Ozone
30           Ambient pollution in most areas is a complex mix of more than two  chemicals. A number
31     of new studies have examined the effects of exposure to multicomponent atmospheres

       August 2005                              5-67       DRAFT-DO NOT  QUOTE OR CITE

-------
 1      containing O3.  Some of these studies attempted to simulate photochemical reaction products
 2      occurring under actual atmospheric conditions. However, the results of these studies are often
 3      difficult to interpret because of chemical interactions between the components, as well as the
 4      resultant production of variable amounts of numerous secondary reaction products, and a lack of
 5      precise control over the ultimate composition of the exposure environment.  In addition, the role
 6      of O3 in the observed biological responses is often obscure. Prior studies using irradiated
 7      automobile exhaust mixtures containing total oxidant concentrations (expressed as O3) in the
 8      range of 0.2 to 1.0 ppm have demonstrated pulmonary function changes in several species.
 9           A more recent attempt has been made to examine multicomponent mixtures resulting from
10      the reaction of O3 with unsaturated hydrocarbons [e.g., isoprene (C5H8) and terpene (C10H16)],
11      producing HCHO, formic acid, acetone, acrolein, acetic acid, and other oxidation products, many
12      of which are strong airway irritants.  Wilkins et al. (2001) evaluated sensory irritation by
13      measuring mean fB in the mouse bioassay and found a 50% reduction after 30 min of exposure to
14      reaction products of O3 and isoprene. The mixture at this time period contained <0.2 ppm O3, so
15      the authors attributed the observed effects to the oxidation products. Clausen et al. (2001), using
16      the same mouse model, evaluated the reaction products of O3 and limonene. A 33% reduction in
17      mean fB was produced  after 30 min of exposure to the mixture containing <0.3 ppm O3, again
18      implicating the effects  of strong irritant products. Further work needs to be done with these
19      complex reaction mixtures because of their potential impact on the respiratory tract.  The results
20      would be particularly important, however, to the reaction of O3 indoors (see Chapter 3).
21           Pollutant mixtures containing acid aerosols comprise another type of commonly examined
22      exposure  atmosphere (studies summarized in Table AX5-18). Earlier studies that employed
23      simultaneous single,  repeated,  or continuous exposures of various animal species to mixtures of
24      acid sulfates and O3 found responses for several endpoints, including tracheobronchial
25      mucociliary clearance, alveolar clearance, pulmonary mechanics, and lung morphology, to be
26      due solely to O3  Some synergism was noted for bacterial infectivity, response to antigen, and
27      effects  on lung protein content and the rate of collagen synthesis.
28           More recent studies found some differences in airway responses to inhaled acid particle-O3
29      mixtures that may have been partly due to airway dosimetry. Various physical and chemical
30      mechanisms may be responsible (see Schlesinger, 1995). For example, physical adsorption or
31      absorption of O3 or its reaction products on a particle could result in transport to more sensitive

        August 2005                               5-68         DRAFT-DO NOT QUOTE OR CITE

-------
 1      sites, or to sites where O3, by itself, would not normally be reactive (Madden et al., 2000).
 2      Chemical reactions on the surface of particles can form secondary products that are more
 3      lexicologically active, or chemical characteristics of the particle may change the residence time
 4      or reactivity of oxidation products at the site of deposition.  The hypothesis that synergism
 5      between O3 and sulfates is due to decreased pH changing the residence time or reactivity of
 6      reactants, such as free radicals, was tested by Chen et al. (1995) and El-Fawal et al. (1995).
 7      Male New Zealand white rabbits were exposed for 3 h to 125 |ig/m3 H2SO4, 0.1, 0.3, or
 8      0.6 ppm O3, and to combinations.  Chen et al. (1995) demonstrated that decreased pH following
 9      exposure to acid aerosol was correlated with phagocytic activity and capacity of harvested
10      macrophages and that exposure to O3/ H2SO4 removed this relationship. El-Fawal et al. (1995)
11      showed that responsiveness of rabbit harvested bronchial rings to ACh was increased following a
12      3 h O3 exposure, but that 0.1 to 0.6 ppm O3/0.5 to 0.125 mg/m3 H2SO4 combinations resulted in
13      antagonism.
14           As discussed in Section 5.2.2.1, Churg et al. (1996) demonstrated increased uptake of
15      asbestos or TiO2 in response to 10 min O3 (up to 1.0 ppm) pre-exposure suggesting that low
16      concentrations of O3 may increase the penetration of some types of PM into epithelial cells.
17      Using human epithelial cell cultures, Madden et al. (2000) demonstrated a greater potency for
18      ozonized diesel PM to induce prostaglandin E2 production.  This suggests that 0.1 ppm O3 for
19      24 h can modify the biological activity of PM derived from diesel exhaust.
20           Effects of combined exposures of O3 and resuspended urban particles on cell proliferation
21      in epithelial cells  of the terminal bronchioles  and the alveolar ducts were examined by Vincent
22      et al. (1997) and Adamson et al. (1999). Rats exposed to 0.8 ppm O3 in combination with 5 or
23      50 mg/m3 particles for 4 h demonstrated greatly potentiated proliferative effects compared to O3
24      exposure alone. These findings using resuspended dusts, although at high concentrations, are
25      consistent with the studies demonstrating interaction between H2SO4 aerosols and O3. Effects of
26      acute coexposure to 0.6 ppm O3 and fine or ultrafine H2SO4 (0.5 to 0.3 mg/m3) aerosols on lung
27      morphology were examined by Kimmel et al. (1997). They demonstrated that alveolar septal
28      volume was increased in animals co-exposed to O3 and ultrafine, but not fine, H2SO4.
29      Interestingly, cell proliferation was increased only in animals co-exposed to fine H2SO4 and O3,
30      as compared to animals exposed to O3 alone.  Subchronic exposure to acid aerosols (20 to
31      150 |ig/m3 H2SO4) had no interactive effect on the biochemical and morphometric changes

        August 2005                              5-69        DRAFT-DO NOT QUOTE OR CITE

-------
 1      produced by either intermittent or continuous exposure to 0.12 to 0.2 ppm O3 for up to 90 days,
 2      which suggests that the interactive effects of O3 and acid aerosol coexposure in the lung
 3      disappeared during the long-term exposure (Last and Pinkerton, 1997).  Sindhu et al. (1998)
 4      observed an increase in rat lung putrescine levels after repeated, combined exposures to O3 and a
 5      nitric acid vapor for 40 weeks.
 6           Other studies have examined interactions between carbon particles and O3.  The
 7      interactions of intratracheally instilled carbon particles  followed by either a 7-day or 60-day
 8      exposure to 0.5 ppm O3 in rats was evaluated by Creutzenberg et al. (1995).  The carbon
 9      particles caused diminished phagocytotic capacity and chemotactic migration capability of AMs
10      that was stimulated by the subsequent O3 exposure.  Inflammatory responses following
11      exposures to low- and high-concentration mixtures of O3 and acidic aerosols (0.2 ppm O3 +
12      50 |ig/m3 carbon + 100 |ig/m3 H2SO4; 0.4 ppm O3 + 250 |ig/m3 carbon + 500 |ig/m3 H2SO4,
13      respectively) for 1 or 5 days was examined by Kleinman et al. (1999).  The response with
14      the O3-particle mixture was greater after 5 days (4 h/day) than after day 1. This contrasted
15      with O3 exposure alone (0.4 ppm), which caused marked inflammation on acute exposure,
16      but no inflammation after 5 consecutive days of exposure.
17           The effects of a mixture of elemental carbon particles, 0.2 ppm O3, and 0.5  mg/m3
18      ammonium bisulfate on rat lung collagen content and macrophage activity was examined by
19      Kleinman et al. (2000). Decreases in lung collagen, and increases in macrophage respiratory
20      burst and phagocytosis were observed relative to other pollutant combinations. Mautz et al.
21      (2001) used a similar mixture (i.e., elemental carbon particles, 0.16 to 0.59 ppm O3, ammonium
22      bisulfate 0.5 to 0.22 mg/m3, but with 0.11 to 0.39 ppm NO2 also) and exposure regimen as
23      Kleinman et al. (2000). Also observed were decreases in pulmonary macrophage Fc-receptor
24      binding and phagocytosis and increases in acid phosphatase staining. Bronchoalveolar epithelial
25      permeability and cell proliferation were increased.  Altered breathing-patterns  also were
26      observed, with some adaptations occurring.
27           Bolarin et al. (1997) exposed rats to 50 or 100 |ig/m3 carbon particles in combination with
28      ammonium bisulfate and  0.2 ppm O3. Despite 4 weeks of exposure, they observed no changes in
29      protein concentration in lavage fluid or blood prolyl 4-hydroxylase, an enzyme involved in
30      collagen metabolism.  Slight decreases in plasma fibronectin were present in animals exposed to
31      the combined pollutants versus O3 alone.  Thus, the potential for adverse effects in the lungs of

        August 2005                               5-70        DRAFT-DO NOT  QUOTE OR CITE

-------
 1      animals challenged with a combined exposure to particles and gaseous pollutants is dependent
 2      on numerous factors, including the gaseous co-pollutant, concentration, and time.
 3           In a complex series of studies, Oberdorster and colleagues examined the interaction of
 4      several pulmonary oxidative stress pollutants. Elder et al. (2000a,b) reported the results of
 5      combined exposure to ultrafine carbon particles (100 |ig/m3) and O3 (1 ppm for 6 h) in young
 6      and old Fischer 344 rats that were pretreated with aerosolized endotoxin.  In old rats, exposure to
 7      carbon and O3 produced an interaction that resulted in a greater influx in neutrophils than that
 8      produced by either agent alone. This interaction was not seen in young rats. Oxidant release
 9      from lavage fluid cells also was assessed and the combination of endotoxin, carbon particles,
10      and O3 produced an increase in oxidant release in old rats. This mixture produced the opposite
11      response in the cells recovered from the lungs of the young rats, indicating that the lungs of the
12      aged animals underwent greater oxidative stress in response to a complex pollutant mix of
13      particles, O3, and a biogenic agent. Johnston et al. (2000a; 2002) reported the results of
14      combined exposure to O3 (1.0  and 2.5 ppm for 4, 20, or 24 h) and low-dose endotoxin, or to O3
15      and endotoxin separately, in newborn and adult C57BL/6J mice. In the first study, adult (8 wk
16      old) mice showed greater sensitivity to  O3 than newborn (36 h old) mice on the basis of mRNAs
17      encoding for various chemokines and cytokines. In contrast, adult and newborn mice responded
18      similarly 2 h after endotoxin exposure (10 ng for 10 min), suggesting that age differences
19      in O3-generated inflammation  is secondary to epithelial cell injury. In the second study, 8 wk
20      old mice exposed to O3 (1 ppm for 24 h) followed by endotoxin (37.5 EU for 10 min) showed
21      increased responsiveness over either exposure alone, on the basis of increased expression of
22      chemokine and cytokine messages and increased BAL fluid levels of protein and PMNs.
23           Fanucchi et al. (1998) and Wagner et al. (2001a,b) examined the synergistic effect of
24      coexposure to O3 and endotoxin on the  nasal transitional epithelium of rats that also was
25      mediated, in part, by neutrophils.  Fisher 344 rats intranasally instilled with endotoxin and
26      exposed to 0.5 ppm O3, 8 h per day, for 3 days developed mucous cell metaplasia in the nasal
27      transitional epithelium, an area normally devoid of mucous cells; whereas, intratracheal
28      instillation of endotoxin (20 jig) caused mucous cell metaplasia rapidly in the respiratory
29      epithelium of the conducting airways. A synergistic increase of intraepithelial mucosubstances
30      and morphological evidence of mucous cell metaplasia were found in rat maxilloturbinates upon
31      exposure to both O3 and endotoxin, compared to each pollutant alone. A similar response was

        August 2005                               5-71        DRAFT-DO NOT QUOTE OR CITE

-------
 1      reported in OVA-sensitized Brown Norway rats exposed to 0.5 ppm O3, 8 h/day for 3 days
 2      (Wagner et al., 2002), indicating that coexposure to O3 and inflammatory biogenic substances
 3      like allergens (e.g., OVA) or bacterial endotoxin can augment epithelial and inflammatory
 4      responses in rat nasal passages.
 5           In follow-up studies, Wagner et al. (2003) reported that coexposure of rats to O3 and
 6      endotoxin also enhanced epithelial and neutrophilic inflammatory responses in the pulmonary
 7      airways. Fisher 344 rats were intranasally instilled with endotoxin and exposed to 1.0 ppm O3
 8      for 8 h, which was repeated 24 h later. Three days after the last exposure, BALF was analyzed
 9      for inflammatory cells and secreted mucosubstances (mucin SAC), and lung tissue was
10      processed for morphometric analysis. Endotoxin instillation alone caused a dose-dependent
11      increase in BALF neutrophils that was further increased 2-fold in O3-exposed rats given 20 jig
12      endotoxin, the highest dose.  Mucin glycoprotein SAC also was increased in the BALF at this
13      dose but not at lower endotoxin doses. Ozone exposure alone did not cause mucus
14      hypersecretion, but it did potentiate mucus secretion in rats given both 2 and 20 jig endotoxin
15      and increased intraepithelial mucosub stances 2-fold, which was further substantiated by
16      significant increases in mucin gene (rMucSAC) mRNA levels in the conducting airways.
17           The effect of O3 modifying the biological  potency  of PM (diesel PM and carbon black) was
18      examined by Madden et al. (2000) in rats. Reaction of NIST Standard Reference Material
19      # 2975 diesel PM with 0.1 ppm O3 for 48 hr increased the potency (compared to unexposed or
20      air-exposed diesel PM) to induce neutrophil influx,  total protein, and LDH in lung lavage fluid in
21      response to intratracheal instillation. Exposure  of the diesel PM to high, nonambient O3
22      concentration (1.0 ppm) attenuated the increased potency, suggesting destruction of the bioactive
23      reaction products. Unlike the diesel particles, carbon black particles exposed to 0.1 ppm O3 did
24      not exhibit an increase in biological potency, which suggested that the reaction of organic
25      components of the diesel PM with O3 were responsible for the increased potency.
26           Ulrich et  al. (2002) investigated the effect of ambient PM from Ottawa Canada (EHC-93)
27      on O3-induced inflammation.  Male Wistar rats  were exposed to 0.8 ppm O3 for 8 h and allowed
28      to recover before intratracheal instillation of 0.5, 1.5, and 5 mg of EHC-93 in 0.3 ml of saline.
29      The high concentrations of PM used were sufficient to induce pulmonary inflammation, which
30      was not exacerbated by pre-exposure to O3.  Rats from the combined exposure group did have
        August 2005                              5-72        DRAFT-DO NOT QUOTE OR CITE

-------
 1      higher and more persistent lung lavage protein and albumin levels, as well as increased plasma
 2      fibrinogen levels when compared to PM exposure alone.
 3           The interaction of PM and O3 was further examined in a murine model of OVA-induced
 4      asthma. Kobzik et al. (2001) investigated whether coexposure to inhaled, concentrated ambient
 5      particles (CAPs) from Boston, MA and to O3 could exacerbate asthma-like symptoms.  On days
 6      7 and 14 of life, half of the BALB/c mice used in this study were sensitized by intraperitoneal
 7      (ip) injection of OVA and then exposed to OVA aerosol on three successive days to create the
 8      asthma phenotype. The other half received the ip OVA but were exposed to a phosphate-
 9      buffered saline aerosol (controls).  The mice were further subdivided (n > 6I/group) and exposed
10      for 5  h to  CAPs, ranging from 63 to 1,569 |ig/m3, 0.3 ppm O3, CAPs + O3, or to FA. Pulmonary
11      resistance and airway responsiveness to an aerosolized MCh challenge were measured after
12      exposures. A small, statistically significant increase in pulmonary resistance and airway
13      responsiveness, respectively, was found in both normal and asthmatic mice immediately after
14      exposure to CAPs alone and to CAPs + O3 but not to O3 alone or to FA. By 24 h after exposure,
15      the responses returned to baseline levels.  There were no significant increases in  airway
16      inflammation after any of the pollutant exposures. In this well-designed study of a small-animal
17      model of asthma, O3 and CAPs did not appear to be synergistic. In further analysis of the data
18      using specific elemental groupings of the CAPs, the acutely increased pulmonary resistance
19      was found to be associated with the AISi fraction of PM. Thus, some components of
20      concentrated PM2 5 may affect airway caliber in sensitized animals, but the results are difficult
21      to extrapolate to people with asthma.
22           Animal studies have examined the adverse cardiopulmonary effects of complex mixtures in
23      urban and rural environments of Italy (Gulisano et al., 1997), Spain (Lorz and Lopez, 1997), and
24      Mexico (Vanda et al., 1998; Moss et al., 2001). Some of these studies have taken advantage of
25      the differences in pollutant mixtures of urban and rural environments to report primarily
26      morphological changes in the nasopharynx and lower respiratory tract (Gulisano et al.,  1997;
27      Lorz  and Lopez, 1997) of lambs and pigeons, respectively, after natural, continuous exposures to
28      ambient pollution.  Each study has provided evidence that animals living in urban air pollutants
29      have  greater pulmonary changes than those that would occur in a rural and presumably cleaner,
30      environment.  However, these studies either did not report ambient O3 levels, or reported only
31      annual means.

        August 2005                              5-73       DRAFT-DO NOT QUOTE OR CITE

-------
 1           The study by Moss et al. (2001) examined the nasal and lung tissue of rats exposed
 2      (23 h/day) to Mexico City air for up to 7 weeks and compared them to controls similarly exposed
 3      to FA. No inflammatory or epithelial lesions were found using quantitative morphological
 4      techniques; however, the concentrations of pollutants were low. Extrapolation of these results to
 5      humans is restricted, however, by uncontrolled exposure conditions, small sample sizes, and
 6      other unknown exposure and nutritional factors in the studies in mammals and birds, and the
 7      negative studies in rodents. They also bring up the issue of which species  of "sentinel" animals
 8      is more useful for predicting urban pollutant effects in humans. Thus, in these field studies, it is
 9      difficult to assign a specific role to any specific component of the mixture  for the significant
10      cardiopulmonary effects reported.
11           Similar morphological changes (Calderon-Garciduefias et al., 2000a; 2001) and chest X-ray
12      evidence of mild lung hyperinflation (Calderon-Garciduefias et al., 2000b) have been reported in
13      children residing in urban and rural areas of Mexico City. (See Chapter 7 for details of these
14      studies.) The ambient air in urban areas, particularly in southwestern Mexico City, is a complex
15      mixture of particles and gases, including high concentrations of O3 and aldehydes that previously
16      have been shown to cause airway inflammation and epithelial lesions in humans (e.g., Calderon-
17      Garciduefias et al., 1992, 1994, 1996) and laboratory animals (Morgan et al., 1986; Heck et al.,
18      1990; Harkema et al.,  1994, 1997a,b).  The described effects demonstrate a persistent, ongoing
19      upper and lower airway inflammatory process and chest X-ray abnormalities in children residing
20      predominantly in highly polluted areas. Again, extrapolation of these results to urban
21      populations of the United States is difficult because of the unique complex mixture of urban
22      air in Mexico City, uncontrolled exposure conditions, and other unknown exposure and
23      nutritional factors.
24
25      5.4.4   Summary and Conclusions - Interactions of Ozone with other
26              Co-occurring Pollutants
27           It is difficult to summarize the role that O3 plays in exposure responses to binary mixtures,
28      and  even harder to determine its role in responses to multicomponent, complex atmospheres.
29      Though the specific mechanisms of action of the individual pollutants within a mixture may be
30      known, the exact bases for toxic interactions have not been elucidated clearly.  Certain generic
31      mechanisms that may underlie pollutant interactions: (1) physical, involving adsorption  of one

        August 2005                              5-74        DRAFT-DO NOT QUOTE OR CITE

-------
 1      pollutant onto another and subsequent transport to more or less sensitive sites or to sites where
 2      one of the components of the mixture normally would not deposit in concentrated amounts
 3      (probably not involved in O3-related interactions); (2) production of secondary products that may
 4      be more lexicologically active than the primary materials, demonstrated or suggested in a
 5      number of studies as a basis for interaction between O3 and NO2 and between O3  and PM;
 6      (3) biological or chemical alterations at target sites that affect response to O3 or the copollutant,
 7      which which has been suggested to underlie interactions with mixtures of O3 and acid sulfates;
 8      4) O3- or copollutant-induced physiological change, such as alteration in ventilation pattern,
 9      resulting in changes in the penetration or deposition of one pollutant when another is present.
10      This has been implicated in enhanced responses to various O3-containing mixtures with exercise.
11           Evaluation  of interactions between O3 and copollutants is a complex procedure. Responses
12      are dependent on a number of host and environmental factors,  such that different studies using
13      the same copollutants may show different types or magnitudes of interactions. The occurrence
14      and nature of any interaction is dependent on the  endpoint being examined and is also highly
15      related to the specific conditions of each study, such as animal species, health status, exposure
16      method, dose, exposure sequence, and the physicochemical characteristics of the copollutants.
17      Because of this, it is difficult to compare studies,  even those examining similar endpoints, that
18      were performed under different exposure conditions.  Thus, any  description of interactions is
19      really valid only for the specific conditions of the study in question and cannot be generalized to
20      all conditions of exposure to a particular chemical mixture.  Furthermore, it is generally not
21      possible to extrapolate the effect of pollutant mixtures from studies on the effects of each
22      component when given separately. In any case, what can  be concluded from the  database is that
23      interactions of O3-containing mixtures are generally synergistic (antagonism has been noted in a
24      few studies), depending on the various factors noted above, and that O3 may produce more
25      significant biological  responses as a component of a mixture than when inhaled alone.
26      Furthermore, although most studies have shown that interaction  occurs only at higher than
27      ambient concentrations with acute exposure, some have demonstrated interaction at more
28      environmentally relevant levels (e.g., 0.05 to 0.1  ppm O3 with NO2) and with repeated exposures.
29
30
        August 2005                               5-75        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.5   EFFECTS OF OTHER PHOTOCHEMICAL OXIDANTS
 2           Peroxyacetyl nitrate (PAN) and peroxypropionyl nitrate (PPN) are the most abundant
 3      non-O3 oxidants in ambient air of industrialized areas, other than the inorganic nitrogenous
 4      oxidants such as NO2, and possibly HNO3. Ambient levels of PAN and PPN were reported to be
 5      decreasing over the 1990's and available air quality data (Grosjean et al., 2001; Grosjean, 2003;
 6      Jakobi and Fabian, 1997) indicate that present peak concentrations of PAN and PPN in ambient
 7      air from urban areas are in the low ppb range (e.g., <1 to 10 ppb). The levels found in nonurban
 8      areas are considerably lower (Gaffney et al., 1993).
 9           Reactions occur in the troposphere between O3 and hydrocarbons (e.g., d-limonene) to
10      produce epoxides, hydroperoxides, and peroxides. The majority  of the measured ambient
11      hydroperoxides produced is hydrogen peroxide (H2O2), although a small amount of organic
12      hydroperoxides (ROOH) also may be formed. Friedlander and Yeh (1998) have estimated that
13      atmospheric aerosols can carry as high as 1 mM of H2O2 and organic hydroperoxides (e.g.,
14      hydroxymethylhydroperoxide) in the associated water. In vitro cell and tissue damage are
15      induced by high concentrations of liquid phase H2O2 (50 jiM to 1 mM). Morio et al. (2001)
16      (see Table AX5-19) demonstrated  that a 2 h exposure of 10 and 20 ppb of inhaled H2O2 vapor
17      can penetrate the lower lung where it causes inflammation. It is likely that hygroscopic
18      components of PM transport ambient H2O2 into the lower lung and induce tissue injury as well.
19      Exposure of rats to a H2O2-fine particle mixture (215 or 429 |ig/m3 ammonium sulfate) resulted
20      in increased neutrophil influx, and production of inflammatory mediators by AMs (Morio et al.,
21      2001). Hygroscopic secondary organic aerosols generated by the O3/hydrocarbon reactions and
22      their co-occurrence with H2O2 also provides another possible mechanism, yet to be validated,
23      whereby H2O2 can be transported into the lower respiratory tract  (e.g., Friedlander and Yeh,
24      1998). Interaction of inhaled O3 with unsaturated fatty acids on cell membranes and mucus in
25      the airways generates epoxides, hydroperoxides, and secondary ozonation products such as
26      4-hydroxynonenal (see Section 5.2.1)
27           Inhalation toxicological information on the effects of the non-O3 oxidants has been limited
28      to a few studies on PAN, but at concentrations much higher (approximately 100- to 1,000 fold)
29      than levels typically found in ambient air.  Such high acute levels cause changes in lung
30      morphology, behavioral modifications, weight loss, and susceptibility to pulmonary infections.
31      Therefore, acute toxicity of PAN is much lower than O3, and it is unlikely that present ambient

        August 2005                              5-76        DRAFT-DO NOT QUOTE OR CITE

-------
 1     PAN levels would affect pulmonary function responses to O3 (reviewed in Vyskocil et al., 1998).
 2     Cytogenetic studies indicate that PAN is not a potent mutagen, clastogen, or DNA damaging
 3     agent in mammalian cells in vivo or in vitro at concentrations several orders of magnitude higher
 4     than the generally encountered ambient air levels in most cities (Vyskocil et al., 1998;
 5     Kligerman et al., 1995; Heddle et al., 1993).  Some studies suggest that PAN may be a weak
 6     bacterial mutagen at concentrations much higher than exist in present urban atmospheres
 7     (DeMarini et al., 2000; Kleindienst et al., 1990).
 8           An additional  level of complexity exists due to the possibility that other ambient oxidants
 9     may contribute to effects attributed to O3. As discussed in Chapter 2, both short-lived radicals
10     and secondary particles containing highly polar compounds are generated in the troposphere by
11     the same photochemical mechanisms that produce O3.  It is plausible that, in addition to the
12     direct effects of O3,  health effects are produced by ambient exposures to these gaseous and
13     particulate secondary compounds. Little is known regarding the composition of these reaction
14     products, and little research has been undertaken evaluating their toxicologic effects. Due to the
15     many oxidizing species present in the atmosphere, interpretation of toxicology data based on O3
16     exposures alone have the potential for underestimating health effects of ambient oxidant
17     mixtures.
18
19     5.5.1   Summary and Conclusions - Effects of Other Photochemical Oxidants
20           Concentrations of PAN and PPN (<1 to 10 ppb) in ambient air are unlikely to affect
21     pulmonary function or cause DNA damage. Levels of 10-20 ppm H2O2 can penetrate to the
22     lower lung directly or be transported there by PM, where inflammation can result; however,
23     ambient H2O2 levels of are typically < ~5 ppb.  As toxicology studies of other photochemical
24     oxidants are rare, quantitative scientific evaluations of possible health effects of environmental
25     exposures cannot be completed at this time.
26
        August 2005                              5-77        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES

  2       Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. J.; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
  3             Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir. Crit.
  4             Care Med. 159: 373-382.
  5       Adamson, I. Y. R.; Vincent, R.; Bjarnason, S. G. (1999) Cell injury and interstitial inflammation in rat lung after
  6             inhalation of ozone and urban particulates. Am. J. Respir. Cell Mol. Biol. 20: 1067-1072.
  7       Aizawa, H.; Shigyo, M; Nakano, H.; Matsumoto, K.; Inoue, H.; Hara, N. (1999a) Effect of the Chinese herbal
  8             medicine, Bakumondo-to, on airway hyperresponsiveness induced by ozone exposure in guinea-pigs.
  9             Respirology 4: 349-354.
10       Aizawa, H.; Shigyo, M.; Matsumoto, K.; Inoue, H.; Koto, H.; Hara, N. (1999b) PACAP reverses airway
11             hyperresponsiveness induced by ozone exposure in guinea pigs. Respiration 66: 538-542.
12       Alfaro, M. F.; Putney, L.; Tarkington, B. K.; Hatch, G. E.; Hyde, D. M.; Schelegle, E. S. (2004) Effect of rapid
13             shallow breathing on the distribution of 18O-labeled ozone reaction product in the respiratory tract of the rat.
14             Inhalation Toxicol.: in press.
15       Anderson, S. D.; Daviskas, E. (2000) The mechanism of exercise-induced asthma is ... J. Allergy Clin. Immunol.
16             106:453-459.
17       Arito, H.; Takahashi, M.; Iwasaki, T.; Uchiyama, I. (1997) Age-related changes in ventilatory and heart rate
18             responses to acute ozone exposure in the conscious rat. Ind.  Health 35:  78-86.
19       Arsalane, K.; Cosset, P.; Vanhee, D.; Voisin, C.; Hamid, Q.; Tonnel, A.-B.; Wallaert, B. (1995) Ozone stimulates
20             synthesis of inflammatory cytokines by alveolar macrophages in vitro. Am. J. Respir. Cell Mol. Biol.
21             13:60-68.
22       Avila-Costa, M. R.;  Colin-Barenque, L.;  Fortoul, T. L; Machado-Salas, J. P.; Espinosa-Villanueva, J.;
23             Rugerio-Vargas, C.; Rivas-Arancibia, S. (1999) Memory deterioration in an oxidative stress model and its
24             correlation with cytological changes on rat hippocampus CA1. Neurosci. Lett. 270: 107-109.
25       Avital, A.; Springer, C.; Bar-Yishay, E.;  Godfrey, S. (1995a) Adenosine, methacholine, and exercise challenges in
26             children with asthma or paediatric chronic obstructive pulmonary disease. Thorax 50:  511-516.
27       Avital, A.; Picard, E.; Uwyyed, K.; Springer, C. (1995b) Comparison of adenosine 5'-monophosphate and
28             methacoline for the differentiation of asthma from chronic airway diseases with the use of the auscultative
29             method in very young children. J.  Pediatr. 127: 438-440.
30       Ballinger, C. A.; Cueto, R.; Squadrito, G.; Coffin, J. F.; Velsor, L. W.; Pryor, W. A.; Postlethwait, E. M. (2005)
31             Antioxidant-mediated augmentation of ozone-induced membrane oxidation. Free Radical Biol. Med.
32             38:515-526.
33       Bassett, D.; Elbon-Copp, C.; Otterbein, S.; Barraclough-Mitchell, H.; DeLorme, M.; Yang, H. (2001) Inflammatory
34             cell availability affects ozone-induced lung damage. J. Toxicol. Environ. Health A 64: 547-565.
35       Becker, S.; Quay, J.; Koren, H.  S. (1991) Effect of ozone on immunoglobulin production by humanB cells in vitro.
36             J. Toxicol. Environ. Health 34: 353-366.
37       Beeson, W. L.; Abbey, D. E.; Knutsen, S. F. (1998) Long-term concentrations of ambient air pollutants and incident
38             lung cancer in California adults: results from the AHSMOG study. Environ. Health Perspect. 106: 813-823.
39       Bermudez, E.  (2001) Detection of poly(ADP-ribose) synthetase activity in alveolar macrophages of rats exposed to
40             nitrogen dioxide and ozone. Inhalation Toxicol. 13: 69-84.
41       Bermudez, E.; Ferng, S. F.; Castro, C. E.; Mustafa, M. G. (1999) DNA strand breaks caused by exposure to ozone
42             and nitrogen dioxide. Environ. Res. 81: 72-80.
43       Bhalla, D. K. (1996) Alteration of alveolar macrophage chemotaxis, cell adhesion, and cell adhesion molecules
44             following ozone exposure of rats.  J. Cell. Physiol. 169: 429-438.
45       Bhalla, D. K.; Gupta, S. K. (2000) Lung injury, inflammation, and inflammatory stimuli in rats exposed to ozone.
46             J. Toxicol. Environ. Health 59: 211-228.
47       Bhalla, D. K.; Gupta, S. K.; Reinhart, P. G. (1999)  Alteration of epithelial integrity, alkaline phosphatase activity,
48             and fibronectin expression in lungs of rats exposed to ozone. J. Toxicol. Environ. Health A  56: 329-343.
49       Bhalla, D. K.; Reinhart, P. G.; Bai, C.; Gupta,  S. K. (2002) Amelioration of ozone-induced lung injury by anti-tumor
50             necrosis factor-a. Toxicol. Sci. 69: 400-408.
51       Bignami, G.; Musi, B.; DeH'Omo, G.; Laviola, G.; Alleva, E. (1994) Limited effects of ozone exposure during
52             pregnancy on physical and neurobehavioral  development of CD-I mice. Toxicol. Appl. Pharmacol.
53             129:264-271.
54       Bimonte, H. A.; Nelson, M. E.;  Granholm, A. C. (2003) Age-related deficits as working memory load increases:
55             relationships with growth factors.  Neurobiol. Aging 24: 37-48.
         August 2005                                     5-78          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Bolarin, D. M; Bhalla, D. K.; Kleinman, M. T. (1997) Effects of repeated exposures of geriatric rats to ozone and
  2             particle-containing atmospheres: an analysis of bronchoalveolar lavage and plasma proteins. Inhalation
  3             Toxicol. 9: 423-434.
  4       Bonham, A. C.; Chen, C. Y.; Mutoh, T.; Joad, J. P. (2001) Lung C-fiber CNS reflex: role in the respiratory
  5             consequences of extended environmental tobacco smoke exposure ini young guinea pigs. Environ. Health
  6             Perspect. 109(suppl. 4): 573-578.
  7       Boorman, G. A.; Hailey, R.; Grumbein, S.; Chou, B. J.; Herbert, R. A.; Goehl, T.; Mellick, P. W.; Roycroft, J. H.;
  8             Haseman, J. K.; Sills, R. (1994) Toxicology and carcinogenesis studies of ozone and ozone
  9             4-(A^-nitrosomethylamino)-l-(3-pyridyl)-l-butanone in Fischer-344/N rats. Toxicol. Pathol. 22: 545-554.
10       Bornholdt, J.; Dybdahl, M.; Vogel, U.; Hansen, M.; Loft, S.; Wallin, H. (2002) Inhalation of ozone induces DNA
11             strand breaks and inflammation in mice. Mutat. Res. 520: 63-71.
12       Brannan, J. D.; Koskela, H.; Anderson, S. D.; Chew, N. (1998) Responsiveness to Mannitol in asthmatic subjects
13             with exercise- and hyperventiliation-induced asthma. Am. J. Respir. Crit. Care Med. 158: 1120-1126.
14       Bridges, J. P.; Davis, H. W.; Demodarasamy, M.; Kuroki, Y.; Howies, G.; Hui, D. Y.; McCormack, F. X. (2000)
15             Pulmonary surfactant proteins A and D are potent endogenous inhibitors of lipid peroxidation and oxidative
16             cellular injury. J. Biol.  Chem. 275: 38848-38855.
17       Broeckaert, F.; Clippe, A.; Wattiez, R.; Falmagne, P.; Bernard, A. (2003) Lung hyperpermeability, Clara-cell
18             secretory potein (CC16), and susceptibility to ozone of five inbred strains of mice. Inhalation Toxicol.
19             15: 1209-1230.
20       Bush, M. L.; Zhang, W.; Ben-Jebria, A.; Ultman, J. S. (2001) Longitudinal distribution of ozone and chlorine in the
21             human respiratory tract: simulation of nasal and oral breathing with the single-path diffusion model. Toxicol.
22             Appl. Pharmacol. 173:  137-145.
23       Calderon-Garciduenas, L.; Osorno-Velaquez, A.; Bravo-Alvarez, H.; Delgado-Chavez, R.; Barrios-Marquez, R.
24             (1992) Histopathologic changes of the nasal mucosa in Southwest metropolitan Mexico City inhabitants. Am.
25             J. Pathol. 140: 225-232.
26       Calderon-Garciduenas, L.; Rodriguez-Alcaraz, A.; Garcia, R.; Sanchez, G.; Barragan, G.; Camacho,  R.; Ramirez, L.
27             (1994) Human nasal mucosal changes after exposure to urban pollution. Environ. Health Perspect.
28             102: 1074-1080.
29       Calderon-Garciduenas, L.; Osnaya-Brizuela, N.; Ramirez-Martinez, L.; Villarreal-Calderon, A. (1996) DNA strand
30             breaks in human nasal respiratory epithelium are induced upon exposure to urban pollution. Environ. Health
31             Perspect. 104: 160-168.
32       Calderon-Garciduenas, L.; Wen-Wang, L.; Zhang, Y.-J.; Rodriguez-Alcaraz, A.; Osnaya, N.; Villarreal-Calderon,
33             A.; Santella, R. M. (1999) 8-hydroxy-2'-deoxyguanosine, a major mutagenic oxidative DNA lesion, and DNA
34             strand breaks in nasal respiratory epithelium of children exposed to urban pollution. Environ. Health Perspect.
35             107:469-474.
36       Calderon-Garciduenas, L.; Devlin, R. B.; Miller, F. J. (2000a) Respiratory tract pathology and cytokine imbalance in
37             clinically healthy children chronically and sequentially exposed to air pollutants. Med. Hypotheses
38             55: 373-378.
39       Calderon-Garciduenas, L.; Mora-Tiscareno, A.; Chung, C. J.; Valencia, G.; Fordham, L. A.; Garcia, R.;  Osnaya, N.;
40             Romero, L.; Acuna, H.; Villarreal-Calderon, A. (2000b) Exposure to air pollution is associated with lung
41             hyperinflation in healthy children and adolescents in southwest Mexico City: a pilot study. Inhalation Toxicol.
42             12:537-561.
43       Calderon-Garciduenas, L.; Rodrigues-Alcaraz, A.; Valencia-Salazar, G.; Mora-Tascareno, A.; Garcia, R.; Osnaya,
44             N.; Villarreal-Calderon, A.; Devlin, R. B.; Van Dyke, T. L. (2001) Nasal biopsies of children exposed to air
45             pollutants. Toxicol. Pathol. 29: 558-564.
46       Cassee, F. R.; Feron, V. J. (1994) Biochemical and histopathological changes in nasal epithelium of rats after 3-day
47             intermittent exposure to formaldehyde and ozone alone or in combination. Toxicol. Lett. 72: 257-268.
48       Chang, M. M.-J.; Wu, R.; Plopper, C. G.; Hyde, D. M. (1998) IL-8 is one of the major chemokines produced by
49             monkey airway epithelium after ozone-induced injury. Am. J. Physiol. 275: L524-L532.
50       Cheek, J. M.; McDonald, R. J.; Rapalyea, L.; Tarkington, B. K.; Hyde, D. M. (1995) Neutrophils enhance removal
51             of ozone-injured alveolar epithelial cells in vitro. Am. J. Physiol. 269: L527-L535.
52       Chen, L.  C.; Qu, Q.; Amdur, M. O.; Schlesinger, R. B. (1995) Alteration of pulmonary macrophage intracellular pH
53             following inhalation exposure to sulfuric acid/ozone mixtures. Exp. Lung Res. 21: 113-128.
54       Chen, L.; Yang, W.; Jennison, B. L.; Goodrich, A.; Omaye, S. T. (2002) Air pollution and birth weight in northern
55             Nevada, 1991-1999. Inhalation Toxicol. 14: 141-157.
         August 2005                                     5-79          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Chen, C.-Y.; Bonham, A. C.; Plopper, C. G.; Joad, J. P. (2003) Plasticity in respiratory motor control: selected
  2             contribution: neuroplasticity in nucleus tractus solitarius neurons following episodic ozone exposure in infant
  3             primates. J. Appl. Physiol. 94: 819-827.
  4       Cho, H. Y.; Hotchkiss, J. A.; Harkema, J. R. (1999a) Inflammatory and epithelial responses during the development
  5             of ozone-induced mucous cell metaplasia in the nasal epithelium of rats. Toxicol. Sci.  51: 135-145.
  6       Cho, H. Y.; Hotchkiss, J. A.; Bennett, C. B.; Harkema, J. R. (1999b) Effects of pre-existing rhinitis on
  7             ozone-induced mucous cell metaplasia in rat nasal epithelium. Toxicol. Appl. Pharmacol. 158: 92-102.
  8       Cho, H. Y.; Hotchkiss, J. A.; Bennett, C. B.; Harkema, J. R. (2000) Neutrophil-dependent and
  9             neutrophil-independent alterations in the nasal epithelium of ozone-exposed rats. Am.  J. Respir. Crit. Care
10             Med. 162: 629-636.
11       Cho, H.-Y.; Zhang, L.-Y.; Kleeberger, S. R. (2001) Ozone-induced lung inflammation and hyperreactivity are
12             mediated via tumor necrosis factor-a receptors. Am. J. Physiol. 280: L537-L546.
13       Churg, A.; Brauer,  M; Keeling, B. (1996) Ozone enhances the uptake of mineral particles by tracheobronchial
14             epithelial cells in organ culture. Am.  J. Respir. Crit. Care Med. 153: 1230-1233.
15       Clausen, P. A.; Wilkins, C. K.;  Wolkoff, P.; Nielsen, G. D. (2001) Chemical and biological evaluation of a reaction
16             mixture of R-(+)-limonene/ozone: formation of strong airway irritants. Environ. Int. 26: 511-522.
17       Cohen, M. D.; Zelikoff, J. T.; Qu, Q.; Schlesinger, R. B. (1996) Effects of ozone upon macrophage-interferon
18             interactions. Toxicology 114: 243-252.
19       Cohen, M. D.; Sisco, M.; Li, Y.; Zelikoff, J. T.; Schlesinger, R.  B. (2001) Ozone-induced modulation of
20             cell-mediated immune responses in the lungs.  Toxicol. Appl. Pharmacol. 171: 71-84.
21       Cohen, M. D.; Sisco, M.; Baker, K.; Li, Y.;  Lawrence, D.; VanLoveren, H.; Zelikoff, J. T.; Schlesinger, R. B.
22             (2002) Effects of inhaled ozone on pulmonary immune cells critical to antibacterial responses in situ.
23             Inhalation Toxicol. 14: 599-619.
24       Colin-Barenque, L.; Avila-Costa, M. R.; Fortoul, T.; Rugerio-Vargas, C.; Machado-Salas, J. P.;
25             Espinosa-Villanueva, J.; Rivas-Arancibia, S. (1999) Morphologic alteration of the olfactory bulb after acute
26             ozone exposure in rats. Neurosci. Lett. 274: 1-4.
27       Connor, L. M.; Ballinger, C. A.; Albrecht, T. B.; Postlethwait, E. M. (2004) Interfacial phospholipids inhibit ozone
28             reactive absorption-mediated cytotoxicity in vitro.  Am. J. Physiol.: 10.1152/ajplung.00397.2003.
29       Cottet-Emard, J.-M.; Dalmaz, Y.; Pequignot, J.; Peyrin, L.; Pequignot, J.-M. (1997) Long-term exposure to ozone
30             alters peripheral and central catecholamine activity in rats. Pfluegers Arch. 433: 744-749.
31       Creutzenberg, O.; Bellmann, B.; Klingebiel, R.; Heinrich, U.; Muhle, H. (1995) Phagocytosis and chemotaxis of rat
32             alveolar macrophages after a combined or separate exposure to ozone and carbon black. Exp. Toxicol. Pathol.
33             47:202-206.
34       Cross, C. E.; Van Der Vliet, A.; Louie, S.; Thiele, J. J.; Halliwell, B. (1998) Oxidative stress  and antioxidants at
35             biosurfaces:  plants, skin, and respiratory tract surfaces. Environ. HealthPerspect. 106(suppl. 5): 1241-1251.
36       Custodio-Ramirez,  V.; Paz,  C. (1997) Ozone produces functional deficits in the rat visual pathway.
37             Electroencephalogr. Clin. Neurophysiol. 104:  269-273.
3 8       Daly, C.; Fox, K.; Henein, M. (2002) Natriuretic peptides in the diagnosis of heart disease-first amongst equals? Int.
39             J. Cardiol. 84: 107-113.
40       Delaunois, A.; Segura, P.; Montano, L. M.; Vargas, M. H.; Ansay, M.; Gustin, P. (1998) Comparison of
41             ozone-induced effects on lung mechanics and hemodynamics in the rabbit. Toxicol.  Appl. Pharmacol.
42             150:58-67.
43       DeH'Omo, G.; Fiore, M.; Petruzzi, S.; Alleva, E.; Bignami, G. (1995a) Neurobehavioral development of CD-I mice
44             after combined gestational and postnatal exposure  to ozone. Arch. Toxicol. 69: 608-616.
45       DeH'Omo, G.; Wolfer, D.; Alleva, E.; Lipp,  H.-P. (1995b) Developmental exposure to ozone  induces subtle changes
46             in swimming navigation of adult mice. Toxicol. Lett. 81:91-99.
47       DeLorme, M. P.; Yang, H.; Elbon-Copp, C.; Gao, X.; Barraclough-Mitchell, H.; Bassett, D. J. P. (2002)
48             Hyperresponsive airways correlate with lung tissue inflammatory cell changes in ozone-exposed rats. J.
49             Toxicol. Environ. Health Part A 65: 1453-1470.
50       DeMarini, D. M.; Shelton, M. L.; Kohan, M. J.; Hudgens, E. E.; Kleindienst, T. E.; Ball, L. M.; Walsh, D.; de Boer,
51             J. G.; Lewis-Bevan, L.; Rabinowitz, J. R.; Claxton, L. D.; Lewtas, J. (2000) Mutagenicity in lung of Big
52             Blue(R) mice and induction of tandem-base substitutions in Salmonella by the air pollutant peroxyacetyl
53             nitrate (PAN): predicted formation of intrastrand cross-links. Mutat. Res. 457: 41-55.
54       Depuydt, P.; Joos, G. F.; Pauwels, R. A. (1999) Ambient ozone  concentrations induce airway hyperresponsiveness
55             in some rat strains. Eur. Respir. J. 14: 125-131.
         August 2005                                    5-80         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Depuydt, P. O.; Lambrecht, B. N.; Joos, G. F.; Pauwels, R. A. (2002) Effect of ozone exposure on allergic
  2             sensitization and airway inflammation induced by dendritic cells. Clin. Exp. Allergy 32: 391-396.
  3       Dorado-Martinez, C.; Parades-Carbajal, C.; Mascher, D.; Borgonio-Perez, G.; Rivas-Arancibia, S. (2001) Effects of
  4             different ozone doses on memory, motor activity and lipid peroxidation levels, in rats. Int. J. Neurosci.
  5             108: 149-161.
  6       Dormans, J. A. M. A.; Boere, A. J. F.; van Loveren, H.; Rombout, P. J. A.; Marra, M; van Bree, L.  (1996)
  7             Age-related toxicity in rat lungs following acute and repeated ozone exposure. Inhalation Toxicol. 8: 903-925.
  8       Dormans, J. A. M. A.; Van Bree, L.; Boere, A. J. F.; Marra, M.; Rombout, P. J. A. (1999) Interspecies differences in
  9             time course of pulmonary toxicity following repeated exposure to ozone. Inhalation Toxicol.  11: 309-329.
10       Driscoll, K. E.;  Simpson, L.; Carter, J.; Hassenbein, D.; Leikauf, G. D. (1993) Ozone inhalation stimulates
11             expression of a neutrophil chemotactic protein, macrophage inflammatory protein 2. Toxicol. Appl.
12             Pharmacol. 119: 306-309.
13       Dye, J. A.; Madden, M. C.; Richards, J. H.; Lehmann, J. R.; Devlin, R. B.;  Costa, D. L. (1999) Ozone effects on
14             airway responsiveness, lung injury, and inflammation. Comparative rat strain and in vivo/in vitro
15             investigations. Inhalation Toxicol. 11: 1015-1040.
16       El-Fawal, H. A. N.;  McGovern, T.; Schlesinger, R. B. (1995) Nonspecific bronchial responsiveness assessed in vitro
17             following acute inhalation exposure to ozone and ozone/sulfuric acid mixtures. Exp. Lung Res. 21: 129-139.
18       Elder, A. C. P.;  Gelein, R.; Finkelstein J. N.; Cox, C.; Oberdorster, G. (2000a) Endotoxin priming affects the lung
19             response to ultrafine particles and ozone in young and old rats. In: Phalen, R. F., ed. Inhalation toxicology:
20             proceedings of the third colloquium on paniculate air pollution and human health (first special issue); June,
21             1999; Durham, NC. Inhalation Toxicol. 12(suppl. 1): 85-98.
22       Elder, A. C. P.;  Gelein, R.; Finkelstein, J. N.; Cox, C.; Oberdorster, G. (2000b) Pulmonary inflammatory response to
23             inhaled ultrafine particles is modified by age, ozone exposure, and bacterial toxin. In: Grant, L. D., ed.
24             PM2000: paniculate matter and health. Inhalation Toxicol. 12(suppl. 4): 227-246.
25       Elsayed, N. M. (2001) Diet restriction modulates lung response and survivability of rats exposed to ozone.
26             Toxicology 159:  171-182.
27       Evans, M. J.; Fanucchi, M. V.; Baker, G. L.; Van Winkle, L. S.; Pantle, L. M.; Nishio, S. J.; Schelegle, E. S.;
28             Gershwhin, L. J.; Miller, L. A.; Hyde, D. M.; Sannes, P. L.; Plopper, C. G. (2003) Atypical development of
29             the trachea! basement membrane zone of infant rhesus monkeys exposed to ozone and allergen. Am. J.
30             Physiol. 285:L931-L939.
31       Fakhrzadeh, L.; Laskin, J. D.; Laskin,  D. L. (2002) Deficiency in inducible nitric oxide synthase protects mice from
32             ozone-induced lung inflammation and tissue injury. Am. J. Respir. Cell Mol. Biol. 26: 413-419.
33       Fanucchi, M. V.; Hotchkiss, J. A.; Harkema, J. R. (1998) Endotoxin potentiates ozone-induced mucous cell
34             metaplasia in rat nasal epithelium. Toxicol. Appl. Pharmacol. 152: 1-9.
35       Farman, C. A.; Pinkerton, K. E.; Rajini, P.; Witschi, H.; Last, J.  A. (1997) Evolution of lung lesions in rats exposed
36             to mixtures of ozone and nitrogen dioxide. Inhalation Toxicol. 9: 647-677'.
37       Farman, C. A.; Watkins, K.; Van Hoozen, B.; Last, J. A.; Witschi, H.; Pinkerton, K. E.  (1999) Centriacinar
3 8             remodeling and sustained procollagen gene expression after exposure to ozone and nitrogen dioxide. Am. J.
39             Respir. Cell Mol. Biol. 20: 303-311.
40       Ferng, S.-F.; Castro, C. E.; Afifi, A. A.; Bermudez, E.; Mustafa, M. G. (1997) Ozone-induced DNA strand breaks in
41             guinea pig tracheobronchial epithelial cells. J. Toxicol. Environ. Health 51: 353-367.
42       Folkerts, G.; Busse,  W. W.; Nijkamp,  F. P.; Sorkness, R.; Gern, J. E. (1998) Virus-induced airway
43             hyperresponsiveness and asthma. Am. J. Respir. Crit. Care Med. 157: 1708-1720.
44       Foster, W. M.; Freed, A. N. (1999) Regional clearance of solute from peripheral airway epithelia: recovery after
45             sublobar exposure to ozone. J. Appl. Physiol. 86: 641-646.
46       Frampton, M. W.; Pryor, W. A.; Cueto, R.; Cox, C.; Morrow, P. E.; Utell, M. J. (1999) Aldehydes (nonanal and
47             hexanal)  in rat and human bronchoalveolar lavage fluid after ozone exposure. Cambridge, MA: Health Effects
48             Institute; research report no. 90. Available: www.healtheffects.org/Pubs/Frampton-C.pdf [2000, February 9].
49       Freed, A. N.; Chou,  C. L.; Fuller, S. D.;  Croxton, T. L. (1996) Ozone-induced vagal reflex modulates airways
50             reactivity in rabbits. Respir. Physiol. 105: 95-102.
51       Freed, A. N.; Cueto, R.; Pryor, W. A. (1999) Antioxidant transport modulates peripheral airway reactivity and
52             inflammation during ozone exposure. J. Appl. Physiol. 87: 1595-1603.
53       Friedlander,  S. K.; Yeh, E. K. (1998) The submicron atmospheric aerosol as a carrier of reactive chemical species:
54             case of peroxides. Appl. Occup. Environ. Hyg. 13: 416-420.
55       Gaffney, J. S.; Marley, N. A.; Prestbo, E. W. (1993) Measurements of peroxyacetyl nitrate at a remote site in the
56             southwestern United States: tropospheric implications. Environ. Sci. Technol. 27: 1905-1910.
         August 2005                                    5-81         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Garssen, I; Van Bree, L.; Van Der Vliet, H.; Van Loveren, H. (1997) Ozone-induced impairment of pulmonary type
  2             IV hypersensitivity and airway hyperresponsiveness in mice. Inhalation Toxicol. 9: 581-599.
  3       Gohil, K.; Cross, C. E.; Last, J. A. (2003) Ozone-induced disruptions of lung transcriptomes. Biochem. Biophys.
  4             Res.  Commun. 305: 719-728.
  5       Goldsmith,  C.-A. W.; Ning, Y.-Y.; Qin, G.; Imrich, A.; Lawrence, J.; Murthy, G. G., K.;  Catalano, P. J.; Kobzik, L.
  6             (2002) Combined air pollution particle and ozone exposure increases airway responsiveness in mice.
  7             Inhalation Toxicol. 14: 325-347.
  8       Gonzalez-Pina, R.; Paz, C. (1997) Brain monoamine changes in rats after short periods of ozone exposure.
  9             Neurochem. Res. 22: 63-66.
10       Grosjean, D. (2003) Ambient PAN and PPN in southern California from 1960 to the SCOS97-NARSTO. Atmos.
11             Environ. 37(suppl. 2): S221-S238.
12       Grosjean, E.; Grosjean, D.; Woodhouse, L. F. (2001) Peroxyacetyl nitrate and peroxypropionyl nitrate during SCOS
13             97-NARSTO. Environ. Sci. Technol. 35: 4007-4014.
14       Guerrero, A. L.; Dorado-Martinez, C.; Rodriguez, A.; Pedroza-Rios, K.; Borgonio-Perez, G.; Rivas-Arancibia, S.
15             (1999) Effects of vitamin E on ozone-induced memory deficits and lipid peroxidation in rats. NeuroReport
16             10: 1689-1692.
17       Gulisano, M; Marceddu, S.; Barbara, A.; Pacini, A.; Buiatti, E.; Martini, A.; Pacini, P. (1997) Damage to the
18             nasopharyngeal mucosa induced by current levels of urban air pollution: a field study in lambs. Eur. Respir. J.
19             10:567-572.
20       Gtinther, T.; Hollriegl, V.; Vormann, J. (1993) Perinatal development of iron and antioxidant defence systems.
21             J. Trace Elem. Electrolytes Health Dis. 7: 47-52.
22       Gupta, S. K.; Reinhart, P. G.; Bhalla, D. K. (1998) Enhancement of fibronectin expression in rat lung by ozone and
23             an inflammatory stimulus. Am. J. Physiol. 275: L330-L335.
24       Haddad, E.-B.; Liu, S. F.; Salmon, M.; Robichaud, A.; Barnes, P. J.; Chung, K. F.  (1995) Expression of inducible
25             nitric oxide synthase mRNA in Brown Norway rats exposed to ozone: effect of dexamethasone. Eur. J.
26             Pharmacol. Environ. Toxicol. Pharmacol. Sect. 293: 287-290.
27       Hamilton, R. F.; Li, L.; Eschenbacher, W. L.; Szweda, L.; Holian, A. (1998) Potential involvement of
28             4-hydroxynonenal in the response of human lung cells to ozone. Am.  J. Physiol. 274: L8-L16.
29       Harkema, J. R.; Morgan, K. T.; Gross, E. A.; Catalano, P.  J.; Griffith, W. C.  (1994) Consequences of prolonged
30             inhalation of ozone on F344/N rats: collaborative studies. Part VII: effects on the nasal mucociliary apparatus.
31             Cambridge, MA: Health Effects Institute; research report no. 65.
32       Harkema, J. R.; Catalano, P. J.; Hotchkiss, J. A. (1997a) Consequences of prolonged inhalation of ozone on F344/N
33             rats:  collaborative studies. Part XII. Atrophy of bone in nasal turbinates. Cambridge, MA: Health Effects
34             Institute; research report no. 65.
35       Harkema, J. R.; Hotchkiss, J. A.; Griffith, W. C. (1997b) Mucous cell metaplasia in rat nasal epithelium after a
36             20-month exposure to ozone: a morphometric study of epithelial differentiation. Am. J. Respir. Cell Mol.
37             Biol. 16: 521-530.
38       Harkema, J. R.; Hotchkiss, J. A.; Barr, E. B.; Bennett, C. B.; Gallup, M.; Lee, J. K.; Basbaum, C. (1999)
39             Long-lasting effects of chronic ozone exposure on rat nasal epithelium. Am. J. Respir. Cell Mol. Biol.
40             20:517-529.
41       Kara, R.; Paz, C. (1993) Effects of ozone exposure during pregnancy on ontogeny of sleep in rats. Neurosci. Lett.
42             164:67-70.
43       Hawgood, S.; Poulain, F. R. (2001) The pulmonary collectins and surfactant metabolism. Annu. Rev. Physiol.
44             63:495-519.
45       Hawgood, S.; Ochs, M.; Jung, A.; Akiyama, J.; Allen, L.; Brown, C.; Edmondson, J.; Levitt, S.; Carlson, E.;
46             Gillespie, A. M.; Villar, A.; Epstein, C. J.; Poulain, F. R.  (2002) Sequential targeted deficiency of SP-A and
47             -D leads to progressive alveolar lipoproteinosis and emphysema. Am. J. Physiol. 283: L1002-L1010.
48       Heck, H. d'A.; Casanova, M.; Starr, T. B. (1990) Formaldehyde toxicity-new understanding. Crit. Rev. Toxicol.
49             20:397-426.
50       Heddle, J. A.; Shepson, P. B.; Gingerich, J. D.; So, K. W.  (1993) Mutagenicity of peroxyacetyl nitrate (PAN)
51             in vivo: tests for somatic mutations and chromosomal aberrations. Environ.  Mol. Mutagen. 21: 58-66.
52       Herbert, R.  A.; Hailey, J. R.; Grumbein, S.; Chou, B. J.; Sills, R. C.; Haseman, J. K.; Goehl, T.; Miller, R. A.;
53             Roycroft, J. H.; Boorman, G. A. (1996) Two-year and lifetime toxicity and carcinogenicity studies of ozone in
54             B6C3F1 mice. Toxicol. Pathol. 24: 539-548.
55       Highfill, J. W.; Watkinson, W. P. (1996) Ozone toxicity in the rat. II. Modeling changes due to ambient temperatures
56             and duration. J. Appl. Physiol.  80: 1811-1818.
         August 2005                                    5-82         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Hoffer, E.; Baum, Y.; Tabak, A.; Frevert, C. (1999) Adhesion molecules of blood polymorphonuclear leukocytes
  2             and alveolar macrophages in rats: modulation by exposure to ozone. Hum. Exp. Toxicol. 18: 547-551.
  3       Hotchkiss, J. A.; Harkema, J. R.; Johnson, N. F. (1997) Kinetics of nasal epithelial cell loss and proliferation in F344
  4             rats following a single exposure to 0.5 ppm ozone. Toxicol. Appl. Pharmacol. 143: 75-82.
  5       Hotchkiss, J. A.; Hilaski, R.; Cho, H.; Regan, K.; Spencer, P.; Slack, K.; Harkema, J. R. (1998) Fluticasone
  6             propionate attenuates ozone-induced rhinitis and mucous cell metaplasia in rat nasal airway epithelium.
  7             Am. J. Respir. Cell Mol. Biol. 18: 91-99.
  8       Huffman, L. J.; Judy, D. J.; Brumbaugh, K.; Frazer, D. G.; Reynolds, J. S.; McKinney, W. G.; Goldsmith, W. T.
  9             (2001) Hyperthyroidism increases the risk of ozone-induced lung toxicity in rats. Toxicol. Appl. Pharmacol.
10             173: 18-26.
11       Huitron-Resendiz, S.; Custodio-Ramirez, V.; Escalante-Membrillo, C.; Gonzalez-Pina, R.; Paz, C. (1994) Sleep
12             alterations and brain regional changes of serotonin and its metabolite in rats exposed to ozone. Neurosci. Lett.
13             177: 119-122.
14       Hyde, D. M; Miller, L. A.; McDonald, R. J.; Stovall, M. Y.; Wong, V.; Pinkerton, K. E.; Wegner, C. D.; Rothlein,
15             R.; Plopper, C. G. (1999) Neutrophils enhance clearance of necrotic epithelial cells in ozone-induced lung
16             injury in rhesus monkeys. Am. J. Physiol. 277: L1190-L1198.
17       Igarashi, A.; lijima, H.; Tamura,  G.; Shirato, K. (1998) Tazanolast inhibits ozone-induced airway
18             hyperresponsiveness in guinea pigs. Am. J. Respir. Crit. Care  Med. 157: 1531-1535.
19       lijima, M. K.; Kobayashi, T.; Kamada, H.; Shimojo, N. (2001) Exposure to ozone aggravates nasal allergy-like
20             symptoms in guinea pigs.  Toxicol. Lett.  123: 77-85.
21       Ishii, Y.; Yang, H.; Sakamoto, T.; Nomura, A.; Hasegawa, S.; Hirata, F.; Bassett, D. J. P. (1997) Rat alveolar
22             macrophage cytokine production and regulation of neutrophil  recruitment following acute ozone exposure.
23             Toxicol. Appl. Pharmacol. 147: 214-223.
24       Ishii, Y.; Hashimoto, K.; Hirano, K.; Morishima, Y.; Mochizuki, M.; Masuyama, K.; Nomura, A.; Sakamoto, T.;
25             Uchida, Y.; Sagai, M.; Sekizawa, K. (2000) Ebselen decreases ozone-induced pulmonary inflammation in
26             rats. Lung 178: 225-234.
27       Iwasaki, T.; Takahashi, M.; Saito, H.; Arito, H. (1998) Adaptation of extrapulmonary  responses to ozone exposure in
28             conscious rats. Ind. Health 36: 57-60.
29       Jakobi, G.; Fabian, P. (1997) Indoor/outdoor concentrations of ozone and peroxyacetyl nitrate (PAN). Int. J.
30             Biometeorol. 40: 162-165.
31       Jang, A.-S.; Choi, I.-S.; Koh, Y.-L; Park, C.-S.; Lee, J.-S. (2002) The relationship between alveolar epithelial
32             proliferation and airway obstruction after ozone exposure. Allergy 57: 737-740.
33       Jimba, M.; Skornik, W. A.; Killingsworth, C. R.; Long, N. C.; Brain, J. D.; Shore, S. A. (1995) Role of C fibers in
34             physiological responses to ozone in rats. J. Appl. Physiol. 78:  1757-1763.
35       Joad, J. P.; Kott, K. S.; Bonham, A. C. (1998) Exposing guinea pigs to ozone for 1 wk enhances responsiveness of
36             rapidly adapting receptors. J. Appl. Physiol. 84: 1190-1197.
37       Joad, J. P.; Brie, J. M.; Weir, A. J.; Putney, L.; Hyde, D. M.; Postlewait, E. M.; Plopper, C. G. (2000) Effect of
38             respiratory pattern on ozone injury to the airways of isolated rat lungs. Toxicol. Appl. Pharmacol.  169: 26-32.
39       Johnston, C. J.; Stripp, B. R.; Piedbeouf, B.; Wright, T. W.; Mango, G. W.; Reed, C. K.; Finkelstein, J. N. (1998)
40             Inflammatory and epithelial responses in mouse strains that differ in sensitivity to hyperoxic injury. Exp.
41             Lung Res. 24: 189-202.
42       Johnston, C. J.; Stripp, B. R.; Reynolds, S. D.; Avissar, N. E.; Reed,  C. K.; Finkelstein, J. N. (1999a) Inflammatory
43             and antioxidant gene expression in C57BL/6J mice after lethal and sublethal ozone exposures. Exp. Lung
44             Res. 25: 81-97.
45       Johnston, C. J.; Finkelstein, J. N.; Oberdorster,  G.; Reynolds, S. D.; Stripp, B. R. (1999b) Clara cell  secretory
46             protein-deficient mice differ from wild-type mice in inflammatory chemokine expression to oxygen and
47             ozone, but not to endotoxin. Exp. Lung Res. 25: 7-21.
48       Johnston, C. J.; Oberdorster, G.;  Gelein, R.; Finkelstein, J. N. (2000) Newborn mice differ from adult mice in
49             chemokine and cytokine expression to ozone, but not to endotoxin. Inhalation Toxicol. 12: 205-224.
50       Johnston, C. J.; Oberdorster, G.;  Gelein, R.; Finkelstein, J. N. (2002) Endotoxin potentiates ozone-induced
51             pulmonary chemokine and inflammatory responses. Exp. Lung Res. 28: 419-433.
52       Kafoury, R. M.; Pryor, W. A.; Squadrito, G. L.; Salgo, M.  G.; Zou, X.; Friedman, M.  (1999) Induction of
53             inflammatory mediators in human airway epithelial cells by lipid ozonation products. Am. J. Respir. Crit.
54             Care Med. 160:  1934-1942.
55       Kenyon, N. J.; Van Der Vliet, A.; Schock, B.C.; Okamoto, T.; McGrew, G. M.; Last, J. A. (2002) Susceptibility to
56             ozone-induced acute lung injury in iNOS-deficient mice. Am. J.  Physiol. 282: L540-L545.
         August 2005                                    5-83          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Kim, M. Y.; Son, J. W.; Cho, M. H.; Choi, C. S.; Chae, C. H.; Lee, M. H. (2001) Oviductural carcinoma in B6C3F1
  2             female mice exposed to 0.5 ppm ozone. Vet. Hum. Toxicol. 43: 370-372.
  3       Kimmel, T. A.; Chen, L. C.; Bosland, M. C.; Nadziejko, C. (1997) Influence of acid aerosol droplet size on
  4             structural changes in the rat lung caused by acute exposure to sulfuric acid and ozone. Toxicol. Appl.
  5             Pharmacol. 144: 348-355.
  6       Kleeberger, S. R.; Levitt, R. C.; Zhang, L.-Y.; Longphre, M.; Harkema, J.; Jedlicka, A.; Eleff, S. M.; DiSilvestre, D.;
  7             Holroyd, K. J. (1997) Linkage analysis of susceptibility to ozone-induced lung inflammation in inbred mice.
  8             Nat. Genet. 17: 475-478.
  9       Kleeberger, S. R.; Reddy, S.;  Zhang, L.-Y.; Jedlicka, A. E. (2000) Genetic susceptibility to ozone-induced lung
10             hyperpermeability: role of toll-like receptor 4. Am. J. Respir. Cell Mol. Biol. 22: 620-627.
11       Kleeberger, S. R.; Reddy, S. P.; Zhang, L.-Y.; Cho, H.-Y.; Jedlicka, A. E. (2001a) Toll-like receptor 4 mediates
12             ozone-induced murine  lung hyperpermeability via inducible nitric oxide synthase. Am. J. Physiol.
13             280: L326-L333.
14       Kleeberger, S. R.; Ohtsuka, Y.; Ahang, L.-Y.; Longphre, M. (200Ib) Airway responses to chronic ozone exposure
15             are partially mediated through mast cells. J. Appl. Physiol. 90: 713-723.
16       Kleindienst,  T. E.; Shepson, P. B.; Smith, D. F.; Hudgens, E. E.; Nero, C. M.; Cupitt, L. T.; Bufalini, J. J.; Claxton,
17             L. D.  (1990) Comparison of mutagenic activities of several peroxyacyl nitrates. Environ. Mol. Mutagen.
18             16:70-80.
19       Kleinman, M. T.; Mautz, W. J.; Bjarnason, S. (1999) Adaptive and non-adaptive responses in rats exposed to ozone,
20             alone and in mixtures,  with acidic aerosols. Inhalation Toxicol. 11: 249-264.
21       Kleinman, M. T.; Bufalino, C.; Rasmussen, R.; Hyde, D.; Bhalla, D. K.; Mautz, W. J. (2000) Toxicity of chemical
22             components of ambient fine paniculate matter (PM2 5) inhaled by aged rats.  J. Appl. Toxicol. 20: 357-364.
23       Kligerman, A. D.; Mottus, K.; Erexson, G.  L. (1995) Cytogenetic analyses of the in vitro and in vivo responses of
24             murine cells to peroxyacetyl nitrate (PAN). Mutat. Res. 341: 199-206.
25       Kobzik, L.; Goldsmith, C.-A.  W.; Ning, Y. Y.; Qin, G.; Morgan, B.; Imrich, A.; Lawrence J.; Murthy, G. G. K.;
26             Catalano, P. J. (2001) Effects of combined ozone and air pollution particle exposure in mice. Boston, MA:
27             Health Effects Institute; research report no. 106.  Available: http://www.healtheffects.org/Pubs/Kobzik.pdf
28             [27 January 2003].
29       Kodavanti, U. P.; Hatch, G. E.; Starcher, B.; Giri, S. N.; Winsett, D.; Costa, D. L. (1995) Ozone-induced pulmonary
3 0             functional, pathological, and biochemical changes in normal and vitamin C-deficient guinea pigs. Fundam.
31             Appl. Toxicol. 24:  154-164.
32       Koto, H.; Salmon, M.; Haddad el-B.; Huang, T.-J.; Zagorski, J.; Chung, K. F. (1997) Role of cytokine-induced
33             neutrophil chemoattractant (CINC) in ozone-induced airway inflammation and hyperresponsiveness. Am. J.
34             Respir. Crit. Care Med. 156: 234-239.
3 5       Koyama, Y.; Hayaishi, O. (1994) Modulation by prostaglandins of activity of sleep-related neurons in the
36             preoptic/anterior hypothalamic areas in rats. Brain Res. Bull. 33: 367-372.
37       Krishna, M. T.; Chauhan, A. J.; Frew, A. J.; Holgate, S. T. (1998) Toxicological mechanisms underlying oxidant
38             pollutant-induced airway injury. Rev. Environ. Health 13: 59-71.
39       Larson, S. D.; Schelegle, E. S.; Walby, W.  F.; Gershwin, L. J.; Fanuccihi, M. V.; Evans, M. J.; Joad, J. P.;
40             Tarkington, B. K.; Hyde, D. M.; Plopper, C. G. (2004) Postnatal remodeling of the neural components of the
41             epithelial-mesenchymal trophic unit in the proximal airways of infant rhesus monkeys exposed to ozone and
42             allergen. Toxicol. Appl. Pharmacol.  194: 211-220.
43       Laskin, D. L.; Laskin, J. D. (2001) Role of macrophages and inflammatory mediators in chemically induced toxicity.
44             Toxicology (Ireland) 160: 111-118.
45       Laskin, D. L.; Pendino, K. J.;  Punjabi, C. J.; del Valle, M. R.; Laskin, J. D. (1994) Pulmonary and hepatic effects of
46             inhaled ozone in rats. Environ. Health Perspect. 102(suppl. 10): 61-64.
47       Laskin, J. D.; Heck, D. E.; Laskin, D. L. (1996) Nitric oxide production in the lung and liver following inhalation of
48             the pulmonary irritant ozone. In: Snyder, R.; Kocsis, J.  J.; Sipes, I. G.; Kalf, G. F.;  Jollow, D. J.; Greim, H.;
49             Monks, T. J.;  Witmer,  C. M., eds.  Biological Reactive Intermediates V: Basic Mechanistic Research in
50             Toxicology and Human Risk Assessment: proceedings of the Fifth International Symposium; January 1995;
51             Munich, Germany. Adv. Exp. Med. Biol. 387: 141-146.
52       Laskin, D. L.; Sunil, V.; Guo, Y.; Heck, D. E.; Laskin, J. D. (1998) Increased nitric oxide synthase in the lung after
53             ozone inhalation is associated with activation of NF-KB. Environ. Health Perspect. 106(suppl. 5): 1175-1178.
54       Laskin, D. L.; Fakhrzadeh, L.; Heck, D. E.; Gerecke, D.; Laskin, J. D. (2002) Upregulation of phosphoinositide
55             3-kinase and protein kinase B in alveolar macrophages following ozone inhalation. Role of NF-KB and
56             STAT-1 in ozone-induced nitric oxide production and toxicity. Mol. Cell. Biochem. 234-235: 91-98.
         August 2005                                     5-84          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Last, J. A.; Pinkerton, K. E. (1997) Chronic exposure of rats to ozone and sulfuric acid aerosol: biochemical and
  2             structural responses. Toxicology 116: 133-146.
  3       Lavnikova, N.; Prokhorova, S.; Lakhotia, A. V.; Gordon, R.; Laskin, D. L. (1998) Distinct inflammatory responses
  4             of adherent vascular lung neutrophils to pulmonary irritants. J. Inflammation 48: 56-66
  5       Lee, C.; Watt, K. C.; Chang, A. M; Plopper, C. G.; Buckpitt, A. R.; Pinkerton, K. E. (1998) Site-selective
  6             differences in cytochrome P450 isoform activities: comparison of expression in rat and rhesus monkey lung
  7             and induction in rats. Drug Metab. Dispos. 26: 396-400.
  8       Lemos, M.; Lichtenfels, A. J. F. C.; Amaro, E., Jr.; Macchione, M.; Martins, M. A.; King, M.; Bohm, G. M.;
  9             Saldiva, P. H. N. (1994) Quantitative pathology of nasal passages in rats exposed to urban levels of air
10             pollution. Environ. Res. 66:  87-95.
11       Longphre, M.; Zhang, L.-Y.; Harkema, J. R.; Kleeberger, S. R. (1999) Ozone-induced pulmonary inflammation and
12             epithelial proliferation are partially mediated by PAF. J. Appl. Physiol. 86: 341-349.
13       Lorz, C.; Lopez, J. (1997) Incidence of air pollution in the pulmonary surfactant system of the pigeon (Columbia
14             livid). Anat. Rec. 249:  206-212.
15       Madden, M.  C.; Richards, J. H.; Dailey, L. A.; Hatch, G. E.; Ohio, A. J. (2000) Effect of ozone ondiesel exhaust
16             particle toxicity in rat lung. Toxicol. Appl. Pharmacol. 168: 140-148.
17       Mango,  G. W.; Johnston, C. J.; Reynolds, S. D.; Finkelstein, J. N.; Plopper, C. G.; Stripp, B. R. (1998) Clara cell
18             secretory protein deficiency  increases oxidant stress response in conducting airways. Am. J. Physiol.
19             275: L348-L356.
20       Matsubara, S.; Kikkawa, H.; Kaminuma, O.; Ikezawa, K. (1997a) Angiotensin-converting enzyme inhibitors can
21             potentiate ozone-induced airway hyperresponsiveness. Eur. J. Pharmacol. 337: 259-265.
22       Matsubara, S.; Fushimi, K.; Kaminuma, O.; Kikkawa, H.; Ikezawa, K.; Naito, K. (1997b) Prevention of
23             ozone-induced airway  hyperresponsiveness and epithelial injury by phosphodiesterase inhibitors in guinea
24             pigs. Environ. Toxicol. Pharmacol. 3: 201-209.
25       Mautz, W. J. (2003) Exercising animal models in inhalation toxicology: interactions with ozone and formaldehyde.
26             Environ. Res. 92:  14-26.
27       Mautz, W. J.; Kleinman, M. T.; Bhalla, D. K.; Phalen, R. F. (2001) Respiratory tract  responses to repeated inhalation
28             of an oxidant and acid gas-particle air pollutant mixture. Toxicol. Sci. 61: 331-341.
29       McGraw, D. W.; Forbes, S. L.; Mak, J. C. W.; Witte, D. P.; Carrigan, P. E.; Leikauf, G. D.; Liggett, S. B. (2000)
30             Transgenic overexpression of beta(2)-adrenergic receptors in airway epithelial cells decreases
31             bronchoconstriction. Am. J.  Physiol. 279: L379-L389.
32       McKinney, W. J.; Jaskot, R. H.; Richards, J. H.; Costa,  D. L.; Dreher, K. L. (1998) Cytokine mediation of
33             ozone-induced pulmonary adaptation. Am. J. Respir. Cell. Mol. Biol. 18: 696-705.
34       Miller, L. A.; Barnett, N. L.; Sheppard, D.; Hyde, D. M. (2001) Expression of the P6 integrin subunit is associated
35             with sites of neutrophil influx in lung epithelium. J. Histochem. Cytochem. 49: 41-48.
36       Morgan, M. S.; Meyer, P.; Holub, R.; Frank, R. (1986)  Overall and regional lung function in dogs exposed acutely to
37             ozone. Environ. Res. 41: 546-557.
38       Morio, L. A.; Hooper, K. A.;  Brittingham, J.; Li, T.-H.; Gordon, R. E.; Turpin, B. J.;  Laskin, D. L. (2001) Tissue
3 9             injury following inhalation of fine paniculate matter and hydrogen perioxide is associated with altered
40             production of inflammatory  mediators and antioxidants by alveolar macrophages. Toxicol. Appl. Pharmacol.
41             177: 188-199.
42       Moss, O. R.; Gross, E. A.; James, R. A.; Janszen, D. B.; Ross, P. W.; Roberts, K. C.; Howard, A. M.; Harkema,
43             J. R.; Calderon-Garciduenas, L.; Morgan, K. T. (2001) Respiratory tract toxicity in rats exposed to Mexico
44             City air. Cambridge, MA:  Health Effects Institute; research report no. 100. Available:
45             http://www.healtheffects.org/pubs-research.htm [15 May, 2003].
46       Mticke, W. (1996) The environment and the eye. Topics of ophthalmic toxicology. Leban. Med. J. 44:  146-150.
47       Mudway, I. S.; Kelly, F. J. (1998) Modeling the interactions of ozone with pulmonary epithelial lining fluid
48             antioxidants. Toxicol. Appl.  Pharmacol. 148: 91-100.
49       Mudway, I. S.; Kelly, F. J. (2000) Ozone and the lung: a sensitive issue. Mol.  Aspects. Med. 21: 1-48.
50       Murphy, D. J. (2002) Assessment of respiratory function in safety pharmacology. Fundam. Clin. Pharmacol.
51             16:183-196.
52       Mutoh, T.; Joad, J. P.; Bonham, A.  C. (2000) Chronic passive  cigarette smoke exposure augments
53             bronchopulmonary C-fibre inputs to nucleus tractus solitarii neurones and reflex output in young guinea-pigs.
54             J. Physiol. (London) 523: 223-233.
         August 2005                                     5-85          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Nakano, H.; Aizawa, H.; Matsumoto, K.; Fukuyama, S.; Inoue, H.; Hara, N. (2000) Cyclooxygenase-2 participates
  2             in the late phase of airway hyperresponsiveness after ozone exposure in guinea pigs. Eur. J. Pharmacol.
  3             403:267-275.
  4       National Toxicology Program. (1994) NTP technical report on the toxicology and carcinogenesis studies of ozone
  5             (CAS no. 10028-15-6) and ozone/NNK (CAS no. 10028-15-6/64091-91-4) in F344/N rats and B6C3F! mice
  6             (inhalation studies). Research Triangle Park, NC: U.S. Department of Health and Human Services, National
  7             Institutes of Health; publication no. 95-3371. (National Toxicology Program technical report series: no. 440).
  8       Neuhaus-Steinmetz, U.; Uffhausen, F.; Herz, U.; Renz, H. (2000) Priming of allergic immune responses by repeated
  9             ozone exposure in mice. Am. J. Respir.  Cell Mol. Biol. 23: 228-233.
10       Nichols, B. G.; Woods, J. S.; Luchtel, D. L.; Corral, J.; Koenig, J. Q. (2001) Effects of ozone exposure on nuclear
11             factor-KB activation and tumor necrosis factor-a expression  in human nasal epithelial cells. Toxicol. Sci.
12             60:356-362.
13       Nielsen, G. D.; Hougaard, K. S.; Larsen, S. T.; Hammer, M;  Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Alarie, Y.
14             (1999) Acute airway effects of formaldehyde and ozone inBALB/c mice. Hum. Exp. Toxicol. 18: 400-409.
15       Nino-Cabrera, H. G.; Colin-Barenque, L.; Avila-Costa, M.  R.; Espinosa-Villanueva, J.; Fortoul, T. L;
16             Rivas-Arancibia, S. (2002) Differences between hippocampus and cerebral cortex in aged rats in an oxidative
17             stress model. Int. J. Neurosci. 112: 373-381.
18       Noviski, N.; Brewer, J. P.; Skornik, W. A.; Galli, S. J.; Drazen, J. M.; Martin, T. R. (1999) Mast cell activation is
19             not required for induction of airway hyperresponsiveness by ozone in mice. J. Appl. Physiol. 86: 202-210.
20       Paige, R. C.; Royce, F. H.; Plopper, C. G.; Buckpitt, A. R. (2000a)  Long-term exposure to ozone increases acute
21             pulmonary centriacinar injury by 1-nitronaphthalene: I. Region-specific enzyme activity. J. Pharmacol. Exp.
22             Ther. 295: 934-941.
23       Paige, R. C.; Wong, V.; Plopper, C. G. (2000b) Long-term exposure to ozone increases acute pulmonary centriacinar
24             injury by 1-nitronaphthalene: II. Quantitative histopathology. J. Pharmacol. Exp. Ther. 295: 942-950.
25       Paquette, N. C.; Tankersley, C. G.; Zhang, L.-Y.; Kleeberger, S.  R. (1994) Repeated subacute ozone exposure of
26             inbred  mice: airway inflammation and ventilation. Exp. Lung Res. 20: 579-594.
27       Paz, C. (1997) Some consequences of ozone exposure  on health. Arch. Med. Res. 28: 163-170.
28       Paz, C.; Bazan-Perkins, B. (1992) Sleep-wake  disorganization in cats exposed to ozone. Neurosci. Lett.
29             140:270-272.
30       Paz, C.; Huitron-Resendiz, S. (1996) The effects of ozone exposure on the sleep-wake cycle and serotonin contents
31             in the pons of the rat.  Neurosci. Lett. 204: 49-52.
32       Pearson, A. C.; Bhalla, D. K. (1997) Effects of ozone onmacrophage adhesion in vitro and epithelial and
33             inflammatory responses in vivo: the role of cytokines.  J. Toxicol. Environ. Health 50: 143-157.
34       Peden, D. B.;  Dailey, L. (1995) Modulation of mast cell functions by in vitro ozone exposure. Am. J. Physiol.
35             268: L902-L910.
36       Pendino, K. J.; Shuler, R. L.; Laskin,  J. D.; Laskin, D.  L. (1994)  Enhanced production of interleukin-1, tumor
3 7             necrosis factor-a, and fibronectin by rat lung phagocytes following inhalation of a pulmonary irritant.  Am. J.
38             Respir. Cell Mol. Biol. 11: 279-286.
39       Petruzzi, S.; Fiore, M.; DeH'Omo, G.; Bignami, G.; Alleva, E. (1995) Medium and long-term behavioral effects in
40             mice of extended gestational exposure to ozone. Neurotoxicol. Teratol. 17: 463-470.
41       Petruzzi, S.; De Acetis, L.; Chiarotti, F.; Sorace,  A.; Alleva, E. (1999) Limited changes in handedness and morphine
42             reactivity in CD-I mice after pre- and postnatal  ozone  exposure. Acta Neurobiol. Exp. 59: 115-122.
43       Pinkerton, K.  E.; Weller, B.  L.; Menache, M. G.; Plopper, C.  G.  (1998) Consequences of prolonged inhalation of
44             ozone on F344/N rats: collaborative studies. Part XIII.  A comparison of changes in the tracheobronchial
45             epithelium and pulmonary acinus in male rats at 3 and 20  months. Cambridge, MA: Health Effects Institute;
46             research report no. 65.
47       Plopper, C. G.; Fanucchi, M. V. (2000) Do urban environmental pollutants exacerbate childhood lung diseases?
48             Environ. Health Perspect. 108: A252-A253.
49       Plopper, C. G.; Hatch, G. E.; Wong, V.; Duan, X.; Weir, A. J.; Tarkington, B. K.; Devlin, R. B.; Becker, S.;
50             Buckpitt, A. R. (1998) Relationship of inhaled ozone concentration to acute tracheobronchial epithelial injury,
51             site-specific ozone dose and glutathione depletion in rhesus monkeys. Am. J. Respir. Cell Mol. Biol.
52             19:387-399.
53       Polosa, R.; Holgate, S. T. (1997) Adenosine bronchoprovocation: a promising marker of allergic inflammation in
54             asthma? Thorax 52: 919-923.
         August 2005                                    5-86          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Pope, C. A., Ill; Burnett, R. T.; Thun, M. I; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
  2             cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA J. Am. Med.
  3             Assoc. 287: 1132-1141.
  4       Postlethwait, E. M.; Cueto, R.; Velsor, L. W.; Pryor, W. A. (1998) O3-induced formation of bioactive lipids:
  5             estimated surface concentrations and lining layer effects. Am. J. Physiol. 274: L1006-L1016.
  6       Postlethwait, E. M.; Joad, J. P.; Hyde, D. M.; Schelegle, E. S.; Brie, J. M.; Weir, A. J.; Putney, L. F.; Wong, V. J.;
  7             Velsor, L. W.; Plopper, C. G. (2000) Three-dimensional mapping of ozone-induced acute cytotoxicity in
  8             tracheobronchial airways of isolated perfused rat lung. Am. J. Respir. Cell Mol. Biol. 22: 191-199.
  9       Pryor, W. A.; Squadrito,  G. L.; Friedman, M. (1995) A new mechanism for the toxicity of ozone. Toxicol. Lett.
10             82/83:287-293.
11       Pryor, W. A.; Bermudez, E.; Cueto, R.; Squadrito, G. L. (1996) Detection of aldehydes in bronchoalveolar lavage of
12             rats exposed to ozone. Fundam. Appl.  Toxicol. 34: 148-156.
13       Pulfer, M. K.; Murphy, R. C. (2004) Formation of biologically active oxysterols during ozonolysis of cholesterol
14             present in lung surfactant. J. Biol. Chem. 279: 26331-26338.
15       Pulfer, M. K.; Taube, C.; Gelfand, E.; Murphy, R. C. (2005) Ozone exposure in vivo and formation of biologically
16             active oxysterols in the lung. J. Pharmacol. Exp. Ther. 312: 256-264.
17       Quinlan, T.; Spivack, S.; Mossman, B. T. (1994) Regulation of antioxidant enzymes in lung after oxidant injury.
18             Environ. Health Perspect. 102(suppl. 2): 79-87.
19       Rehle, D.; Leleux, D.;  Erdelyi, M.; Tittel, F.; Fraser, M.; Friedfeld, S.; et al. (2001) Ambient formaldehyde detection
20             with a laser spectrometer based on difference-frequency generation in PPLN. Appl. Phys. B: Lasers Opt.
21             72: 947-952.
22       Reinhart, P.  G.; Gupta, S. K.; Bhalla, D. K. (1999) Attenuation of ozone-induced lung injury by interleukin-10.
23             Toxicol. Lett. 110:35-42.
24       Renner, R. (2002) Bad air and birth defects. Environ. Health Perspect. 110: A291.
25       Ritz, B.; Yu, F. (1999) The effect of ambient carbon monoxide on low birth weight among children born in southern
26             California between 1989 and 1993. Environ. Health Perspect. 107: 17-25.
27       Rivas-Arancibia, S.; Vazquez-Sandoval, R.; Gonzalez-Kladiano, D.; Schneider-Rivas, S.; Lechuga-Guerrero, A.
28             (1998) Effects of ozone exposure in rats on memory and levels of brain and pulmonary superoxide dismutase.
29             Environ. Res. 76: 33-39.
30       Rivas-Arancibia, S.; Dorado-Martinez, C.; Borgonio-Perez, G.; Hiriart-Urdanivia, M.; Verdugo-Diaz, L.;
31             Duran-Vazquez, A.; Colin-Baranque, L.; Avila-Costa, M. R. (2000) Effects of taurine on ozone-induced
32             memory deficits and lipid peroxidation levels in brains of young, mature, and old rats. Environ. Res. 82: 7-17.
33       Rivas-Arancibia, S.; Dorado-Martinez, C.; Colin-Barenque, L.; Kendrick, K. M.; De la Riva, C.; Guevara-Guzman,
34             R. (2003) Effect of acute ozone exposure on locomotor behavior and striatal function. Pharmacol. Biochem.
35             Behav. 74: 891-900.
36       Rivas-Manzano, P.; Paz, C. (1999) Cerebellar morphological alterations in rats induced by prenatal ozone exposure.
37             Neurosci. Lett. 276: 37-40.
38       Rose, R. C.; Richer,  S. P.; Bode, A. M.  (1998) Ocular oxidants and antioxidant protection. Proc. Soc. Exp. Biol.
39             Med. 217: 397-407.
40       Savov, J. D.; Whitehead, G. S.; Wang, J.; Liao, G.; Usuka, J.; Peltz, G.; Foster, W. M.; Schwartz, D. A. (2004)
41             Ozone-induced  acute pulmonary injury in inbred mouse strains. Am. J. Respir. Cell Mol. Biol. 31: 69-77.
42       Schelegle, E. S.; Alfaro, M. F.; Putney, L.; Stovall, M.; Tyler, N.; Hyde, D. M. (2001) Effect of C-fiber-mediated,
43             ozone-induced rapid shallow breathing on airway epithelial injury in rats. J. Appl. Physiol. 91: 1611-1618.
44       Schelegle, E. S.; Miller, L. A.; Gershwin, L. J.; Fanucchi, M. V; Van Winkle, L. S.; Gerriets, J. E.; Walby, W. F.;
45             Mitchell, V; Tarkington, B. K.; Wong, V J.; Baker, G. L.; Pantle, L. M.; Joad, J. P.;  Pinkerton, K. E.;
46             Wu, R.; Evans,  M. J.; Hyde, D. M.;  Plopper, C. G. (2003a) Repeated episodes  of ozone inhalation amplifies
47             the effects of allergen sensitization and inhalation on airway immune and structural development in Rhesus
48             monkeys. Toxicol. Appl. Pharmacol. 191: 74-85.
49       Schelegle, E. S.; Walby, W. F.; Alfaro,  M. F.; Wong, V. J.; Putney, L.; Stovall, M. Y.; Sterner-Kock, A.; Hyde,
50             D. M.; Plopper, C. G. (2003b) Repeated episodes of ozone inhalation attenuates airway injury/repair and
51             release of substance P, but not adaptation. Toxicol. Appl. Pharmacol. 186: 127-142.
52       Schlesinger, R. B. (1995) Interaction of gaseous and paniculate pollutants in the respiratory  tract: mechanisms and
53             modulators. Toxicology 105: 315-325.
54       Schlesinger, R. B.; Cohen, M. D.; Gordon, T.; Nadziejko, C.; Zelikoff, J. T.; Sisco, M.; Regal, J. F.; Menache, M. G.
55             (2002a) Ozone differentially modulates airway responsiveness in atopic versus nonatopic guinea pigs.
56             Inhalation Toxicol. 14:431-457.
         August 2005                                    5-87          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Schlesinger, R. B.; Cohen, M; Gordon, T.; Nadziejko, C.; Zelikoff, J. T.; Sisco, M.; Regal, J. F.; Menache, M. G.
  2             (2002b) Ozone-induced modulation of airway hyperresponsiveness in guinea pigs. Boston, MA: Health
  3             Effects Institute; research report no. 109.
  4       Schwartz, D. A. (2002) TLR4 and LPS hyporesponsiveness in humans. Int. J. Hyg. Environ. Health 205: 221-227.
  5       Segura, P.; Montano, L. M.; Bazan-Perkins, B.; Gustin, P.; Vargas, M. H. (1997) Ozone at high-pollution urban
  6             levels causes airway hyperresponsiveness to substance P but not to other agonists. Environ. Toxicol.
  7             Pharmacol. 3: 91-95.
  8       Sen, S.; Dulchavsky, S. A.; Dutta, S. (1993) Effects of triiodothyronine (T3)  supplementation upon ozone-induced
  9             lung injury. Free Radic. Res. Commun. 18: 299-308.
10       Shore, S. A.; Abraham, J. H.; Schwartzman, I. N.; Murthy, G. G.; Laporte, J. D. (2000) Ventilatory responses to
11             ozone are reduced in immature rats. J. Appl. Physiol. 88: 2023-2030.
12       Shore, S. A.; Johnston, R. A.; Schwartzman, I. N.; Chism, D.; Krishna Murthy, G. G. (2002) Ozone-induced airway
13             hyperresponsiveness is reduced in immature mice. J. Appl. Physiol. 92: 1019-1028.
14       Shore, S. A.; Rivera-Sanchez, Y. M.; Schwartzman, I. N.; Johnston, R. A. (2003) Responses to ozone are increased
15             in obese mice. J. Appl. Physiol. 95: 938-945.
16       Sindhu, R. K.; Mautz, W. J.; Kikkawa, Y. (1998) Chronic exposure to ozone and nitric acid vapor results in
17             increased levels of rat pulmonary putrescine. Arch. Toxicol. 72: 445-449.
18       Slade, R.; Watkinson, W. P.; Hatch, G. E. (1997) Mouse strain differences in ozone dosimetry and body temperature
19             changes. Am. J. Physiol. 272: L73-L77.
20       Sorace, A.;  De Acetis, L.; Alleva, E.; Santucci, D. (2001) Prolonged exposure to low doses of ozone: short- and
21             long-term changes to behavioral performance in mice. Environ. Res. 85: 122-134.
22       Spannhake, E. W. (1996) Down-regulation of canine airway mast cell function following exposure to ozone in vivo.
23             Exp. Lung Res. 22: 163-178.
24       Sterner-Kock, A.; Kock, M.; Braun, R.; Hyde, D. M. (2000) Ozone-induced epithelial injury in the ferret is similar
25             to nonhuman primates. Am. J. Respir. Crit. Care Med. 162: 1152-1156.
26       Sun, J.; Chung, K. F. (1997) Interaction of ozone exposure with airway hyperresponsiveness and inflammation
27             induced by trimellitic anhydride in sensitized guinea pigs. J. Toxicol. Environ. Health 51: 77-87.
28       Sun, J.; Koto, H.; Chung, K. F. (1997) Interaction of ozone and allergen challenges  on bronchial responsiveness and
29             inflammation in sensitised guinea pigs. Int. Arch. Allergy Immunol. 112: 191-195.
30       Szarek, J. L.; Stewart, N. L.; Zhang, J. Z.; Webb, J. A.; Valentovic, M. A.; Catalano, P. (1995) Contractile responses
31             and structure of small bronchi isolated from rats after 20 months' exposure to ozone. Fundam. Appl. Toxicol.
32             28: 199-208.
33       Tager, I. B. (1999) Air pollution and lung function growth. Is it ozone? [editorial]. Am. J. Respir. Crit. Care Med.
34             160:387-389.
35       Takahashi, T.; Miura, M.; Katsumata, U.; Ichinose, M.; Kimura, K.; Inoue, H.; Takishima, T.; Shirato, K. (1993)
36             Involvement of superoxide in ozone-induced airway hyperresponsiveness in anesthetized cats. Am. Rev.
37             Respir. Dis. 148: 103-106.
38       Takahashi, N.; Yu, X.-Y.; Schofield, B. H.; Kleeberger, S. R.; Scott, A. L.; Hasegawa, S.; Spannhake, E. W. (1995a)
3 9             Expression of ICAM-1 in airway epithelium after acute ozone exposure in the mouse. J. Appl. Physiol.
40             79: 1753-1761.
41       Takahashi, M.; Kleeberger, S. R.; Croxton, T. L. (1995b) Genetic control of susceptibility to ozone-induced changes
42             in mouse trachea! electrophysiology.  Am. J. Physiol. 269: L6-L10.
43       Takebayashi, T.; Abraham, J.; Murthy, G. G. K.; Lilly, C.; Rodger, I.; Shore, S. A. (1998) Role of tachykinins in
44             airway responses to ozone in rats. J. Appl. Physiol. 85: 442-450.
45       Tankersley, C.  G.; Kleeberger, S. R. (1994)  Ozone-induced inflammation and altered ventilation in genetically
46             susceptible mice: a comparison of acute and subacute exposures. Toxicol Lett. 72: 279-289.
47       Tankersley, C.  G.; Fitzgerald, R. S.; Mitzner, W. A.; Kleeberger, S. R. (1993) Hypercapnic ventilatory responses in
48             mice differentially susceptible to acute ozone exposure. J. Appl. Physiol. 75:  2613-2619.
49       Tesfaigzi, J.; Hotchkiss, J. A.; Harkema, J. R. (1998) Expression of the Bcl-2 protein in nasal epithelia of F344/N
50             rats during mucous cell metaplasia and remodeling. Am. J. Respir. Cell Mol.  Biol. 18: 794-799.
51       Thiele, J. J. (2001) Oxidative targets in the stratum corneum. A new basis for antioxidative strategies. Skin
52             Pharmacol. Appl.  Skin Physiol.  14(suppl. 1): 87-91.
53       Thiele, J. J.; Traber, M. G.; Packer, L.  (1998) Depletion of human stratum corneum vitamin E: an early and sensitive
54             in vivo marker of UV induced photo-oxidation. J. Invest. Dermatol. 110: 756-761.
         August 2005                                     5-88          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Thiele, J. I; Hsieh, S. N.; Briviba, K.; Sies, H. (1999) Protein oxidation in human stratum corneum: susceptibility of
  2             keratins to oxidation in vitro and presence of a keratin oxidation gradient in vivo. J. Invest. Dermatol.
  3             113:335-339.
  4       Tsai, J.-J.; Lin, Y.-C.; Kwan, Z.-H.; Kao, H.-L. (1998) Effects of ozone on ovalbumin sensitization in guinea pigs.
  5             J. Microbiol. Immunol. Infect. 31: 225-232.
  6       U.S. Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park,
  7             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
  8             nos. EPA/600/8-9!/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525, and
  9             PB95-124517.
10       U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
11             Research Triangle Park, NC: Office of Research and Development; report nos. EP A/600/ AP-93/004aF-cF.  3v.
12             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available online at:
13             www.epa.gov/ncea/ozone.htm.
14       Uhlson, C; Harrison, K.; Allen, C. B.; Ahmad, S.; White, C. W.; Murphy, R. C. (2002) Oxidized phospholipids
15             derived from ozone-treated lung surfactant extract reduce macrophage and epithelial cell viability. Chem.
16             Res. Toxicol. 15: 896-906.
17       Ulrich, M. M. W.; Alink, G. M; Kumarathasan, P.; Vincent, R.; Boere, A. J.; Cassee, F. R. (2002) Health effects
18             and time course of paniculate matter on the cardiopulmonary system in rats with lung inflammation.
19             J. Toxicol. Environ. HealthPart A65:  1571-1595.
20       Valacchi, G.; Weber, S. U.; Luu, C.; Cross, C. E.; Packer, L. (2000) Ozone potentiates vitamin E depletion by
21             ultraviolet radiation in the murine stratum corneum. FEES Lett. 466: 165-168.
22       Valacchi, G.; Pagnin, E.; Okamoto, T.; Corbacho, A. M.; Olano, E.; Davis, P. A.; Van der Vliet, A.; Packer, L.;
23             Cross, C. E. (2003) Induction of stress proteins and MMP-9 by 0.8 ppm of ozone in murine skin. Biochem.
24             Biophys. Res. Commun. 305: 741-746.
25       Valverde, M.; del Carmen Lopez, M.; Lopez, I.; Sanchez, I.; Fortoul, T. I.; Ostrosky-Wegman, P.; Rojas, E. (1997)
26             DNA damage in leukocytes and buccal and nasal epithelial cells of individuals exposed to air pollution in
27             Mexico City. Environ. Mol. Mutagen. 30:  147-152.
28       Van Bree, L.; Dormans, J. A. M. A.; Boere, A. J. F.; Rombout, P. J. A. (2001) Time study on development and
29             repair of lung injury following ozone exposure in rats. Inhalation Toxicol. 13: 703-717.
30       Van Bree, L.; Dormans, J. A. M. A.; Koren, H. S.; Devlin, R. B.; Rombout, P. J. A. (2002) Attenuation and recovery
31             of pulmonary injury in rats following short-term, repeated daily exposure to ozone. Inhalation Toxicol.
32             14:883-900.
33       Van Hoof, I. H. J. M.; Van Bree, L.; Bast, A.  (1996) Changes in receptor function by oxidative stress in guinea pig
34             trachea! smooth muscle. Cent. Eur. J. Public Health 4(suppl.): 3-5.
35       Van Hoof, H. J. M.; Van Acker, F.  A. A.; Voss, H.-P.; Van Bree, L.; Bast, A. (1997a) Acute exposure to ozone does
36             not influence neuroreceptor density and sensitivity in guinea pig lung. Toxicol. Lett. 90: 53-60.
37       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)
3 8             Changes in neuroreceptor function of tracheal smooth muscle following acute ozone exposure of guinea pigs.
39             Toxicology 120: 159-169.
40       Vanda, B.; de Buen, N.; Jasso, R.; Valero, G.; Vargas, M. H.; Olmos, R.; Arreola, J. L.; Santillan, P.; Alonso, P.
41             (1998) Inflammatory cells and ferruginous bodies  in bronchoalveolar lavage in urban dogs. Acta Cytol.
42             42: 939-944.
43       Vargas, M. H.; Romero, L.; Sommer, B.; Zamudio, P.; Gustin, P.; Montano, L. M. (1998) Chronic exposure to ozone
44             causes tolerance to airway hyperresponsiveness in guiea pigs: lack of SOD  role. J. Appl. Physiol. 84:
45             1749-1755.
46       Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994a) Ozone increases amino- and
47             carboxy-terminal atrial natriuretic factor prohormone peptides in lung, heart, and circulation. J. Biochem.
48             Toxicol. 9: 107-112.
49       Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994b) Increase in atrial natriuretic factor in
50             the lungs, heart, and circulatory system owing to ozone. Chest 105: 1551-1554.
51       Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994c) Ozone increases atrial natriuretic
52             peptides in heart, lung and circulation of aged vs. adult animals. Gerontology (Basel) 40: 227-236.
53       Vesely, K. R.; Schelegle, E. S.; Stovall, M. Y.; Harkema, J. R.; Green, J. F.; Hyde, D. M. (1999a) Breathing pattern
54             response and epithelial labeling in ozone-induced airway injury in neutrophil-depleted rats. Am.  J. Respir.
55             Cell Mol. Biol. 20: 699-709.
         August 2005                                    5-89         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Vesely, K. R.; Hyde, D. M; Stovall, M. Y.; Harkema, J. R.; Green, J. F.; Schelegle, E. S. (1999b)
  2             Capsaicin-sensitive C-fiber-mediated protective responses in ozone inhalation in rats. J. Appl. Physiol.
  3             86:951-962.
  4       Vincent, R.; Janzen, E. G.; Chen, G.; Kumarathasan, P.; Haire, D. L.; Guenette, J.; Chen, J. Z.; Bray, T. M. (1996)
  5             Spin trapping study in the lungs and liver of F344 rats after exposure to ozone. Free Radical Res. 25: 475-488.
  6       Vincent, R.; Bjarnason, S. G.; Adamson, I. Y. R.; Hedgecock, C.; Kumarathasan, P.; Guenette, J.; Potvin, M.;
  7             Goegan, P.; Bouthillier, L. (1997) Acute pulmonary toxicity of urban paniculate matter and ozone. Am. J.
  8             Pathol. 151:  1563-1570.
  9       Vyskocil, A.; Viau, C.; Lamy, S. (1998) Peroxyacetyl nitrate: review of toxicity. Hum. Exp. Toxicol. 17: 212-220.
10       Wagner, J. G.; Hotchkiss, J. A.; Harkema, J. R. (200la) Effects of ozone and endotoxin coexposure on rat airway
11             epithelium: potentiation of toxicant-induced alterations.  Environ. Health Perspect. 109(suppl. 4): 591-598.
12       Wagner, J. G.; VanDyken, S. J.; Hotchkiss, J. A.; Harkema, J. R. (200Ib) Endotoxin enhancement of ozone-induced
13             mucous cell metaplasia is neutrophil-dependent in rat nasal epithelium. Toxicol. Sci. 60: 338-347.
14       Wagner, J. G.; Hotchkiss, J. A.; Harkema, J. R. (2002) Enhancement of nasal inflammatory and epithelial responses
15             after ozone and allergen coexposure in brown Norway rats. Toxicol. Sci.  67: 284-294.
16       Wagner, J. G.; VanDyken, S. J.; Wierenga, J. R.; Hotchkiss, J. A.; Harkema, J.  R. (2003) Ozone exposure enhances
17             endotoxin-induced mucous cell metaplasia in rat pulmonary airways. Toxicol. Sci. 74: 437-446.
18       Wang, G.; Umstead, T. M.; Phelps, D. S.; Al-Mondhiry, H.; Floras, J. (2002) The effect of ozone exposure on the
19             ability of human surfactant protein A variants to stimulate cytokine production. Environ. Health Perspect.
20             110:79-84.
21       Watkinson, W. P.; Wiester, M. J.; Highfill, J. W. (1995) Ozone toxicity in the rat. I. Effect of changes in ambient
22             temperature on extrapulmonary physiological parameters. J. Appl. Physiol. 78: 1108-1120.
23       Watkinson, W. P.; Campen, M. J.; Nolan, J. P.; Costa, D. L. (2001) Cardiovascular and systemic responses to
24             inhaled pollutants in rodents: effects  of ozone and paniculate matter. Environ. Health Perspect.
25             109(suppl. 4): 539-546.
26       Watkinson, W. P.; Campen, M. J.; Wichers, L. B.; Nolan, J. P.; Costa, D. L. (2003) Cardiac and thermoregulatory
27             responses to inhaled pollutants in healthy and compromised rodents: modulation via interaction with
28             environmental factors. Environ. Res. 92: 35-47.
29       Watt, K. C.; Plopper, C. G.; Weir, A. J.; Tarkington, B.; Buckpitt, A. R. (1998)  Cytochrome P450 2E1 in rat
30             tracheobronchial airways: response to ozone  exposure. Toxicol. Appl. Pharmacol. 149: 195-202.
31       Wattiez, R.; Noel-Georis, L; Cruyt, C.; Broeckaert,  F.; Bernard, A.; Falmagne, P. (2003) Susceptibility to oxidative
32             stress: proteomic analysis of bronchoalveolar lavage from ozone-sensitive and  ozone-resistant strains of mice.
33             Proteomics 3: 658-665.
34       Weber, S. U.; Thiele, J. J.; Cross, C. E.; Packer, L. (1999) Vitamin C, uric acid, and glutathione gradients in murine
3 5             stratum corneum and their susceptibility to ozone exposure. J. Invest. Dermatol. 113:1128-1132.
36       Weber, S. U.; Jothi,  S.; Thiele, J. J. (2000) High-pressure liquid chromatography analysis of ozone-induced
37             depletion of hydrophilic and lipophilic antioxidants in murine skin. Methods Enzymol. 319: 536-546.
38       Weber, S. U.; Han, N.; Packer, L. (2001) Ozone: an emerging oxidative stressorto skin. Curr. Probl. in Dermatol.
39             29:52-61.
40       Weller, B. L.; Crapo, J. D.; Slot, J.; Posthuma, G.; Plopper, C. G.; Pinkerton, K. E. (1997) Site- and cell-specific
41             alteration of lung copper/zinc and manganese superoxide dismutases by chronic ozone exposure. Am. J.
42             Respir. Cell Mol. Biol. 17: 552-560.
43       Weller, B. L.; Witschi, H.; Pinkerton, K. E.  (2000) Quantitation and localization of pulmonary manganese
44             superoxide dismutase and tumor necrosis factor a following exposure to ozone and nitrogen dioxide. Toxicol.
45             Sci. 54: 452-461.
46       Wells, C. A.; Ravasi, T.; Faulkner, G. J.; Carninci, P.; Okazaki, Y.; Hayashizaki, Y.; Sweet, M.; Wainwright, B. J.;
47             Hume, D. A. (2003) Genetic control of the innate immune response. BMC Immunol. 4: 5. Available:
48             http://www.biomedcentral.eom/1471-2172/4/5 [18 February, 2003]
49       Wiester, M. J.; Watkinson, W. P.; Costa, D. L.; Crissman, K. M.; Richards, J. H.; Winsett, D. W.; Highfill, J. W.
50             (1996) Ozone toxicity in the rat. III. Effect of changes in ambient temperature on pulmonary parameters.
51             J. Appl. Physiol. 81: 1691-1700.
52       Wilkins, C. K.; Clausen, P. A.; Wolkoff, P.; Larsen, S. T.; Hammer, M.; Larsen, K.; Hansen, V; Nielsen, G. D.
53             (2001) Formation of strong irritants in mixtures of isoprene/ozone and isoprene/ozone/nitrogen dioxide.
54             Environ. Health Perspect. 109: 937-941.
55       Witschi, H.; Espiritu, L; Pinkerton, K. E.; Murphy, K.; Maronpot, R. R. (1999) Ozone carcinogenesis revisited.
56             Toxicol. Sci. 52: 162-167.
         August 2005                                    5-90         DRAFT-DO NOT QUOTE OR CITE

-------
 1       Wu, Z.-X.; Morton, R. F.; Lee, L.-Y. (1997) Role of tachykinins in ozone-induced airway hyperresponsiveness to
 2             cigarette smoke in guinea pigs. J. Appl. Physiol. 83: 958-965.
 3       Yamauchi, T.; Shima, M; Kuwaki, T.; Ando, M; Ohmichi, M.; Fukuda, Y.; Adachi, M. (2002) Acute effects of
 4             ozone exposure on lung function in mice sensitized to ovalbumin. Toxicology (Ireland) 172: 69-78.
 5       Yu, M.; Pinkerton, K. E.; Witschi, H. (2002) Short-term exposure to aged and diluted sidestream cigarette smoke
 6             enhances ozone-induced lung injury inB6C3Fl mice. Toxicol. Sci. 65: 99-106.
 7       Zhang, L.-Y.; Levitt, R. C.; Kleeberger, S. R. (1995) Differential susceptibility to ozone-induced airways
 8             hyperreactivity in inbred strains of mice.  Exp. Lung Res. 21: 503-518.
 9       Zhao, Q.; Simpson, L. G.; Driscoll, K. E.; Leikauf, G. D. (1998) Chemokine regulation of ozone-induced neutrophil
10             and monocyte inflammation. Am.  J. Physiol. 274: L39-L46.
11
         August 2005                                   5-91         DRAFT-DO NOT QUOTE OR CITE

-------
 i           6.  CONTROLLED HUMAN EXPOSURE  STUDIES
 2                           OF OZONE AND RELATED
 3                        PHOTOCHEMICAL OXIDANTS
 4
 5
 6     6.1   INTRODUCTION
 7          In the previous chapter, results of ozone (O3) studies in laboratory animals and in vitro test
 8     systems were presented. The extrapolation of results from animal studies is one mechanism by
 9     which information on potential adverse human health effects from exposure to O3 is obtained.
10     More direct evidence of human health effects due to O3 exposure can be obtained through
11     controlled human exposure studies of volunteers or through field and epidemiologic studies of
12     populations exposed to ambient O3 (see Chapter 7). Controlled human exposure studies
13     typically use fixed concentrations of O3 under carefully regulated environmental  conditions and
14     subject activity levels.  This chapter discusses studies in which volunteers were exposed for up
15     to 8 h to between 0.08 to 0.75 ppm O3 while at rest or during varying intensities of exercise.
16          The majority of controlled human studies have investigated the effects of exposure to O3 in
17     young nonsmoking healthy adults (18 to 35 years of age) performing continuous  exercise (CE)
18     or intermittent exercise (IE).  Varied combinations of O3 concentration, exercise routine, and
19     exposure duration have been used in these studies. The responses to ambient O3  concentrations
20     include  decreased inspiratory capacity; mild bronchoconstriction; rapid, shallow breathing
21     patterns during exercise; and symptoms of cough and pain on deep inspiration. Reflex inhibition
22     of inspiration results in a decrease in forced vital capacity (F VC) and total lung capacity (TLC)
23     and, in combination with mild bronchoconstriction, contributes to a decrease in the forced
24     expiratory volume in 1 s (FEVj). In addition to physiological pulmonary responses and
25     respiratory symptoms, O3 has been shown to result in airway hyperresponsiveness, epithelial
26     permeability, and inflammation.
27          The most salient observations from studies reviewed in the 1996 EPA Ozone Air Quality
28     Criteria Document or O3 AQCD (U.S. Environmental Protection Agency, 1996) were that:
29     (1) young healthy adults exposed to O3 concentrations >0.08 ppm develop significant reversible,
30     transient decrements in pulmonary function if minute ventilation (VE ) or duration of exposure is
       August 2005                              6-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1      increased sufficiently, (2) children experience similar spirometric responses but lesser symptoms
 2      from O3 exposure relative to young adults, (3) O3-induced spirometric responses are decreased in
 3      the elderly relative to young adults, (4) there is a large degree of intersubject variability in
 4      physiologic and symptomatic responses to O3 but responses tend to be reproducible within a
 5      given individual over a period of several months, and (5) subjects exposed repeatedly to O3 for
 6      several days develop a tolerance to successive exposures, as demonstrated by an attenuation of
 7      responses, which is lost after about a week without exposure.
 8           There are several important limitations associated with these clinical studies: (1) the
 9      ability to study only short-term, acute effects; (2) difficulties in trying to link short-term effects
10      with long-term consequences; (3) the use of a small number of volunteers that may not be
11      representative of the general population; and (4) the statistical  limitations associated with the
12      small sample size. Sample size affects the power of a study, and having a small number of
13      samples causes a risk of Type II error, i.e., the incorrect conclusion that no difference exists
14      between treatments or groups when comparisons are not significantly different. This affects the
15      confidence in estimates  of a minimum O3 concentration at which some degree of pulmonary
16      impairment will occur in both the general population and susceptible subpopulations.  As a
17      result, the conclusions drawn from many of the studies cited in this chapter may underestimate
18      the presence of responses at low O3 concentrations and low activity levels.
19           Most of the scientific information summarized in this chapter comes from the literature
20      published since the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996).  In addition
21      to further study of physiological pulmonary responses and symptoms of breathing discomfort,
22      much of this literature has focused on mechanisms of inflammation and cellular responses to
23      injury induced by O3 inhalation. A more thorough discussion and review of this literature
24      appears in Annex AX6 of this document.  In summarizing the literature,  effects are described if
25      they are statistically significant at a probability (p-value) of less than 0.05; otherwise, trends are
26      noted as such.
27           As spirometry typically improves in healthy young adults with exercise exposures to
28      filtered air (FA), the term "O3-induced" is used herein and in the annex to designate effects that
29      have been corrected for responses during FA exposures. For healthy adults, an O3-induced
30      change in lung function is the difference between the decrement experienced with O3 exposure
31      and the improvement observed with FA exposure.  However, the distinction between an O3-

        August 2005                               6-2        DRAFT-DO NOT QUOTE OR CITE

-------
 1     induced change and a post- versus preexposure change is particularly important in individuals
 2     with respiratory disease who may experience exercise-induced decrements in pulmonary
 3     function during both FA and O3 exposures.  Hence, in subjects with respiratory disease, exercise-
 4     induced responses could be mistaken for O3-induced responses in the absence of a correction for
 5     FA responses.
 6
 7
 8     6.2   PULMONARY FUNCTION EFFECTS OF OZONE EXPOSURE
 9           IN HEALTHY SUBJECTS
10     6.2.1   Introduction
11          As reviewed in the 1986 and 1996 O3 AQCD's (U.S. Environmental Protection Agency,
12     1986, 1996), 0.5 ppm is the lowest O3 concentration at which statistically significant reductions
13     in FVC and FEVj have been reported in sedentary subjects. On average, young adults (n = 23;
14     mean age, 22 yrs) exposed at rest for 2 h to 0.5 ppm O3 had O3-induced decrements of-4% in
15     FVC and -7%  in FEVj (Folinsbee et al., 1978; Horvath et al., 1979).  During exercise,
16     spirometric and symptoms responses are observed at lower O3 concentrations. For acute
17     exposures of 2  h or less to >0.12 ppm  O3, if VE is sufficiently increased by exercise, healthy
18     human subjects generally experience decreases in TLC, inspiratory capacity (1C), FVC, FEVl5
19     mean forced expiratory flow from 25% to 75% of FVC (FEF25_75), and tidal volume (VT)
20     and increases in specific airways resistance (sRaw), breathing frequency (fB), and airway
21     responsiveness. These exposures also cause symptoms of cough, pain on deep inspiration,
22     shortness of breath, throat irritation, and wheezing. With exposures of 4- to 8-h in duration,
23     statistically significant pulmonary function and symptoms responses are  observed at
24     lower O3 concentrations and  lower VE than in shorter duration studies.
25
26     6.2.2   Acute Exposure for Up  to 2 h
27          With heavy CE (VE = 89 L/min), an Cyinduced decrement of 9.7% in FEVj has
28     been reported for healthy young adults (n =  17; age, 24 ± 3 yrs) exposed for only 1 h to
29     0.12 ppm O3 (Gong et al., 1986).  With moderate-to-heavy IE (15 min intervals of rest and
30     exercise [VE =  68 L/min]), McDonnell et al. (1983) reported a physiologically small, but


       August 2005                               6-3        DRAFT-DO NOT QUOTE OR CITE

-------
 1      significant, O3-induced decrement of 3.4% in FEVj for young healthy adults (n = 22, age,
 2      22 ± 3 yrs) exposed for 2 h to 0.12 ppm O3. Using the same 2 h IE exposure protocol, Linn et al.
 3      (1986) found no statistically significant spirometic responses at O3 concentrations of 0.16 ppm
 4      and lower. However, the subjects in the Linn et al.  (1986) study were potentially exposed
 5      concurrently in Los Angeles to ambient O3 levels of between 0.12 and 0.16 ppm and were on
 6      average 3 yrs older than the subjects in the McDonnell et al. (1983) study.  (The attenuating
 1      effects of increasing age and repeated O3 exposures are discussed in Sections 6.5.1 and 6.6,
 8      respectively.)  The disparities between the Linn et al. (1986) and McDonnell et al. (1983) studies
 9      demonstrate the difficulty in determining a no-effect-level for O3 based on relatively small study
10      populations.
11           Studies analyzing large data sets (>300 subjects) provide better predictive ability of acute
12      changes in FEVj at low levels of O3 and VE than possible via comparisons between smaller
13      studies. Such an analysis was performed by McDonnell et al. (1997), who examined FEVj
14      responses in 485 healthy white males  (18 to 36 years of age; subjects recruited from the area
15      around Chapel Hill, NC) exposed once for 2 h to O3 concentrations of up to 0.40 ppm at rest or
16      with IE.  Decrements in FEVj were modeled by sigmoid-shaped curve as a function of subject
17      age, O3 concentration, VE, and duration of exposure. Regarding applicability to the general
18      population, the McDonnell et al. (1997) model has an apparent limitation of considering only
19      data for white males. However, two other large studies (n = 372; 18 to 35 yrs of age; subjects
20      recruited from the area around Chapel Hill, NC) found no significant gender or race effects on
21      spirometric responses to O3 exposure  (Seal et al., 1993,  1996).
22           Ultman et al. (2004) recently reported pulmonary responses in 60 young heathy
23      nonsmoking adults (32 M, 28 F) exposed to 0.25 ppm O3 for 1 h with CE at a target VE of
24      30 L/min. Consistent with findings reported in the  1996 O3 criteria document, considerable
25      intersubject variability in FEVj decrements was reported by Ultman et al. (2004) with responses
26      ranging from a 4% improvement to a  56% decrement. One-third of the subjects had FEVj
27      decrements of >15% and 7% of the subjects had decrements of >40%.  It should be pointed out
28      that the McDonnell et al. (1997) model predicts only average responses.  In a more recent study,
29      McDonnell et al. (1999) also reported a model predicting average symptom responses from O3
30      exposure. Unfortunately, neither of these papers (McDonnell et al., 1997, 1999) provide
        August 2005                              6-4        DRAFT-DO NOT QUOTE OR CITE

-------
 1      predictions of intersubject variability in response.  (Section 6.4 of this Chapter discusses
 2      intersubject variability in response to O3 exposure).
 3           In addition to the overt effects of O3 exposure on the large airways as indicated by
 4      spirometric responses, O3 exposure also affects the function of the small airways and
 5      parenchymal lung. Foster et al. (1993,  1997) examined the effect of O3 on ventilation
 6      distribution in healthy adult males. In healthy nonsmoking males (26.7 ± 7 years old) exposed to
 7      FA or 0.33 ppm O3 for 2 h with IE, there was a significant reduction in the ventilation to the
 8      lower-lung (31% of lung volume) and significant increases in ventilation to the upper- and
 9      middle-lung regions relative to the FA values in 7 of the 9 subjects (Foster et al., 1993).
10      In another study, 15 healthy nonsmoking males (25.4 ± 2 years old) were exposed to FA or
11      0.35 ppm O3 for 2.2 h with IE (Foster et al., 1997). Following O3 exposure, an inert gas washout
12      was delayed and resembled a two-compartment washout, whereas pre-O3 exposure a log-linear
13      gas clearance as a function of expired volume resembled a single-compartment washout. The
14      pronounced  slow phase of gas washout occurring post-O3, represented a 24% decrease in the
15      washout rate relative to pre-O3. At 24-h post-O3, 6 of the 12 subjects still had [or developed]
16      a delayed washout relative to the pre-O3 maneuver. This suggests a prolonged O3  effect on the
17      small airways and ventilation distribution in some individuals.
18
19      6.2.3   Prolonged Ozone Exposures
20           In the exposure range of 0.08 to 0.16 ppm O3, a number of studies using moderate
21      quasi continuous exercise (QCE; 50 min exercise and 10 min rest per h) for 4 to 8  h have
22      shown significant responses under the following conditions:  0.16 ppm for 4 h with QCE
23      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
24      40 L/min (Adams, 2002; Adams, 2003a; Folinsbee et al., 1988; Horstman et al., 1990), and
25      0.12 ppm for 8 h of IE (30 min per h) at VE « 40 L/min (Hazucha et al., 1992).  Symptoms and
26      spirometric responses increased with duration of exposure, O3 concentration, and total VE.
27      Airway resistance is only modestly affected with moderate or even heavy exercise combined
28      with O3 exposure (Folinsbee et al., 1978; McDonnell et al., 1983; Seal et al., 1993).
29
        August 2005                              6-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      6.2.3.1   Effect of Exercise Ventilation Rate on FEVt Response to 6.6 h Ozone Exposure
 2           It is well established that response to O3 exposure is a function of VE in studies of 2 h or
 3      less in duration (See Section AX6.2.2).  It is reasonable to expect that response to a prolonged
 4      6.6-h O3 exposure is also a function of VE, although quantitative analyses are lacking. Data
 5      from  five similar prolonged exposure studies are available for evaluation of FEVj responses as a
 6      function of exercise VE (Adams, 2000; Adams and Ollison, 1997; Folinsbee et al., 1988,  1994;
 7      Horstman et al., 1990). Each of these studies exposed similarly aged subjects (mean ages 22 to
 8      25 yrs) to 0.12 ppm O3 for 6.6 h. In total, ten sets of mean FEVj decrements were available for
 9      exercise VE ranging from 20 to 43 L/min, although no data were available for VE between
10      20 and 30 L/min (data illustrated in Figure AX6-2). As in 2 h exposure studies, FEVj
11      decrements are a function of VE in prolonged 6.6-h exposure studies as demonstrated by a
12      significant correlation between  these variables (Pearson, r = 0.95, p < 0.001; Spearman, r = 0.84,
13      p<0.01).
14
15      6.2.3.2   Exercise Ventilation Rate as a Function of Body/Lung  Size on FEVt Response
16              to 6.6 h Ozone Exposure
17           Based on the assumption that the total inhaled O3 dose (product of O3 concentration,
18      exposure duration,  and VE) is proportional to the lung size, exercise VE is typically selected to be
19      a multiple of body  surface area  (BSA) or FVC. Data from several recent studies do not support
20      the contention that VE should be normalized. In an analysis of data from 485 young adults,
21      McDonnell et al. (1997) found that any effect of BSA, height, or baseline FVC on percent
22      decrement in FEVj was small to nonexistent. This is consistent with Messineo and Adams
23      (1990), who compared pulmonary function responses in young adult women having small
24      (n = 14)  or large (n = 14) lung sizes (mean FVC of 3.74 and 5.11 L, respectively) and found no
25      significant group difference in FEVj decrements.  For 30 subjects (15M, 15F) exposed to
26      0.12 ppm O3 for 6.6 h, Adams (2000) also reported that FEVj responses were more closely
27      related to VE than to VE normalized to BSA.  The O3 dosimetry study of Bush et al. (1996)
28      suggested that normalization of the O3 dose might more appropriately be a function of anatomic
29      dead  space.  Ozone penetrates deeper into the lungs of individuals with larger conducting airway
       August 2005                              6-6        DRAFT-DO NOT QUOTE OR CITE

-------
 1     volumes, however, FEVj responses in subjects exposed for 2 h to 0.25 ppm O3 did not appear to
 2     be associated with O3 uptake (Ultman et al., 2004).
 3
 4     6.2.3.3   Comparison of 2 h IE to 6.6 h O3 Exposure Effects on Pulmonary Function
 5           Adams (2003b) examined whether prolonged 6.6-h QCE exposure to a relatively low O3
 6     concentration (0.08 ppm) and the 2-h IE exposure at a relatively high O3 concentration (0.30
 7     ppm) elicited consistent individual subject FEVj responses. Individual subject O3 exposure
 8     reproducibility was first examined via a regression plot of the postexposure FEVj response to the
 9     6.6-h chamber exposure as a function of postexposure FEVj response to the 2-h IE chamber
10     exposure. The R2 of 0.40, although statistically significant, was substantially less than that
11     observed in a comparison of individual FEVj response to the two 2-h IE exposures by chamber
12     and face mask, respectively (R2 = 0.83). The Spearman rank order correlation for the chamber
13     6.6-h and chamber 2-h exposure comparison was also substantially less (0.49) than that obtained
14     for the two 2-h IE exposures (0.85). The primary reason for the greater variability in the
15     chamber 6.6-h exposure FEVj response as a function of that observed for the two 2-h IE
16     exposures is very likely related to the increased variability in response upon repeated exposure
17     to O3 concentrations lower than 0.18 ppm (R = 0.57, compared to a mean R of 0.82 at higher
18     concentrations) reported by McDonnell et al. (1985a). This rationale is supported by the lower r2
19     (0.40) observed by Adams (2003b) for the FEVj responses found in 6.6 h chamber and face
20     mask exposures to 0.08 ppm O3, compared to an r2 of 0.83 observed for responses found at 0.30
21     ppm O3.
22
23     6.2.4  Triangular Ozone Exposures
24           To further explore the factors that determine responsiveness to O3, Hazucha et al. (1992)
25     designed a protocol to examine the effect of varying, rather than constant, O3 concentrations.
26     Subjects were exposed to an O3 level that increased linearly from 0 to 0.24 ppm for the first 4 h
27     and then decreased linearly from 0.24 to 0 ppm over the second 4 h of the 8 h exposure
28     (triangular concentration profile) and to a constant level exposure of 0.12 ppm O3 for 8 h.  While
29     total inhaled O3 doses for the constant and the triangular concentration profile were almost
30     identical, the FEVj response was dissimilar. For the constant 0.12 ppm O3 exposure, FEVj
31     declined -5% by the fifth hour and then remained at that level. With the triangular O3

       August 2005                              6-7        DRAFT-DO NOT QUOTE OR CITE

-------
 1     concentration profile, there was minimal FEVj response over the first 3 h followed by a rapid
 2     decrease in FEVj (-10.3%) over the next 3 h. During the seventh and eighth hours, mean FEVj
 3     decrements improved to -6.3% as the O3 concentration decreased from 0.12 to 0.00 ppm
 4     (mean = 0.06 ppm).
 5          More recently, Adams (2003a) used a less abrupt triangular O3 exposure profile at
 6     concentrations assumed to be typical of outdoor ambient conditions (beginning at 0.03 ppm,
 7     increasing steadily to 0.15 ppm in the fourth hour and decreasing steadily to 0.05 ppm at 6.6 h
 8     (mean = 0.08 ppm). Postexposure values for FEVj and symptoms were not significantly
 9     different between the 6.6 h triangular and a square-wave 0.08 ppm O3 exposure. During the
10     triangular exposure, however, FEVj responses became statistically significant after 4.6 h,
11     whereas, they were not significant until 6.6 h during the square wave exposure (Adams, 2003a).
12     Perhaps due to the lower O3 concentrations, evidence of FEVj response recovery with the
13     triangular exposure was less pronounced than as observed by Hazucha et al. (1992).  Figure 6-1
14     illustrates the average O3-induced FEVj responses and the O3 exposure schemes for the Adams
15     (2003a) and Hazucha et al. (1992) studies.  For completeness, other studies have also used a
16     triangular exposure profile but for a shorter duration of only 130 minutes (Foster and
17     Stetkiewicz, 1996;  Foster et al., 1996).
18          With square-wave O3 exposures between 0.08 to 0.12 ppm, FEVj decrements may increase
19     with time of exposure (and O3 dose) or reach a plateau (Horstman et al., 1990; McDonnell et al.,
20     1991). For the triangular exposures used by Hazucha et al. (1992) and Adams (2003a),
21     maximal FEVj responses occurred 1 h to 2 h after peak O3 concentration and 1 h to 2 h before
22     the maximal O3 dose occurred (at the end of the O3 exposure).  These two studies suggest that
23     depending upon the profile of the exposure, the triangular exposure can potentially lead to
24     higher FEVj responses than  square wave exposures at overall equivalent ozone doses.
25
26     6.2.5   Mechanisms of Pulmonary Function Responses
27          Inhalation of O3 for several hours while physically  active elicits both subjective respiratory
28     tract symptoms and acute pathophysiologic changes.  The typical symptomatic response
29     consistently reported in studies is that of tracheobronchial airway irritation.  Depending on the
30     individual's responsiveness to O3, this is accompanied by several pathophysiologic changes such
31     as decrements in lung capacities and volumes, bronchoconstriction, airway hyperresponsiveness,

       August 2005                             6-8        DRAFT-DO NOT QUOTE OR CITE

-------
                   u
                   0)
                   UJ
                   u.
                   TJ
                   
-------
 1           Available information on recovery from O3 exposure indicates that an initial phase of
 2      recovery proceeds relatively rapidly, and some 40 to 65% of the acute spirometric and symptom
 3      response appears to occur within about 2 h (Folinsbee and Hazucha, 1989). Following a 2 h
 4      exposure to 0.4 ppm O3 with IE, Nightingale et al. (2000) observed a 13.5% decrement in FEVj.
 5      By 3 h postexposure, however, only a 2.7% FEVj decrement persisted as illustrated in
 6      Figure 6-2.  A similar postexposure recovery in FVC was also observed. Gerrity et al. (1993)
 7      suggested that for healthy young adults transient increases in mucus clearance (mediated by
 8      cholinergic receptors) due to O3 exposure may be coincident to pulmonary function responses,
 9      i.e., the transient increases in clearance and decrements in lung function return to baseline values
10      within 2 to 3 h postexposure.  However, there is some indication that the spirometric responses,
11      especially at higher O3 concentrations, are not fully recovered within 24 h (Folinsbee and
12      Horvath, 1986; Folinsbee et al., 1998). In hyperresponsive individuals, the recovery takes
13      longer,  as much as 48 hours, to return to baseline values.  Collectively, these observations
14      suggest that there is a rapid recovery of O3-induced spirometric responses and symptoms, which
15      may occur during resting exposure to O3 (Folinsbee et al., 1977) or as O3 concentration is
16      reduced during exposure (Hazucha et al., 1992), and a slower phase, which may take at least
17      24 h to  complete (Folinsbee and Hazucha, 2000). Repeated exposure studies at higher
18      concentrations typically show that FEVj response to O3 is enhanced on the second of several
19      days of exposure (Table AX6-8). This enhanced response suggests a residual effect of the
20      previous exposure, about 22 h earlier, even though the preexposure spirometry may be the
21      same as on the previous day. The absence of the enhanced response with repeated exposure at
22      lower O3 concentrations may be the result of a more complete recovery or less damage to
23      pulmonary tissues (Folinsbee et al., 1994).
24
25      6.2.5.1   Pathophysiologic Mechanisms
26      Breathing pattern changes
27           Human studies consistently report that inhalation of O3 alters the breathing pattern without
28      significantly affecting minute ventilation.  A progressive decrease in tidal volume and a
29      "compensatory" increase in frequency of breathing to maintain steady minute ventilation during
30      exposure suggests a direct modulation of ventilatory control. These changes parallel a response
31      of many animal  species exposed to O3 and other lower airway irritants (Tepper et al., 1990).

        August 2005                              6-10       DRAFT-DO NOT QUOTE OR CITE

-------
                              4.0-,
                          LLI
                              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.
                   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).
 1     Bronchial C-fibers and rapidly adapting receptors appear to be the primary vagal afferents
 2     responsible for O3-induced changes in ventilatory rate and depth in both humans (Folinsbee and
 3     Hazucha, 2000) and animals (Coleridge et al., 1993; Hazucha and Sant'Ambrogio, 1993;
 4     Schelegle et al., 1993).
 5          The potential modulation of breathing pattern by activation of sensory afferents located in
 6     extrathoracic airways by O3 has not yet been studied in humans. Nasal only O3 exposure of rats
 7     produces changes in breathing pattern that are similar to changes observed in humans (Kleinman
 8     etal., 1999).
 9
10     Symptoms and lung function changes
11          As discussed, in addition to changes in ventilatory control, O3 inhalation by humans will
12     also induce a variety of symptoms, reduce vital capacity (VC) and related functional measures,
13     and increase airway  resistance.
14          Schelegle et al. (2001) demonstrated that the reduction in VC due to O3 exposure is a reflex
15     action and not a voluntary termination of inspiration as result of discomfort.  They reported
       August 2005
6-11
DRAFT-DO NOT QUOTE OR CITE

-------
 1      that O3-induced symptom responses (mediated in part by bronchial C-fibers) are substantially
 2      reduced by inhaled topical anesthetic. However, the anesthetic had a minor and irregular effect
 3      on pulmonary function decrements and tachypnea.  Since respiratory symptom responses were
 4      largely abolished, these findings support reflex inhibition of VC due to stimulation of both
 5      bronchial and pulmonary C-fibers.
 6           The involvement of nociceptive bronchial C-fibers modulated by opioid receptors
 7      in limiting maximal inspiration and eliciting subjective symptoms in humans was studied
 8      by Passannante et al. (1998). Sufentanil (an opioid agonist and analgesic) rapidly
 9      reversed O3-induced symptom responses and reduced spirometric decrements in "strong"
10      responders.  The incomplete recovery in FEVj following sufentanil administration, however,
11      suggests involvement of non-opioid receptor modulated mechanisms as well.  Interestingly,
12      naloxone (opioid receptor antagonist) had no significant effect on FEVj decrements in "weak"
13      responders.  Plasma levels of p-endorphin (a potent pain suppressor) were not related with O3
14      responses.
15
16      Airway hyperreactivity
17           In addition to limitation of maximal inspiration and its effects on other spirometric
18      endpoints, activation of airway sensory afferents also plays a role in receptor-mediated
19      bronchoconstriction and an increase in airway  resistance.  Despite this common mechanism,
20      post-O3 pulmonary function changes and either early or late bronchial hyperresponsiveness
21      (BHR) to inhaled aerosolized methacholine or histamine are poorly correlated either in time or
22      magnitude.  Fentanyl and indomethacin, the drugs that have been shown to attenuate O3-induced
23      lung function decrements in humans, did not prevent induction of BHR when administered to
24      guinea pigs  prior to O3 exposure (Yeadon et al., 1992). Neither does post-O3 BHR seem to be
25      related to airway baseline reactivity. These findings imply that the mechanisms are either not
26      related or are activated independently in time.  Animal studies (with limited support from human
27      studies) have suggested that an early post-O3 BHR is, at least in part, vagally mediated (Freed,
28      1996) and that stimulation of C-fibers can lead to increased responsiveness of bronchial smooth
29      muscle independently of systemic and inflammatory changes which may be even absent (load
30      et al.,  1996). In vitro study of isolated human bronchi have reported that O3-induced airway
31      sensitization involves changes in smooth muscle excitation-contraction coupling (Marthan,

        August 2005                              6-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1      1996). Characteristic O3-induced inflammatory airway neutrophilia which at one time was
 2      considered a leading BHR mechanism, has been found in a murine model, to be only
 3      coincidentally associated with BHR, i.e., there was no cause and effect relationship (Zhang et al.,
 4      1995). However, this observation does not rule out involvement of other cells such as
 5      eosinophils or T-helper cells in BHR modulation.  There is some evidence that release of
 6      inflammatory mediators by these cells can sustain BHR and bronchoconstriction. In vitro and
 7      animal studies have also suggested that airway neutral endopeptidase activity can be a strong
 8      modulator of BHR (Marthan et al., 1996; Yeadon  et al., 1992).  Late BHR observed in some
 9      studies is plausibly due to a sustained damage of the airway epithelium and continual release of
10      inflammatory mediators (Foster et al., 2000).  Thus, O3-induced BHR appears to be a product of
11      many mechanisms acting at different time  periods and levels of the bronchial smooth muscle
12      signaling pathways (The effects ofO3 on BHR are described in Section 6.8}.
13
14      6.2.5.2    Mechanisms at a Cellular and Molecular Level
15           Stimulation of vagal  afferents by O3  and reactive products, the primary mechanism of lung
16      function impairment, is enhanced and sustained by what can be considered in this context to be
17      secondary mechanisms activated at a cellular and molecular level.  The complexity of these
18      mechanisms is beyond the scope of this section  and the reader is directed to Section 6.9 of this
19      chapter for greater detail. A comprehensive review by Mudway and Kelly (2000) discusses the
20      cellular and molecular mechanisms of O3-induced pulmonary response in great detail.
21           Stimulation of bronchial C-fibers by  O3 not only inhibits maximal inspiration but, through
22      local axon reflexes, induces neurogenic inflammation. This pathophysiologic process is
23      characterized by release of tachykinins and other proinflammatory neuropeptides.  Ozone
24      exposure has been shown to elevate the C-fiber-associated tachykinin, substance P, in human
25      bronchial lavage fluid (Hazbun et al. 1993) and to deplete neuropeptides synthesized and
26      released  from C-fibers in human airway epithelium rich in substance P-immunoreactive axons.
27      Substance P and other transmitters are known to induce granulocyte adhesion and subsequent
28      transposition into the airways, increase vascular permeability and plasma protein extravasation,
29      cause bronchoconstriction, and promote mucus  secretion (Solway and Leff, 1991). Although the
30      initial pathways of neurogenic, antigen-induced, and innate immune-mediated inflammation are
31      not the same, they eventually converge leading to further amplification of airway inflammatory

        August 2005                              6-13       DRAFT-DO NOT QUOTE OR CITE

-------
 1     processes by subsequent release of cytokines, eicosanoids, and other mediators.  Significantly
 2     negative correlations between O3-induced leukotriene (LTC4/D4/E4) production and spirometric
 3     decrements (Hazucha et al., 1996), and an increased level of postexposure PGE2, a mediator
 4     known to stimulate bronchial C-fibers, show that these mediators play an important role in
 5     attenuation of lung function due to O3 exposure (Mohammed et al.,  1993; Hazucha et al., 1996).
 6     Moreover, because the density of bronchial C-fibers is much lower in the small than large
 7     airways, the reported post O3 dysfunction of small airways assessed by decrement in FEF25.75
 8     (Weinman et al., 1995; Frank et al., 2001) may be due in part to inflammation.  Also, because of
 9     the relative slowness of inflammatory responses as compared to reflex effects, O3-triggered
10     inflammatory mechanisms are unlikely to initially contribute to progressive lung function
11     reduction. It is plausible, however, that when fully activated, they sustain and possibly further
12     aggravate already impaired lung function. Indeed, a prolonged recovery of residual spirometric
13     decrements following the initial rapid improvement after exposure termination could be due to
14     slowly resolving airway inflammation. Bronchial biopsies performed 6 h postexposure have
15     shown that O3 caused a significant decrease in immunoreactivity to  substance P in the
16     submucosa (Krishna et al., 1997). A strong negative correlation with FEVj also suggests that the
17     release of substance P may be a contributing mechanism to persistent post-O3
18     bronchoconstriction (Krishna et al., 1997).  Persistent spirometry changes observed for up to
19     48 h postexposure could plausibly be sustained by the inflammatory mediators, many of which
20     have bronchoconstrictive properties (Blomberg et al., 1999).
21
22
23     6.3  SUBJECTS WITH PREEXISTING DISEASE
24          Individuals with respiratory disease are of primary concern in  evaluating the health effects
25     of O3 because even a small change in function is likely to have more impact on a person with
26     reduced reserve, i.e., O3-induced effects are superimposed on preexisting pulmonary impairment.
27
28     6.3.1    Subjects with Chronic  Obstructive Pulmonary Disease
29          For patients with COPD performing light to moderate IE, no decrements in pulmonary
30     function were observed after 1- and 2-h exposures to <0.30 ppm O3  (Kehrl et al., 1985; Linn
31     et al., 1982a, 1983a; Solic et al., 1982), and only small decreases in  forced expiratory volume

       August 2005                             6-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1     were observed for 3-h exposures of chronic bronchitics to 0.41 ppm O3 (Kulle et al., 1984).
 2     More recently, Gong et al. (1997a) found no significant difference in response between age-
 3     matched controls and COPD patients to a 4 h exposure to 0.24 ppm O3 with IE. Although the
 4     clinical significance is uncertain, small transient decreases in arterial blood oxygen saturation
 5     have also been observed in some of these studies.
 6
 7     6.3.2    Subjects with Asthma
 8          Based on studies reviewed in the 1996 criteria document (U.S. Environmental Protection
 9     Agency, 1996), asthmatic subjects appear to be at least as sensitive to acute effects of O3 as
10     healthy nonasthmatic subjects.
11          Several recent studies support a tendency for slightly increased spirometric responses in
12     mild asthmatic versus healthy subjects. Alexis et al. (2000) reported reductions in FVC (12%,
13     10%) and FEVj (13%, 11%) for 13 mild asthmatic and 9 healthy subjects, respectively, exposed
14     to 0.4 ppm O3 for 2 h with IE ( VE = 30 L/min).  The FVC and FEVj responses were attenuated
15     by indomethacin in the healthy  subjects but not the asthmatics. As assessed by the magnitude of
16     reductions in mid-flows (viz. FEF25, FEF50, FEF60p, FEF75) following O3 exposure, the small
17     airways tended to be more affected in asthmatics than healthy subjects. In a larger study, Torres
18     et al. (1996) exposed 24 asthmatics, 12 allergic rhinitics, and 10 healthy subjects  to 0.25 ppm O3
19     for 3 h with IE. The O3-induced FEVj decrements tended to be greater in the diseased
20     populations (allergic rhinitis, 14.1%; asthmatic,  12.5%; healthy controls,  10.2%).  Scannell et al.
21     (1996) exposed 18  asthmatics to 0.2 ppm O3 for 4 h with IE (VE « 25 L/min/m2 BSA).
22     An O3-induced increase in sRaw tended to  be greater in the asthmatics compared to 81 healthy
23     subjects who underwent similar experimental protocols (Aris et al., 1995; Balmes et al., 1996).
24          Increased sensitivity of asthmatics to O3 was also demonstrated in the epidemiological
25     study by Hoppe et al. (2003). Relevant pulmonary function responses (>10% drop in FEVl5
26     FVC, or PEF, and/or >20%  increase in sRaw) subsequent to O3 exposure were experienced
27     by 22 of 43 young asthmatics (mean age, 15 yrs) versus only 6 of 43 young athletes (mean
28     age, 18 yrs).  Participants were  asked to engage in their normal activities  for 2 h in the
29     afternoon (61-62 ppb O3, on average) prior to pulmonary function testing. The estimated
30     activity level during O3 exposures was lower in the asthmatics (VE « 25 L/min) than the
31     athletes (VE « 80 L/min). As discussed in  Sections 6.2.2 and 6.2.3.1, responses to O3 increase

       August 2005                             6-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1     with VE. Hence, in the absence of some underlying susceptibility to adverse O3 effects, the
 2     asthmatics would actually be expected to respond far less than the athletes who had a 3.2-fold
 3     greater VE.
 4           Similar O3-induced spirometric responses are suggested by some studies. The Scannell
 5     et al.  (1996) study of 18 asthmatics reported FEVj and FVC decrements that were similar to 81
 6     healthy subjects (Aris et al., 1995; Balmes et al., 1996).  Similar group decrements in FEVj and
 7     FVC  were reported by Hiltermann et al. (1995), who exposed 6 asthmatics and 6 healthy
 8     subjects to 0.4 ppm O3 for 2 h with light IE. Basha et al. (1994) also reported similar spirometric
 9     responses between 5 asthmatic and 5 healthy subjects exposed to 0.2 ppm O3 for 6 h with IE.
10     The lack of significant differences in the Hiltermann et al. (1995) and Basha et al. (1994) studies
11     is not compelling given the extremely small sample sizes and corresponding lack of statistical
12     power. The Basha et al. (1994) study was also confounded by the asthmatics having an average
13     preexposure FEVj that was about 430 mL lower (a 12% difference) on the O3-day relative to the
14     air-day. Hence, only the Scannell et al. (1996) study supports similar O3-induced spirometric
15     responses in asthmatics versus healthy subjects.
16           One study has reported that asthmatics tend to have smaller O3-induced FEVj decrements
17     relative healthy subjects (3% versus 8%, respectively) when exposed to 0.2 ppm  O3 for 2 h with
18     IE (Mudway et al., 2001). However, the asthmatics in the Mudway et al. (2001)  study also
19     tended to be older than the healthy subjects, which could partially explain their lesser response.
20           In a longer exposure duration (7.6 h) study, Horstman et al. (1995) exposed 17 mild-to-
21     moderate asthmatics and 13 healthy controls to 0.16 ppm O3 or FA with quasi continuous
22     exercise (VE «30 L/min). The FEVj decrement observed in the asthmatics was significantly
23     greater than in the healthy subjects (19% versus 10%, respectively).  There was also tendency for
24     a greater O3-induced decrease in FEF25.75 in asthmatics relative to the healthy subjects (24%
25     versus 15%, respectively). A significant positive correlation in asthmatics was also reported
26     between O3-induced spirometric responses and baseline lung function, i.e., responses increased
27     with severity of disease.
28           With repeated O3 exposures asthmatics, like healthy  subjects (see Section 6.6\
29     develop tolerance. Gong et al. (1997b) exposed 10 asthmatics to 0.4 ppm O3, 3 h per day with
30     IE (VE -32 L/min), for 5 consecutive days. Symptom and spirometric responses were greatest
31     on the first (-35 % FEVj) and second (-34 % FEVj) exposure days, and progressively

       August 2005                              6-16       DRAFT-DO NOT QUOTE OR CITE

-------
 1      diminished toward baseline levels (-6 % FEVj) by the fifth exposure day.  Similar to healthy
 2      subjects, asthmatics lost their tolerance 4 and 7 days later.
 3           Some, but not all, studies have reported that asthmatics have a somewhat exaggerated
 4      inflammatory response to acute O3 exposure relative to healthy controls (e.g., McBride et al.,
 5      1994; Basha et al., 1994; Peden et al.,  1995, 1997; Peden, 2001a; Scannell et al., 1996;
 6      Hiltermann et al., 1997, 1999; Michelson et al., 1999; Vagaggini et al., 1999; Newson et al.,
 7      2000; Holz et al., 2002) also (see Section 6.9 and Tables AX6-3 and-12}. For example, at 18-h
 8      post-O3 exposure (0.2 ppm, 4-h with IE) and corrected for FA responses, Scannell et al. (1996)
 9      found significantly increased neutrophils in 18 asthmatics (12%) compared to 20 healthy
10      subjects (4.5%). This inflammatory response difference was observed despite no group
11      differences in spirometric responses to O3.  Inflammatory responses do not  appear to be
12      correlated with lung function responses in either asthmatic or healthy subjects (Balmes et al.,
13      1996, 1997; Holz et al.,  1999). This lack of correlations between inflammatory and spirometric
14      responses may be due to differences in the time kinetics of these responses  (Stenfors et al.,
15      2002).  In addition, airway responsiveness to inhaled allergens is increased by O3 exposure in
16      subjects with allergic asthma for up to 24 h (see Section 6.8).
17
18      6.3.3   Subjects with Allergic Rhinitis
19           Allergic rhinitis is a condition defined by inflammation of the nasal membranes. Nayak
20      (2003) recently reviewed the commonalities between asthma and allergic rhinitis. Clinically,
21      greater than 60% of asthmatics have allergic rhinitis and slightly  less than 40%  of allergic
22      rhinitics have asthma. Leukotrienes and histamine are well-recognized mediators of responses
23      (viz., inflammation, hyperresponsiveness, and bronchoconstriction) in both asthma and allergic
24      rhinitis.  Although, rhinitis and asthma are distinguished as affecting the upper and lower
25      airways, respectively, it has been suggested that these diseases are manifestations of the same
26      disease entity.
27           Given the prevalence of concomitant asthma and rhinitis and their common response
28      mediators, it should be expected that allergic rhinitics might respond more similarly to
29      asthmatics than healthy individuals.  Regarding spirometric responses, Torres et al. (1996)
30      provide the only data demonstrating a trend in support of this supposition.
        August 2005                              6-17        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Studies demonstrating the interaction between air pollutants and allergic processes in the
 2      human nasal airways and rhinoconjunctival tissue have been reviewed by Peden (200Ib) and
 3      Riediker et al. (2001), respectively. Ozone exposure of subjects with allergic rhinitis has been
 4      shown to induce nasal inflammation and increase airway responsiveness to nonspecific
 5      bronchoconstrictors.
 6           Peden et al. (1995), who studied allergic asthmatics exposed to O3  found that O3 causes an
 7      increased response to nasal allergen challenge in addition to nasal inflammatory responses.
 8      Their data suggested that allergic subjects have an increased immediate response to allergen
 9      after O3  exposure. In a follow-up study, Michelson et al. (1999) reported that 0.4 ppm O3 did not
10      promote early-phase-response mediator release or enhance the response to allergen challenge in
11      the nasal airways of mild, asymptomatic dust mite-sensitive asthmatic subjects. Ozone did,
12      however, promote an inflammatory cell influx, which helps induce a more significant late-phase
13      response in this population.
14           Torres et al. (1996) found that O3 causes an increased response to bronchial allergen
15      challenge in  subjects with allergic rhinitis.  This study also measured  responses in healthy
16      subjects and  mildly allergic asthmatics {see Sections AX6.3.2 andAX6.8). All subjects were
17      exposed to 0.25 ppm O3 for 3 h with IE. Statistically significant O3-induced decrements in FEVj
18      occurred in rhinitics (14.1%), asthmatics (12.5%), and the healthy controls (10.2%), but these
19      responses did not differ statistically between groups.  Methacholine responsiveness was
20      significantly increased in asthmatics, but not in subjects with allergic  rhinitis.  Airway
21      responsiveness to an individual's historical allergen (either grass and  birch pollen, house dust
22      mite, or  animal dander) was significantly increased after O3 exposure when compared to FA
23      exposure.  The authors concluded that subjects with allergic rhinitis, but without asthma, could
24      be at risk if a high O3 exposure is followed by a high dose of allergen.
25           Holz et al. (2002) extended the results of Torres et al. (1996) by  demonstrating that
26      repeated daily exposure to lower concentrations of O3 (0.125 ppm for 4 days) causes an
27      increased response to bronchial allergen challenge in subjects with preexisting allergic airway
28      disease,  with or without asthma.  These investigators observed no major difference in the pattern
29      of bronchial  allergen response between asthmatics or rhinitics, except for a 10-fold increase in
30      the dose of allergen required to elicit a similar response (>20% decrease in FEVj) in the
31      asthmatic subjects. Early phase responses were more consistent in subjects with rhinitis and

        August 2005                               6-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1      late-phase responses were more pronounced in subjects with asthma. There also was a tendency
 2      towards a greater effect of O3 in subjects with greater baseline response to specific allergens
 3      (chosen on the basis of skin prick test and history, viz., grass, rye, birch, or alder pollen, house
 4      dust mite, or animal dander). These data suggest that the presence of allergic bronchial
 5      sensitization, but not a history of asthma, may be a key determinant  of increased airway allergen
 6      responsiveness following exposure to O3 (for a more complete discussion of airway
 1      responsiveness)  see Section AX6.8.
 8
 9      6.3.4   Subjects with Cardiovascular Disease
10           Possibly due to the age of subjects studied, O3 exposure does not appear to result in
11      significant pulmonary function impairment or evidence of cardiovascular strain in patients with
12      cardiovascular disease relative to healthy controls. Gong et al. (1998) exposed 10 hypertensive
13      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
14      all subjects combined (no significant group differences), there was an O3-induced decrement of
15      7% in FEVj and an 70% increase in the alveolar-arterial oxygen tension gradient. The overall
16      results did not indicate any major acute cardiovascular effects of O3  in either the hypertensive or
17      normal subjects. Gong et al. (1998) suggested that by impairing alveolar-arterial oxygen
18      transfer, the O3 exposure could potentially lead to adverse cardiac events by decreasing oxygen
19      supply to the myocardium. However, the subjects in their study had sufficient functional reserve
20      so as to not experience significant  ECG changes or myocardial ischemia and/or injury (see
21      Section 6.10 for  additional discussion).
22
23
24      6.4  INTERSUBJECT VARIABILITY AND REPRODUCIBILITY
25           OF RESPONSE
26           Analysis of factors that contribute to intersubj ect variability is important for the
27      understanding of individual responses, mechanisms of response, and health risks associated with
28      acute O3 exposures. A large intersubj ect variability in response to O3 has been reported by
29      numerous investigators (Adams et al., 1981; Aris et al., 1995; Folinsbee et al., 1978; Kulle et al.,
30      1985; McDonnell et al., 1983).  The magnitude  of individual variability in FEVj response in 2 h
31      IE exposures increases at higher O3 concentrations (Kulle et al., 1985; McDonnell et al., 1983).

        August 2005                              6-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1     McDonnell (1996) examined the FEVj response data from three 6.6-h exposure studies
 2     conducted at the EPA Health Effects Research Laboratory and showed that the FEVj responses
 3     in FA were small with most tightly grouped around zero. With increasing O3 concentrations
 4     between 0.08 and 0.12 ppm, the mean response became asymmetrical with a few individuals
 5     experiencing quite large decrements in FEVj (Intersubject variability observed in O3 dosimetry
 6     studies is discussed in Chapter 4.2).
 1           As an example of the variation in spirometric responses to O3 exposure, Hazucha et al.
 8     (2003) analyzed the distribution of O3 responsiveness in 240 subjects (18 to 60 years of age)
 9     exposed to 0.42 ppm  O3 (on 3 occasions) for 1.5 h with IE at VE = 20 L/min/m2 BSA. Across
10     all ages, 18% of subjects were weak responders (<5% FEVj decrement), 39% were moderate
11     responders, and 43%  were strong responders (> 15% FEVj decrement). Younger subjects
12     (<35 years of age) were predominately strong responders, whereas, older subjects (>35 years of
13     age) were mainly weak responders.  The influence of age on intersubject variability was also
14     noted by Passannante et al. (1998) who found that subjects under 35  years of age were more like
15     to be strong responders than older individuals. In contrast to these clinical studies, Hoppe et al.
16     (2003) observed relevant pulmonary function responses (>10% drop in FEVb FVC, or PEF,
17     and/or >20% increase in sRaw) subsequent to O3 exposure in 27% of elderly adults (n = 41;
18     mean age, 81 yrs; VE «  10 L/min) versus only 14% of young athletes (n = 43; mean age,
19     18 yrs; VE « 80 L/min).  In the absence of some underlying susceptibility to adverse O3 effects,
20     the elderly adults would be expected to respond far less than the athletes who had an estimated
21     8-fold greater VE.
22           For repeated exposures, Hazucha et al. (2003) reported that the reproducibility of FEVj
23     responses was related to the length of time between exposures. The  Spearman correlation
24     coefficient of 0.54 was found between responses for exposures separated by  105 days (median),
25     whereas, a correlation coefficient of 0.85 was found between responses for exposures separated
26     by only 7  days (median). The more reproducible the subject's response, the  more precisely it
27     indicates his/her intrinsic responsiveness. In 2 h IE O3 exposures, McDonnell et al. (1985b)
28     found a relatively poor FEVj reproducibility (R = 0.58) at the lowest concentration, 0.12 ppm,
29     due, in part, to a lack of specific O3 response or a uniformly small response in the majority of
30     subjects. It was concluded that for 2 h IE O3 exposures equal to or greater than 0.18 ppm, the
31     intersubject differences in magnitude of change in FVC and FEVj are quite reproducible over

       August 2005                              6-20       DRAFT-DO NOT QUOTE OR CITE

-------
 1     time (21 to 385 days; mean = 33 days) and are due primarily to differences in intrinsic
 2     responsiveness of individual subjects to O3 exposure.
 3           Intersubject variability, mechanisms of response, and health risks associated with acute O3
 4     exposures are complicated by a poor association between various O3-induced responses.  In a
 5     retrospective study of 485 male subjects (ages 18 to 36 yrs) exposed to one of six O3
 6     concentrations at one of three activity levels for 2 h, McDonnell et al. (1999) observed
 7     significant, but low, Spearman rank order correlations between FEVj response and symptoms of
 8     cough (R = 0.39), shortness of breath (R = 0.41), and pain on deep inspiration (R = 0.30). These
 9     authors concluded that these responses  are related mechanistically to some degree, but indicated
10     that there does not appear to be a single factor which is responsible for the observed individual
11     differences in O3 responsiveness across the spectrum of symptom and lung function responses.
12           The effect of large intersubject variability on the ability to predict individual
13     responsiveness to O3 was demonstrated by McDonnell et al. (1993). These investigators
14     analyzed the data of 290 male subjects (18 to 32 years of age) who underwent repeat 2 h IE
15     exposures to one or more O3 concentrations ranging from 0.12 to 0.40 ppm.  They attempted to
16     identify personal characteristics (i.e., age, height, baseline pulmonary function, presence of
17     allergies, and past smoking history) that might predict individual differences in FEVj response.
18     Only age contributed significantly to intersubject responsiveness (younger subjects were more
19     responsive), accounting for just 4% of the observed variance. Interestingly, O3 concentration
20     accounted for only 31% of the variance, strongly suggesting the importance of as yet undefined
21     individual characteristics that determine FEVj responsiveness to O3. The authors concluded that
22     much individual variability in FEVj response to O3 remains unexplained.
23
24
25     6.5  FACTORS MODIFYING RESPONSIVENESS TO OZONE
26     6.5.1  Influence of Age
27           Children, adolescents, and young  adults (<18 yrs of age) appear, on average, to have nearly
28     equivalent spirometric responses to O3, but have greater responses than middle-aged and older
29     adults when exposed to a comparable O3 doses (U.S. Environmental Protection Agency,  1996).
30     Symptomatic responses to O3 exposure, however, appear to increase with age until early
31     adulthood and then gradually decrease with increasing age (U.S. Environmental Protection

       August 2005                             6-21       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Agency, 1996).  In contrast to young adults, the diminished symptomatic responses in children
 2      and the elderly may put them at an increased risk for continued exposure. Although no new
 3      laboratory studies investigating O3 responses in children have been published since the last O3
 4      AQCD, the epidemiological studies published during the last decade (see section 7.2.3.1 for
 5      details) are generally in agreement with the earlier laboratory studies.
 6           The ensuing discussion in this section will provide information on average FEVj responses
 7      to O3 exposure as a function of age in healthy adults ranging from 18 to 50 years of age. As was
 8      specifically addressed in Section 6.4, however, there is considerable intersubject variability in
 9      responses and this between subject variability increases with  increasing O3 dose (see Figure
10      AX6-6).  Epidemiological studies also report such intersubject variability. For instant, Hoppe
11      et al. (2003) observed relevant pulmonary function responses (>10% drop in FEVl3 FVC,
12      or PEF, and/or >20% increase in sRaw) subsequent to  O3 exposure in 50% of healthy
13      children (n = 44; mean age, 7 yrs; VE « 30 L/min), 14% of young athletes (n = 43; mean age,
14      18 yrs; VE « 80 L/min), and 27% of elderly adults (n = 41; mean age, 81 yrs; VE « 10 L/min).
15           Beyond approximately 18 years of age, spirometric and symptom responses to O3 exposure
16      begin to decline with increasing age.  In healthy individuals, the rate of decline in O3
17      responsiveness appears to be greater in younger (18  to 35 yrs) versus middle aged (35 to 55 yrs)
18      individuals (Passannante et al., 1998; Hazucha et al., 2003). Beyond this age (>55 yrs), acute O3
19      exposure elicits minimal spirometric changes.  An average FEVj decrement of-3% has been
20      reported by Gong et al. (1997a) for this older population under a "worst case" exposure scenario
21      (0.24 ppm O3 with 4 h IE).  Although Gong et al. (1997a) and others have examined responses
22      to O3 exposure in subjects of various ages, the exposure conditions differ between most studies
23      so that age effects remain uncertain.
24           Three recent  studies, which analyzed large data sets (>240 subjects) of similarly exposed
25      subjects, show clearly discernable changes in FEVj responses to O3 as a function of age.
26      Seal et al. (1996) analyzed O3-induced spirometric responses  in 371 young nonsmokers
27      (18 to 35 years of age) exposed for 2.3 h during IE at a VE  of 25 L/min/m2 BSA. On average,
28      for the same O3 concentration (C), the response of 25,  30,  and 35 year old individuals are
29      predicted to be 83, 65, and 48%, respectively, of the response in a 20 year old.  For example,
30      a 5.4% decrement in FEVj is predicted for 20 year old exposed to 0.12 ppm O3 for 2.3 h
        August 2005                              6-22        DRAFT-DO NOT QUOTE OR CITE

-------
 1     IE (VE = 25 L/min/m2 BSA), whereas, a similarly exposed 35 yr old is predicted to have only a
 2     2.6% decrement.
 3           McDonnell et al. (1997) examined FEVj responses in 485 healthy white males (18 to
 4     36 years of age) exposed once for 2 h to an O3 concentration of 0.0, 0.12, 0.18, 0.24, 0.30, or
 5     0.40 ppm at rest or one of two levels of IE (VE of 25 and 35 L/min/m2 BSA). For the same
 6     exposure conditions (C, VE, and duration), the average responses of 25, 30, and 35 year old
 7     individuals are predicted to be 69, 48, and 33%, respectively, of the response in 20 year olds.
 8     Hazucha et al. (2003) analyzed the distribution of O3 responsiveness in 240 subjects (18 to
 9     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,
10     the FEVj responses of 25, 35, and 50 year olds are predicted to be 94, 83, and 50% ,
11     respectively, of the average response in 20 year old males. In females, the FEVj responses of 25,
12     35, and 50 year olds are predicted to be 82, 46, and 18%, respectively, of the average response in
13     20 year old females.
14           For subjects aged 18 to 36 yrs, McDonnell et al. (1999) recently reported that symptom
15     responses from O3 exposure also decrease with increasing age.  Whether the same age-dependent
16     pattern of O3 sensitivity decline also holds for airway reactivity or inflammatory endpoints has
17     not been determined.
18
19     6.5.2   Gender and Hormonal Influences
20           Several studies have suggested that physiological differences between the genders may
21     predispose  females to a greater susceptibility to O3.  Lower plasma and nasal lavage fluid levels
22     of uric acid (the most prevalent antioxidant) in females relative to males may be a contributing
23     factor (Housley et al., 1996). Consequently, reduced absorption of O3 in the upper airways may
24     promote its deeper penetration.  Dosimetric measurements have shown that the absorption
25     distribution of O3 is independent of gender when absorption is normalized to anatomical dead
26     space (Bush et al., 1996). More recently,  Ultman et al. (2004) reported that the whole lung
27     uptake fraction of O3 was significantly greater in males (91.4%) than females (87.1%).  But, this
28     increase in  O3 uptake in the males was consistent with their larger tidal volume and slower
29     breathing frequency relative to the females.  Furthermore, O3 uptake was not correlated with
30     spirometric responses. Thus, a differential removal of O3 by uric acid seems to have minimal


       August 2005                               6-23       DRAFT-DO NOT QUOTE OR CITE

-------
 1      effect. In general, the spirometric responses of young healthy females to O3 exposure appears
 2      comparable to the responses of young males (Hazucha et al., 2003). Although, during the
 3      follicular phase of the menstrual cycle, lung function response to O3 is enhanced (Fox et al.,
 4      1993).
 5
 6      6.5.3   Racial, Ethnic, and Socioeconomic Status Factors
 7           A few epidemiologic studies have implied that minorities are more responsive to O3 than
 8      Caucasians. However, this may be more of a consequence of the overall quality of health care
 9      and socioeconomic status (SES) than an innate sensitivity to oxidants (Gwynn and Thurston,
10      2001; Seal et al, 1996). The paucity of data has prevented making any definitive conclusions on
11      the influence of race, ethnic or other related factors on the responsiveness to O3.
12
13      6.5.4   Influence of Physical Activity
14           Any physical activity will increase minute ventilation and therefore the dose of inhaled O3.
15      Consequently, the intensity of physiological response following an acute exposure will be
16      strongly associated with minute ventilation (see Figures 6-3 andAX6-2).
17
18      6.5.5   Environmental Factors
19           Since the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996) few human
20      laboratory studies have examined the potential influence of environmental factors such as rural
21      versus urban  environment, passive cigarette smoke exposure, and bioactive agents such as
22      endotoxin on healthy individual's pulmonary function changes due to O3.
23           New controlled human exposure studies have confirmed that smokers are less responsive
24      to O3 than nonsmokers. Spirometric and plethysmographic  pulmonary function decline,
25      nonspecific airway hyperreactivity, and inflammatory response of smokers to O3 were all weaker
26      than those reported for nonsmokers. Although all these responses are intrinsically related, the
27      functional association between them, as in nonsmokers, has been weak.  Similarly, the time
28      course of development and recovery of these effects, as well their reproducibility, was not
29      different from nonsmokers. Chronic airway inflammation with desensitization of bronchial
30      nerve endings and an increased production of mucus may plausibly explain the pseudo-
31      protective effect of smoking (Frampton et al.,  1997; Torres  et al., 1997).

        August 2005                               6-24       DRAFT-DO NOT QUOTE OR CITE

-------
                                                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 young healthy 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).
 1          The effect of environmental tobacco smoke (ETS) on O3 responses has received very little
 2     attention. In one study, preexposure of mice to sidestream cigarette smoke (ETS surrogate)
 3     elicited no immediate effects, but potentiated subsequent O3-induced inflammatory responses
 4     (Yu et al., 2002) (See Chapter 5.4.2 for additional ETS details). Endotoxin is a biologically
 5     active component of both mainstream and sidestream tobacco smoke (Hasday et al., 1999) which
 6     might contribute to the potentiation of O3 effects.
 7          The influence of ambient temperature on pulmonary effects induced by O3 exposure in
 8     humans has been studied infrequently under controlled laboratory conditions. Several
 9     experimental human studies have reported additive effects of heat and O3 exposure (see U.S.
10     Environmental Protection Agency, 1986, 1996). Foster et al. (2000) exposed 9 young healthy
11     subjects for 130 min (IE 10 min at 36 to 39 1/min) to filtered air and to ramp profile O3 at 22 °C
12     and 30 °C, 45-55% RH. The O3 exposure started at 0.12 ppm, reached the peak of 0.24 ppm
13     midway through and subsequently declined to 0.12 ppm at the end of exposure.  At the end of
14     exposure FEVj decreased significantly (p < 0.5) by -8% at 22 °C and -6.5% at 30 °C. One day
       August 2005
6-25
DRAFT-DO NOT QUOTE OR CITE

-------
 1      (19 h) later, the decline of 2.3% from baseline was still significant (p < 0.05) at both
 2      temperatures.  FVC decrements were smaller and significant only for the 22 °C condition
 3      immediately postexposure.  There was a decline in specific airway conductance (sGaw; p < 0.05)
 4      at 30 °C but not at 22 °C. The nonspecific bronchial responsiveness to methacloline assessed
 5      as PC50 sGaw was significantly (p < 0.05) higher one day following O3 exposure at both
 6      temperatures but more so at 30 °C.  Thus, these findings suggest that elevated temperature may
 7      partially attenuate spirometric responses but enhance airway reactivity.
 8
 9      6.5.6  Oxidant-Antioxidant Balance
10           The first line of defense against oxidative stress is antioxidant present in epithelial lining
11      fluid (ELF) which scavenge free radicals and limit lipid peroxidation. Exposure to O3 depletes
12      the antioxidant level in nasal ELF probably due to scrubbing of O3 (Mudway et al., 1999),
13      however, the concentration and the activity of antioxidant enzymes either in ELF or plasma do
14      not appear to be related to O3 responsiveness (Avissar et al., 2000; Blomberg et al., 1999; Samet
15      et al., 2001). Carefully controlled studies of dietary antioxidant supplementation have
16      demonstrated some protective effects of a-tocopherol and ascorbate on spirometric lung function
17      from O3 but not on the intensity of subjective symptoms and inflammatory response including
18      cell recruitment, activation and release of mediators (Samet et al., 2001; Trenga et al., 2001).
19      Dietary antioxidants have also afforded partial protection to asthmatics by attenuating post-
20      exposure bronchial hyperresponsiveness (Trenga et al., 2001). The field studies performed in
21      Mexico City {described in Section 7.2.3.1) and animal studies (described in Section 5.2.1.3) have
22      also demonstrated the protective effects of ELF antioxidants during O3 exposures.
23
24      6.5.7  Genetic Factors
25           Several recent studies (Yang et al., 2005; David et al., 2003; Romieu et al., 2004) have
26      reported that genetic polymorphism of antioxidant enzymes and inflammatory genes may
27      modulate pulmonary function and inflammatory response to O3 challenge.  It appears that
28      healthy carriers of NAD(P)H:quinone oxidoreductase wild type (NQOlwt) in combination with
29      glutathione S-transferase |i-l (GSTMlnull) genotype are more responsive to O3.  The authors
30      have implied that the interindividual variability in O3 responsiveness (FEVj changes) is related
        August 2005                              6-26        DRAFT-DO NOT QUOTE OR CITE

-------
 1     to the polymorphism of these enzymes. Adults with GSTMlnull only genotype did not show O3
 2     hyperresponsiveness (Bergamaschi et al., 2001). A subsequent study from the same laboratory
 3     reported a positive association between O3 responsiveness, as characterized by the level of
 4     oxidative stress and inflammatory mediators (8-isoprostane, LTB4 and TEARS) in EEC fluid,
 5     and the antioxidant enzyme polymorphism. However, none of the spirometric lung function
 6     endpoints were affected by ozone exposure (Corradi et al., 2002). It is of interest to note, that
 7     human nasal mucosa biopsies of GSTM1 deficient subjects showed higher antioxidant enzymes
 8     activity than biopsies of GSTM1 positive individuals when exposed to ozone (Otto-Knapp et al.,
 9     2003).
10          Asthmatic children with a genetic deficiency of GSTM1 were reported to be more
11     responsive to ambient O3 exposure, as assessed by decrements in FEF25.75, in this field study.
12     Antioxidant supplementation (vit. C and E) attenuated post-ozone lung function response in
13     these children (Romieu et al., 2004).  More specific genotyping has shown that ozone
14     responsiveness of asthmatic children may be related to the presence of variant Ser allele for
15     NQO1. The presence of at least one NQO1 Ser allele in combination with GSTM1 null
16     genotype lowered the risk of asthma in ozone exposed asthmatic children relative to Pro/Pro
17     genotype (David et al., 2003).
18          The influence  of functional polymorphism in TNF-a, lymphotoxin- a (LTA), TLR4, SOD2
19     and GPX1 genes on ozone-induced lung function changes in healthy individuals, mild asthmatics
20     and subjects with rhinitis was varied. Of the inflammatory genes studied only TNF-a has
21     appeared to show some promise as one of the genetic factors of susceptibility. However,  as the
22     authors stated "the functional significance of individual TNF-a polymorphisms remains
23     controversial" (Yang et al., 2005).
24          These recent studies have shown that individual's innate susceptibility to ozone may be
25     linked to genetic background of an individual.  Although a number of potential ozone
26     susceptibility genes have been identified, additional better designed and controlled studies are
27     needed to ascertain the link between susceptibility and polymorphism.
28
29
30
       August 2005                             6-27        DRAFT-DO NOT QUOTE OR CITE

-------
 1      6.6  REPEATED O3 EXPOSURE EFFECTS
 2          Based on studies reviewed here and in the previous O3 criteria documents (U.S.
 3      Environmental Protection Agency, 1986,  1996), several conclusions can be drawn about
 4      repeated 1 to 2-h O3 exposures. Repeated exposures to O3 can cause an enhanced (i.e., greater)
 5      pulmonary function response on the second day of exposure (see Tables AX6-8 andAX6-9for
 6      added detail).  This enhancement appears to be dependent on the interval between the exposures
 7      (24 h is associated with the greatest increase) and is absent with intervals >3 days (Bedi et al.,
 8      1985; Folinsbee and Horvath, 1986; Schonfeld et al., 1989). An enhanced response also appears
 9      to depend to some extent on the magnitude of the initial response (Horvath et al.,  1981).  Small
10      responses to the first O3 exposure are less likely to result in an enhanced response on the second
11      day of O3 exposure (Folinsbee et al., 1994). With continued daily exposures (i.e., beyond the
12      second day) there is an attenuation of pulmonary function responses, typically after 3 to 5 days
13      of repeated exposure.  This attenuated response persists for less than 1 week (Kulle et al., 1982;
14      Linn et al., 1982b) or as long as 2 weeks (Horvath et al., 1981). In temporal conjunction with
15      pulmonary function changes, symptoms induced by O3, such as cough, pain on deep inspiration,
16      and chest discomfort, are increased on the second exposure day and attenuated with repeated
17      exposure thereafter (Folinsbee et al., 1980, 1998; Foxcroft and Adams, 1986; Linn et al., 1982b).
18      O3-induced changes in airway responsiveness persist longer and attenuate more slowly than
19      pulmonary function and symptoms responses (Dimeo et al., 1981; Kulle et al., 1982), although
20      this has been studied only on a limited basis (Folinsbee et al., 1994).  In longer-duration (4 h to
21      6.6 h), lower-concentration studies that do not cause an enhanced second-day response, the
22      attenuation of response to O3 appears to proceed more rapidly (Folinsbee et al., 1994)  [Effects of
23      repeated exposures on inflammatory responses are discussed in Section 6.9.4).
24
25
26      6.7  EFFECTS  ON EXERCISE PERFORMANCE
27          The effects of acute O3 inhalation on endurance exercise performance have been examined
28      in numerous controlled laboratory studies. These studies were discussed in the 1996 O3 AQCD
29      (U.S. Environmental Protection Agency, 1996) and  can be divided into two categories: (1) those
30      that examined the effects of acute O3 inhalation on maximal oxygen uptake (VO2max) and
       August 2005                             6-28        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (2) those that examined the effects of acute O3 inhalation on the ability to complete strenuous
 2      continuous exercise protocols of up to 1 h in duration.
 3           In brief, endurance exercise performance and VO2max may be limited by acute exposure
 4      to O3 (Adams and Schelegle, 1983; Schelegle and Adams, 1986; Gong et al.,  1986; Foxcroft and
 5      Adams, 1986; Folinsbee et al., 1977; Linder et al., 1988). Gong et al. (1986) and Schelegle and
 6      Adams (1986) found that significant reductions in maximal endurance exercise performance may
 7      occur in well-conditioned athletes while they perform CE (VE  > 80 L/min) for 1 h at O3
 8      concentrations >0.18 ppm. Reports from studies of exposure to O3 during high-intensity
 9      exercise indicate that breathing discomfort associated with maximal ventilation may be an
10      important factor in limiting exercise performance in some, but  not all, subjects.
11
12
13      6.8  EFFECTS ON AIRWAY RESPONSIVENESS
14           Airway or bronchial hyperresponsiveness (BHR) refers to a condition in which the
15      propensity for the airways to bronchoconstrict due to a variety  of stimuli becomes augmented.
16      Airway responsiveness is typically quantified by measuring the decrement in pulmonary
17      function (i.e., spirometry or plethysmography)  following the inhalation of small amounts of an
18      aerosolized bronchoconstrictor agent (specific [antigen, allergen] or nonspecific [methacholine,
19      histamine]) or a measured stimulus (e.g., exercise, cold air).
20           Ozone exposure causes an increase in nonspecific airway responsiveness as indicated by a
21      reduction in the concentration of methacholine or histamine required to produce a given
22      reduction in FEVj or increase in SRaw.  Increased airway responsiveness is an important
23      consequence of exposure to O3 because its presence means that the airways are predisposed to
24      narrowing on inhalation of a variety of stimuli (e.g., specific allergens, SO2, cold air).
25           Ozone exposure of asthmatic subjects, who characteristically have increased airway
26      responsiveness at baseline, can cause further increases in responsiveness (Kreit et al., 1989).
27      Similar relative changes in airway responsiveness are seen in asthmatics exposed to O3 despite
28      their markedly different baseline airway responsiveness. Several studies (Torres et al., 1996;
29      Kehrl et al., 1999; Molfino et al.,  1991) have been published suggesting an increase in specific
30      (i.e., allergen-induced) airway reactivity. An important aspect of increased airway
        August 2005                              6-29       DRAFT-DO NOT QUOTE OR CITE

-------
 1      responsiveness after O3 exposure is that this represents a plausible link between ambient O3
 2      exposure and increased hospital admissions for asthma.
 3           Changes in airway responsiveness after O3 exposure appear to be resolved more slowly
 4      than changes in FEVj or respiratory symptoms (Folinsbee and Hazucha, 2000).  Furthermore, in
 5      studies of repeated exposure to O3, changes in airway responsiveness tend to be somewhat less
 6      susceptible to attenuation with consecutive exposures than changes in FEVj (Dimeo et al., 1981;
 7      Folinsbee et al., 1994; Gong et al., 1997b; Kulle et al.,  1982). Increases in airway
 8      responsiveness do not appear to be strongly associated with decrements in lung function or
 9      increases in symptoms.
10           The mechanism of O3-induced increases in airway responsiveness is only partially
11      understood, but it appears to be associated with a number of cellular and biochemical changes
12      in airway tissue. Although inflammation could play a role in the increase in airway
13      responsiveness, cyclooxygenase inhibitors have not been effective at blocking the O3-induced
14      influx of PMNs into bronchoalveolar lavage (BAL) fluid (Hazucha et al., 1996;  Ying et al.,
15      1990). Therefore, O3-induced airway responsiveness may not be due to the presence of PMNs
16      in the airway or to the release  of arachidonic acid metabolites.  Rather, it seems  likely that the
17      mechanism for this response is multifactorial,  possibly involving the presence of cytokines,
18      prostanoids, or neuropeptides; activation of macrophages, eosinophils, or mast cells; and
19      epithelial damage that increases direct access of mediators to the smooth muscle or receptors in
20      the airways that are responsible for reflex bronchoconstriction.
21
22
23      6.9   EFFECTS ON INFLAMMATION AND HOST DEFENSE
24      6.9.1  Introduction
25           Short-term exposure of humans to O3 can cause acute inflammation and long-term
26      exposure of laboratory animals results in a chronic inflammatory state (see Chapter 5).  The
27      relationship between repetitive bouts of acute inflammation in humans caused by O3 and the
28      development of chronic respiratory disease is unknown.
29           The presence of neutrophils (PMNs) in the lung has long been  accepted as a hallmark of
30      inflammation and is an important indicator that O3 causes inflammation in the lungs. It is
31      apparent, however, that inflammation within airway tissues may persist beyond the point that

        August 2005                              6-30       DRAFT-DO NOT QUOTE OR CITE

-------
 1     inflammatory cells are found in BAL fluid (BALF). Soluble mediators of inflammation such as
 2     the cytokines (IL-6, IL-8) and arachidonic acid metabolites (e.g., PGE2, PGF2a, thromboxane,
 3     and leukotrienes [LTs] such as LTB4) have been measured in the BAL fluid of humans exposed
 4     to O3.  In addition to their role in inflammation, many of these compounds have
 5     bronchoconstrictive properties and may be involved in increased airway responsiveness
 6     following O3 exposure.
 7          Some recent evidence suggests that changes in small airways function may provide a
 8     sensitive indicator of O3 exposure and effect, despite the fact that inherent variability in their
 9     measurement by standard spirometric approaches make their assessment difficult (Frank et al.,
10     2001). Observations of increased functional responsiveness of these areas relative to the more
11     central airways, and of persistent effects following repeated exposure, may indicate that further
12     investigation of inflammatory processes in these regions is warranted.
13
14     6.9.2   Inflammatory Responses in the Upper Respiratory Tract
15          The nasal passages constitute the primary portal for inspired air at rest and, therefore,
16     the first region of the respiratory tract to come in contact with  airborne pollutants.  Nikasinovic
17     et al. (2003) recently reviewed the literature of laboratory-based nasal inflammatory studies
18     published since 1985.  Nasal lavage (NL) has provided a useful tool for assessing O3-induced
19     inflammation in the nasopharynx. Increased levels of PMNs in the NL fluid of humans exposed
20     to 0.5 ppm O3 at rest for 4 h has been reported (Graham et al.,  1988; Bascom et al., 1990).
21          Graham and Koren (1990) compared inflammatory mediators present in  both the NL and
22     BAL fluids of humans exposed to 0.4 ppm O3 for 2 h.  Similar increases in PMN were observed
23     in NL and BAL, suggesting a qualitative correlation between inflammatory changes in the lower
24     airways (BAL) and the upper respiratory tract (NL). Torres et al. (1997) compared NL and BAL
25     in smokers and nonsmokers exposed to 0.22 ppm O3 for 4 h. In contrast to Graham and Koren
26     (1990), they did not find a relationship between numbers or percentages of PMNs in the nose
27     and the lung, perhaps in part due to the variability observed in their NL recoveries. Albumin, a
28     marker of epithelial cell permeability, was increased 18 h later, but not immediately after
29     exposure, as seen by Bascom et al. (1990).
30          McBride et al. (1994) reported that asthmatic subjects were more sensitive than
31     nonasthmatics to upper airway inflammation at an O3 concentration (0.24  ppm for 1.5 h with

       August 2005                             6-31       DRAFT-DO NOT QUOTE OR CITE

-------
 1     light IE) that did not affect pulmonary function.  In the asthmatics, there was a significant
 2     increase in the number of PMNs in NL fluid both immediately and 24 h after exposure. Peden
 3     et al. (1995) also found that exposure to 0.4 ppm O3 had a direct nasal inflammatory effect and a
 4     priming effect on response to nasal allergen challenge.  A subsequent study in dust
 5     mite-sensitive asthmatic subjects indicated that O3 at this concentration enhanced eosinophil
 6     influx in response to allergen but did not promote early mediator release or enhance the nasal
 7     response to allergen (Michelson et al., 1999).  Similar to observations made in the lower airways,
 8     the presence of O3 molecular "targets" in nasal lining fluid is likely to provide some level of
 9     local protection against exposure. In a study of healthy subjects exposed to 0.2 ppm O3 for 2 h,
10     Mudway and colleagues (1999) observed a significant depletion of uric acid in NL fluid at 1.5 h
11     following exposure.
12
13     6.9.3   Inflammatory Response in the Lower Respiratory Tract
14          As reviewed in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996), acute
15     exposure to O3 results in an inflammatory reaction, increased epithelial cell permeability, and
16     may stimulate fibrogenic processes. Inflammatory markers are observed in BALF of healthy
17     subjects by 1 h post-O3 exposure and may persist for at least 18 to 24 h.  Not all inflammatory
18     markers, however, follow the same time course.  Studies published since the  1996 O3 AQCD
19     support these earlier findings.
20          Inflammatory effects have been assessed in vivo by lavage (proximal airway and
21     bronchoalveolar), bronchial biopsy, and more  recently, induced sputum. In vitro  studies of
22     human alveolar macrophages (AM) and airway epithelial cells exposed to O3 suggest that most
23     mediators found in the BALF of O3-exposed humans are produced by epithelial cells (U.S.
24     Environmental Protection Agency, 1996). Recent evidence suggests that the release of
25     mediators from AMs may be modulated by the products of O3-induced oxidation of airway
26     lining fluid components, such as human surfactant protein A (Wang et al., 2002).
27          Spirometric responses to O3 are independent from inflammatory responses and markers of
28     epithelial injury (Balmes et al., 1996; Blomberg et al., 1999; Hazucha et al., 1996; Torres et al.,
29     1997).  Significant inflammatory responses to  O3 exposures that did not elicit significant
30     spirometric responses have been reported (Holz et al., 2005; McBride et al., 1994).  A meta-
       August 2005                             6-32        DRAFT-DO NOT QUOTE OR CITE

-------
 1      analysis of 21 studies (Mudway and Kelly, 2004), showed that PMN influx in health subjects is
 2      associated with total O3 dose (product of O3 concentration, exposure duration, and VE).
 3           The time course of the inflammatory response to O3 in humans has not been fully
 4      characterized. From a review of the literature by Mudway and Kelly (2000), Figure 6-4
 5      illustrates a plausible time course of acute O3 responses. As the figure shows, different markers
 6      have peak responses at different times. Studies in which lavages were performed 1 h after O3
 7      exposure (1 h at 0.4 ppm or 4 h at 0.2 ppm) have demonstrated that the inflammatory responses
 8      are quickly initiated (Devlin et al., 1996; Schelegle et al., 1991; Torres et al., 1997).
 9      Inflammatory mediators and cytokines such as IL-8, IL-6, and PGE2 are greater at 1 h than at
10      18 h post-O3  exposure (Devlin et al., 1996; Torres et al., 1997). However, IL-8 still remains
11      elevated at 18 h post-O3 (4 h at 0.2 ppm O3 versus FA) in healthy subjects and more so in
12      asthmatics (Balmes et al.,  1996; Scannell et al., 1996).  Schelegle et al. (1991) found increased
13      PMNs in the  "proximal airway" lavage at 1, 6, and 24 h after O3 exposure (4 h at 0.2 ppm O3),
14      with a peak response at 6 h. Although, at 18 to 24 h after O3 exposure, PMNs remain elevated
15      relative to 1 h postexposure (Schelegle et al.,  1991; Torres et al., 1997).  In addition to the influx
16      of PMNs and (in allergic asthmatics) eosinophils, lymphocyte numbers in BALF are elevated
17      significantly  at 6 h following exposure (2 h at 0.2 ppm O3) of healthy subjects (Blomberg et al.,
18      1997). Flow cytometry also indicated the increased presence of CD3+, CD4+ and CD8+ T cell
19      subsets. This same laboratory later demonstrated that within 1.5 h following exposure of healthy
20      subjects to the same O3 regimen, expression of human leukocyte antigen (HLA)-DR on lavaged
21      macrophages underwent a significant, 2.5-fold increase (Blomberg et al.,  1999).
22           The inflammatory responses to O3 exposure have  also been studied  in asthmatic subjects
23      (Basha et al., 1994; Scannell et al., 1996; Peden et al., 1997). In these studies,  asthmatics
24      showed significantly more neutrophils in the BALF (18 h post-exposure) than similarly exposed
25      healthy individuals.  In one of these studies (Peden et al., 1997), which included only allergic
26      asthmatics who tested positive for Dematophagoides farinae antigen, there was an eosinophilic
27      inflammation (2-fold increase), as well as neutrophilic inflammation (3-fold increase).  In a
28      study of subjects with intermittent asthma exposed to 0.4 ppm O3 for 2 h, increases in eosinophil
29      cationic protein, neutrophil elastase and IL-8  were found to be significantly increased 16 h post-
30      exposure and comparable in induced sputum and BALF (Hiltermann et al, 1999).  Scannell et al.
31      (1996) also reported that IL-8 tends to be higher in the BALF of asthmatics compared to

        August 2005                               6-33        DRAFT-DO NOT QUOTE OR CITE

-------
Initiation    Progression
                                        Resolution
  B
  C
  D
                                                                     »•  FEVi

                                                                        Shed epithelial cells
                                                                        Albumin/total protein
                                                                        LDII
                                                                        UA
                                                                        AA
                                                                        GSH

                                                                        LOP
              Exposure
          Post-Exposure
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 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:  LOP (lipid ozonation products), GSH (reduced glutathione),
            AA (ascorbic acid), UA (uric acid), LDH (lactate dehydrogenase), PGE2
            (prostaglandin E2), LTB4 (leukotriene B4), LTc4 (leukotriene C4), MPO
            (myeloperoxidase).

Source: Reprinted from Mudway and Kelly (2000) with permission from Elsevier.
August 2005
                            6-34
DRAFT-DO NOT QUOTE OR CITE

-------
 1      nonasthmatics following O3 exposure, suggesting a possible mediator for the significantly
 2      increased neutrophilic inflammation in those subjects.  Bosson et al. (2003) found significantly
 3      greater the epithelial expression of IL-5, IL-8, granulocyte-macrophage colony-stimulating
 4      factor (GM-CSF) and epithelial cell-derived neutrophil-activating peptide 78 (ENA-78) in
 5      asthmatics compared to healthy subjects following exposure to 0.2 ppm O3 for 2 h.  Stenfors and
 6      colleagues (2002) were unable to detect a difference in the increased neutrophil numbers
 7      between 15 mild asthmatic and 15 healthy subjects by bronchial wash at the 6 h post-exposure
 8      time point.  However, the asthmatics were on average 5 years older than the healthy subjects in
 9      this study and it is not yet known how age affects inflammatory responses. It is also possible
10      that the time course of neutrophil influx differs between healthy and asthmatic individuals.
11           Vagaggini et al. (2002) investigated the effect of prior allergen challenge on responses in
12      mild asthmatics exposed for 2 h to 0.27 ppm O3 or filtered air.  At 6  h post-exposure, eosinophil
13      numbers in induced sputum were found to be significantly greater after O3 than after air.  Studies
14      such as these suggest that the time course of eosinophil and neutrophil influx following O3
15      exposure can occur to levels detectable within the airway lumen by as early as 6 h.  They also
16      suggest that the previous or concurrent activation of proinflammatory pathways within the
17      airway epithelium may enhance the inflammatory effects of O3. For example, in an in vitro
18      study of epithelial cells from the upper and lower respiratory tract, cytokine production induced
19      by rhinovirus infection was enhanced synergistically by concurrent exposure to O3 at 0.2 ppm for
20      3 h (Spannhake et al, 2002).
21           Although the release of mediators has been demonstrated to occur at exposure
22      concentrations and times that are minimally cytotoxic to airway cells, potentially detrimental
23      latent effects have been demonstrated in the absence of cytotoxicity. These include the
24      generation of DNA single strand breaks (Kozumbo et al., 1996), the loss of cellular replicative
25      activity (Gabrielson et al., 1994) in bronchial epithelial cells exposed in vitro, and the formation
26      of protein and DNA adducts.  A highly toxic aldehyde formed during O3-induced lipid
27      peroxidation is 4-hydroxynonenal (FINE). Healthy human subjects exposed to 0.4 ppm O3 for
28      1 h underwent BAL 6 h later. Analysis of lavaged AMs by Western Blot indicated increased
29      levels of a 32-kDa HNE-protein  adduct, as well as 72-kDa heat shock protein and ferritin
30      in O3-versus air-exposed subjects (Hamilton et al., 1998).  In a recent study of healthy subjects
31      exposed to 0.1 ppm O3 for 2 h (Corradi et al., 2002), formation of 8-hydroxy-2'-deoxyguanosine

        August 2005                               6-3 5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (8-OHdG), a biomarker of reactive oxidant species (ROS)-DNA interaction, was measured in
 2      peripheral blood lymphocytes.  At 18 h post exposure, 8-OHdG was significantly increased in
 3      cells compared to pre-exposure levels, presumably linked to concurrent increases in chemical
 4      markers of ROS.  Of interest, the increase in 8-OHdG was only significant in a subgroup of
 5      subjects with the wild genotype for NQ01 and the null genotype for GSTM1, suggesting that
 6      polymorphisms in redox enzymes may confer "susceptibility' to O3 in some individuals.
 7           The generation of ROS following exposure to O3 has been shown to be associated with a
 8      wide range of responses. In a recent study, ROS production by alveolar macrophages lavaged
 9      from subjects exposed to 0.22 ppm for 4 h was assessed by flow cytometry  (Voter et al., 2001).
10      Levels were found to be significantly elevated 18 h post exposure and associated with several
11      markers of increased permeability. An in vitro study of human tracheal epithelial cells exposed
12      to O3 indicated that generation of ROS resulted in decrease in synthesis of the bronchodilatory
13      prostaglandin, PGE2, as a result of inactivation of prostaglandin endoperoxide G/H synthase 2
14      (Alpertetal., 1997).
15
16      6.9.4  Adaptation of Inflammatory Responses
17           Physiologic and symptomatic responses in humans following  repeated exposure to O3 were
18      discussed in Section 6.6. Inflammatory responses upon repeated O3 exposures are discussed in
19      this section. Animal studies suggest that while inflammation may be diminished with repeated
20      exposure, underlying damage to lung epithelial cells continues (Tepper et al., 1989). Markers
21      from HALF following both 2-h (Devlin et al., 1997) and 4-h (Christian et al., 1998; Torres et al.,
22      2000) repeated O3 exposures (up to 5 days) indicate that there is ongoing cellular damage
23      irrespective of the attenuation of some cellular inflammatory responses of the airways,
24      pulmonary function, and symptom responses.
25           Devlin et al. (1997) examined the inflammatory responces of humans  repeatedly exposed
26      to 0.4 ppm O3 for 5 consecutive days.  Several indicators of inflammation (e.g., PMN influx,
27      IL-6,  PGE2, BAL protein, fibronectin) were attenuated after 5 days  of exposure (i.e., values were
28      not different from FA). Several markers (LDH,  IL-8, total protein,  epithelial cells) did not show
29      attenuation, indicating that tissue damage probably continues to occur during repeated exposure.
30      The recovery of the inflammatory response occurred for some markers after 10 days, but some
31      responses were not normalized  even after 20 days.  The continued presence of cellular injury

        August 2005                             6-36        DRAFT-DO NOT QUOTE OR CITE

-------
 1      markers indicates a persistent effect that may not necessarily be recognized due to the
 2      attenuation of spirometric and symptom responses.
 3           Christian et al. (1998) randomly subjected heathy subjects to a single exposure and to 4
 4      consecutive days of exposure to 0.2 ppm O3 for 4 h.  Both "bronchial" and "alveolar" fractions
 5      of the BAL showed decreased numbers of PMNs and fibronectin concentration at day 4 versus
 6      the single exposure, and a decrease in IL-6 levels in the alveolar fraction.  Following a similar
 7      study design and exposure parameters, Torres et al. (2000) found both functional and BAL
 8      cellular responses to O3 were abolished at 24 h postexposure following the fourth exposure day.
 9      However, levels of total protein, IL-6, IL-8, reduced glutathione and ortho-tyrosine were still
10      increased significantly.  In addition, visual scores for bronchitis, erythema and the numbers of
11      neutrophils in the mucosal biopsies were increased.  Their results indicate that, despite reduction
12      of some markers of inflammation in BAL and measures of large airway function, inflammation
13      within the airways persists following repeated exposure to O3.
14           Holz et al.  (2002) made a comparison of early and late responses to allergen challenge
15      following O3 in subjects with allergic rhinitis or allergic asthma. With some variation, both early
16      and late FEVj and cellular responses in the two subject groups were significantly enhanced by 4
17      consecutive days of exposure to 0.125 ppm O3 for 3 h.
18           In another  study, Frank and colleagues (2001) exposed healthy subjects to FA and to O3
19      (0.25 ppm, 2 h) on 4 consecutive days each, with pulmonary function measurements being made
20      prior to and following each exposure. BAL was performed on day 5, 24 h following the last
21      exposure. On day 5, PMN numbers remained significantly higher following O3 compared to FA.
22      Of particular note in this study was the observation that small airway function, assessed by
23      grouping values  for isovolumetric FEF25.75, VmaxSO and Vmax75 into a single value, showed
24      persistent reduction from day 2 through day 5.  These data suggest that techniques monitoring
25      the function in the small peripheral airway regions, the primary sites of O3 uptake in the lung,
26      may provide important information regarding both acute and cumulative effects of O3 exposure.
27
28      6.9.5  Effect of Anti-Inflammatory  and Other Mitigating Agents
29           Pretreatment of healthy subjects with non-steroidal anti-inflammatory drugs (ibuprofen,
30      etc.) has been found to partially suppress development of airway inflammation and pulmonary
31      function changes (U.S. Environmental Protection Agency, 1996).  Although atropine blocked the

        August 2005                              6-37       DRAFT-DO NOT QUOTE OR CITE

-------
 1      increase in Raw in response to O3 exposure, it did not alter the spirometric or symptom
 2      responses (Beckett et al., 1985).  Similarly, albuterol and salbutamol, which had no effect
 3      on O3-induced changes in spirometry, also had no effect of symptom responses (McKenzie et al.,
 4      1987; Gong et al., 1988). The anti-inflammatory medications indomethacin and ibuprofen,
 5      which partially inhibit the spirometric responses to O3 exposure, also cause a reduction in
 6      respiratory symptoms (Schelegle et al., 1987; Hazucha et al., 1994).  Indomethacin attenuates
 7      decrements in FEVj and FVC in healthy subjects, but not asthmatics (Alexis et al.,
 8      2000).  In contrast, inhalation of the corticosteroid budesonide does not prevent or even
 9      attenuate O3-induced responses in healthy subjects as assessed by measurements of lung
10      function, bronchial reactivity and airway inflammation (Nightingale et al., 2000).  In asthmatic
11      subjects, budesonide decreases airway neutrophil influx following O3 exposure (Vagaggini
12      et al., 2001). This suggests that corticosteroids may be effective only when the inflammation
13      is already present, such as in asthmatics.
14          Holz et al. (2005) studied inflammatory responses in healthy ozone-responders (>10%
15      increase in sputum neutrophils from O3) pretreated with single doses (the highest shown to be
16      safe and well tolerated) of inhaled fluticasone and oral prednisolone. The O3 exposure caused
17      small changes in FEVj (-3.6% ± 6.8%) that were not significantly different from baseline or
18      between treatment groups (i.e., prescreening, placebo, fluticasone, and prednisolone). Relative
19      to placebo, the inhaled or oral corticosteroids significantly reduced O3-induced neutrophil levels.
20      These authors note that their study design was intended to test the anti-inflammatory effects of
21      the steroids and that such high-dose regimens should not be considered for potential long-term
22      patient treatment.
23
24      6.9.6    Changes in Host Defense Capability Following Ozone Exposures
25          A number of studies clearly show that a single acute exposure (1  to 4 h) of humans to
26      moderate concentrations  of O3 (0.2  to 0.6 ppm) while exercising at moderate to heavy levels
27      results in a number of cellular and biochemical changes in the lung including an inflammatory
28      response characterized by increased numbers of PMNs, increased permeability of the epithelial
29      cells lining the respiratory tract, cell damage, and production of proinflammatory cytokines and
30      prostaglandins. This response can be detected as early as 1 h after exposure (Koren et al., 1991;
31      Schelegle et al., 1991) and persists for at least 18 h (Aris et al., 1993; Koren et al., 1989).  The

        August 2005                              6-3 8        DRAFT-DO NOT QUOTE OR CITE

-------
 1      response profile of these mediators is not defined adequately, although it is clear that the time
 2      course of response varies for different mediators and cells (Devlin et al.,  1997; Schelegle et al.,
 3      1991).  These changes also occur in humans exposed to 0.08 and 0.10 ppm O3 for 6 to 8 h
 4      (Devlin et al., 1991; Peden et al., 1997). Ozone also causes inflammatory changes in the nose, as
 5      indicated by increased levels of PMNs and albumin, a marker for increased epithelial cell
 6      permeability. Nasal lavage analyses, however, are not necessarily parallel to BAL analyses.
 7           There appears to be no strong correlation between any of the measured cellular and
 8      biochemical changes and changes in lung function measurements, suggesting that different
 9      mechanisms may be responsible for these processes (Balmes et al., 1996; Devlin et al., 1991).
10      The idea of different mechanisms  is supported by a study in which ibuprofen, a cyclooxygenase
11      inhibitor, blunted the O3-induced decrements in lung function without altering the O3-induced
12      increase in PMNs or epithelial cell permeability (Hazucha et al., 1996). In vitro studies suggest
13      that epithelial cells are the primary target of O3 in the lung and that O3 induces them to produce
14      many of the mediators found in the BAL fluid of humans exposed to O3.  Although O3 does not
15      induce AMs to produce  these compounds in large quantities, it does directly impair the ability of
16      AMs to phagocytize and kill microorganisms.
17           A number of studies have found that O3 exposures increases epithelial cell permeability
18      through direct (technetium-99m labeled diethylene triamine pentaacetic acid, 99mTc-DTPA,
19      clearance) and indirect (e.g, increased BAL albumin, protein) techniques. Kehrl et al. (1987)
20      showed increased 99mTc-DTPA clearance in healthy young adults at 75 minutes postexposure
21      to 0.4 ppm O3 for 2 h. More recently, Foster and Stetkiewicz (1996) have shown that
22      increased 99mTc-DTPA clearance persists for at least 18-20 h post-O3 exposure (130 min to
23      average O3 concentration of 0.24 ppm) and the effect is greater at the lung apices than at the
24      base. Increased BAL protein, suggesting O3-induced changes in epithelial permeability,  have
25      also been reported at 1 h and 18 h postexposure (Balmes et  al., 1996; Devlin et al., 1997). A
26      recent meta-analysis of results from 21 publications (Mudway and Kelly, 2004), showed that
27      increased BAL protein is associated with total ozone dose (product of O3 concentration,
28      exposure duration, and VE ).  Changes in permeability associated with acute inflammation may
29      provide increased access of inhaled antigens, particles, and  other substances to the smooth
30      muscle, interstitial cells, and the blood.
        August 2005                              6-39        DRAFT-DO NOT QUOTE OR CITE

-------
 1           In addition to affecting epithelial permeability and AM-mediated clearance in the
 2      respiratory region of the lung, mucociliary clearance of the tracheobronchial airways is also
 3      affected by O3 exposure.  Only two studies (Foster et al., 1987; Gerrity et al., 1993) have
 4      investigated the effect of O3 exposure on mucociliary particle clearance in humans. Foster et al.
 5      (1987) measured clearance during and after a 2 h exposure to 0.4 ppm O3. Gerrity et al. (1993)
 6      measured clearance at 2 h postexposure (0.4 ppm O3), by which time, sRaw had returned to
 7      baseline and FVC was within 5% of baseline (versus an 11% decrement immediately
 8      postexposure).  Foster et  al. (1987) found a stimulatory effect of acute O3 exposure on
 9      mucociliary clearance. Gerrity et al. (1993), who observed no effect on clearance, suggested that
10      transient clearance increases are coincident to pulmonary function responses. Investigators in
11      both studies suggested that O3-induced increases in mucociliary clearance could be mediated by
12      cholinergic receptors.
13
14
15      6.10  EXTRAPULMONARY EFFECTS  OF OZONE
16           Ozone reacts rapidly on contact with respiratory system tissue and is not absorbed or
17      transported to extrapulmonary sites to any significant degree as  such. Human exposure studies
18      discussed in the previous criteria documents (U.S. Environmental Protection Agency, 1986,
19      1996) failed to  demonstrate any consistent extrapulmonary effects.  More recently, some human
20      exposure studies have attempted to identify specific markers of exposure to O3 in blood. Foster
21      et al. (1996) found a reduction in the serum levels of the free radical scavenger a-tocopherol
22      after O3 exposure. Liu et al. (1997, 1999) used a salicylate metabolite, 2,3, dehydroxybenzoic
23      acid (DHBA), to indicate increased levels of hydroxyl radical which hydroxylates salicylate to
24      DHBA. Increased DHBA levels after exposure to 0.12 and 0.40 ppm suggest that O3 increases
25      production of hydroxyl radical.  The levels of DHBA were correlated with changes in
26      spirometry.
27           Gong et al. (1998) observed a statistically significant O3-induced increase the alveolar-to-
28      arterial PO2 gradient in both healthy (n = 6) and hypertensive (n = 10) adult males (41-78 years
29      old) exposed for 3 h with IE (VE « 30 L/min) to 0.3 ppm O3.  The mechanism for the decrease in
30      arterial oxygen tension in the Gong et al. (1998) study could be  due to an O3-induced ventilation-
31      perfusion mismatch. Foster et al. (1993) has demonstrated that even in relatively young healthy

        August 2005                              6-40        DRAFT-DO NOT QUOTE OR CITE

-------
 1      adults (26.7 ± 7 years old), O3 exposure can cause ventilation to shift away from the well
 2      perfused basal lung.  This effect of O3 on ventilation distribution [and by association the small
 3      airways] may persist beyond 24-h post-exposure (Foster et al., 1997). Gong et al. (1998)
 4      suggested that by impairing alveolar-arterial oxygen transfer, the O3 exposure could potentially
 5      lead to adverse cardiac events by decreasing oxygen supply to the myocardium.  The subjects in
 6      the Gong et al. (1998) study had sufficient functional reserve so as to not experience significant
 7      ECG changes or myocardial ischemia and/or injury.
 8           Effects of O3 exposure on alveolar-arterial oxygen gradients may be more pronounced in
 9      patients with preexisting obstructive lung diseases. Relative to healthy elderly subjects, COPD
10      patients have reduced gas exchange and low SaO2. Any inflammatory or edematous responses
11      due to O3 delivered to the well ventilated regions of the COPD lung could further inhibit gas
12      exchange and reduce oxygen saturation. In addition, O3-induced vasoconstriction could also
13      acutely induce pulmonary hypertension. Inducing pulmonary vasoconstriction and hypertension
14      in these patients would perhaps worsen their condition, especially if their right ventricular
15      function was already compromised.
16
17
18      6.11  EFFECTS OF OZONE MIXED WITH OTHER POLLUTANTS
19           Over the past 10 years only a handful of human controlled studies have examined the
20      effects of pollutant mixtures containing O3. The studies summarized in this section complement
21      the studies reviewed in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996).
22      {The complexities ofO3 and co-pollutant exposures in animal studies are discussed in
23      Section 5.4.4).
24           The results of a controlled study on children (Linn et al., 1997), designed to approximate
25      exposure conditions of an epidemiologic study (Neas et al., 1995) by matching the population
26      and exposure atmosphere (0.1 ppm O3, 0.1 ppm SO2 and 101  |ig/m2 H2SO4), did not support the
27      findings  of this epidemiologic study. The study points out the difficulties in attempting to link
28      the outcomes of epidemiologic and controlled studies.  Another vulnerable population,
29      asthmatics, demonstrated enhanced airway reactivity to house dust mite following exposures
30      to O3, NO2, and the combination of the two gases. Spirometric response, however, was impaired
31      only by O3, and O3 + NO2 at higher concentrations (Jenkins et al., 1999). Continuous exposure

        August 2005                              6-41       DRAFT-DO NOT QUOTE OR CITE

-------
 1      to SO2 and NO2 increases inhaled bolus O3 absorption, while continuous exposure to O3
 2      decreases O3 bolus absorption (Rigas et al., 1997).  Inhalation of a mixture of PM25 and O3 by
 3      healthy subjects increased brachial artery tone and reactivity (Brook et al., 2002).  Since no other
 4      cardiovascular endpoints were affected by the exposure, the pathophysiological importance of
 5      this observation remains uncertain. However, acute pulmonary hypertension due to O3-induced
 6      vasoconstriction could pose a risk to individuals with cardiovascular disease (see Section 6-10).
 1           All in all, the contention that air pollutant mixtures elicit stronger pathophysiologic effects
 8      than individual pollutants of the mix is only weakly supported by human studies of either healthy
 9      or at-risk population.
10
11
12      6.12  CONTROLLED STUDIES OF AMBIENT AIR EXPOSURES
13           A large amount of informative O3 exposure-effects data has been obtained in controlled
14      laboratory exposure studies under a variety of different experimental conditions.  However,
15      laboratory simulation of the variable pollutant mixtures present in ambient air is not practical.
16      Thus, the  exposure effects of one or several artificially generated pollutants (i.e., a simple
17      mixture) on pulmonary function and symptoms may not explain responses to ambient air where
18      complex pollutant mixtures exist.
19
20      6.12.1  Mobile Laboratory Studies
21           Quantitatively useful information on the effects of acute exposure to photochemical
22      oxidants on pulmonary function and symptoms responses from field studies using a mobile
23      laboratory were presented in prior criteria documents (U.S. Environmental Protection Agency,
24      1986, 1996).  Relative to controlled exposure studies, mobile laboratory ambient air studies
25      suffer the additional limation of a dependence on ambient outdoor conditions.  Consistent with
26      controlled exposure studies, mobile studies in California demonstrated that pulmonary effects
27      from exposure to ambient  air in Los Angeles  are related to O3 concentration and level of
28      exercise.  Healthy subjects with a history of allergy also appeared to be more responsive to O3
29      than "nonallergic" subjects (Linn et al., 1980, 1983b), although a standardized evaluation of
30      atopic status was  not performed.
31

        August 2005                              6-42       DRAFT-DO NOT QUOTE OR CITE

-------
 1      6.12.2  Aircraft Cabin Studies
 2           Respiratory symptoms and pulmonary function effects resulting from exposure to O3 in
 3      commercial aircraft flying at high altitudes, and in altitude-simulation studies, have been
 4      assessed previously (U.S. Environmental Protect!on Agency, 1986, 1996).  Commercial aircraft
 5      cabin O3 levels were reported to be very low (average concentration 0.01 to 0.02 ppm) during
 6      92 randomly selected smoking and nonsmoking flights in 1989 (Nagda et al., 1989). None of
 7      these flights recorded O3 concentrations exceeding the 3-h time-weighted average (TWA)
 8      standard of 0.10 ppm promulgated by the U.S. Federal Aviation Administration (FAA, 1980),
 9      probably due to the use of O3-scrubbing catalytic filters (Melton, 1990).
10           Ozone contamination aboard high-altitude aircraft also has been an interest to the U.S. Air
11      Force because of complaints by crew members of frequent symptoms of dryness and irritation of
12      the eyes, nose, and throat and an occasional cough (Hetrick et al., 2000). Despite the lack of
13      ventilation system modifications as used in commercial aircraft, the O3 concentrations  never
14      exceeded the FAA ceiling limit of 0.25 ppm and exceeded the 3-h TWA of 0.10 ppm only 7% of
15      the total monitored flight time (43  h). The authors concluded that extremely low average
16      relative humidity (12%) during flight operations was most likely responsible for the reported
17      symptoms.
18
19
20      6.13  SUMMARY
21           Responses in humans exposed to ambient O3 concentrations include decreased inspiratory
22      capacity; mild bronchoconstriction; rapid, shallow breathing pattern during exercise; and
23      symptoms of cough and pain on deep inspiration. Reflex inhibition of inspiration results in a
24      decrease in forced vital capacity (FVC) and, in combination with mild bronchoconstriction,
25      contributes to a decrease in the forced expiratory volume in 1 s (FEVj).  In addition to
26      physiological pulmonary responses and symptoms of breathing discomfort, O3 exposure also
27      results in airway hyperresponsiveness, inflammation, immune system activation, and epithelial
28      injury.  With repeated O3 exposures over several days, spirometric and symptom responses
29      become attenuated, but this tolerance is lost after about a week without exposure. Airway
30      responsiveness also appears to be attenuated with repeated O3 exposures, but less than  FEVj.
        August 2005                              6-43        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Unlike spirometric and symptom responses, airway inflammation and small airways dysfunction
 2      may not become attenuated by repeated O3 exposures.
 3           Young healthy adults exposed to O3 concentrations >0.08 ppm develop significant
 4      reversible, transient decrements in pulmonary function if minute ventilation (VE ) or duration of
 5      exposure are increased sufficiently. The pattern of FEVj response appears to depend on the O3
 6      exposure profile.  Triangular exposure profiles can potentially lead to greater FEVj responses
 7      than square wave exposures at equivalent average O3 doses.  O3-induced decrements in FEVj do
 8      not appear to depend on gender, race, body surface area, height, lung size, or baseline FVC in
 9      young healthy adults. Healthy children experience similar spirometric responses but lesser
10      symptoms from O3 exposure relative to young adults.  On average, spirometric and symptom
11      responses to O3 exposure appear to decline with increasing age beyond approximately 18 years
12      of age. There is a large degree of intersubject variability in physiologic and  symptomatic
13      responses of heathy adults exposed to O3. However, responses tend to be reproducible within  a
14      given individual over a period of several months.  With increasing O3 concentration, the
15      distribution of FEVj decrements becomes asymmetrical with a few individuals experiencing
16      large decrements.
17           There is a tendency for slightly increased  spirometric responses in mild asthmatics and
18      allergic rhinitics relative to healthy young adults.  Spirometric responses in asthmatics appear  to
19      be affected by baseline lung function, i.e., responses increase with disease severity. With
20      repeated daily O3 exposures, spirometric responses of asthmatics become attenuated;  however,
21      airway responsiveness becomes increased in subjects with preexisting allergic airway disease
22      (with or without asthma).  Possibly due to patient age, O3 exposure does not  appear to cause
23      significant pulmonary function impairment or evidence of cardiovascular strain in patients with
24      cardiovascular disease or chronic obstructive pulmonary disease relative to healthy subjects.
25           Available information on recovery from O3 exposure indicates that an initial phase of
26      recovery in healthy individuals proceeds relatively rapidly, with acute spirometric and symptom
27      responses resolving within about 2 to 4 h. Small residual lung function effects are almost
28      completely resolved within 24 hours.  Effects of O3 on the small airways, assessed by persistent
29      decrement in FEF25.75 and altered ventilation distribution, may be due in part to inflammation.
30      Indeed, a prolonged recovery of residual spirometric decrements following the initial rapid
31      recovery could be due to slowly resolving airway inflammation. In hyperresponsive individuals,

        August 2005                              6-44        DRAFT-DO NOT QUOTE OR CITE

-------
 1      this recovery takes longer, as much as 48 hours, to return to baseline values.  Persistent
 2      spirometry changes observed for up to 48 h postexposure could plausibly be sustained by the
 3      inflammatory mediators.  Cellular responses (e.g., release of immunomodulatory cytokines)
 4      appear to still be active as late as 20 h postexposure. More slowly developing inflammatory and
 5      cellular changes may persist for up to 48 h, but the time course for these parameters in humans
 6      has not been explored fully.
 7           Soluble mediators of inflammation such as the cytokines (IL-6, IL-8) and arachidonic acid
 8      metabolites (e.g., PGE2, PGF2a, thromboxane, and leukotrienes [LTs] such as LTB4) have been
 9      measured in the BAL fluid of humans exposed to O3. Many of these compounds have
10      bronchoconstrictive properties and may be involved in increased airway responsiveness
11      following O3 exposure. Some indicators of inflammation (e.g., PMN influx, IL-6, PGE2,
12      fibronectin) are attenuated with repeated O3 exposures.  However, indicating that tissue  damage
13      probably continues to occur during repeated O3 exposure, other markers (LDH, IL-8, total
14      protein, epithelial cells) do not show attenuation.  There  appears to be no strong correlation
15      between any of the measured cellular and biochemical changes and changes in lung function
16      measurements. A limited number of studies suggest that inflammatory responses may be
17      detected following O3 exposures that are insufficient to cause decrements in pulmonary  function.
18      Whether airway reactivity or inflammatory responses to  O3 are dependent on the age of the
19      exposed individual, such  as spirometric responses, has not been determined.
20           Dietary antioxidant supplementation attenuates O3-induced spirometric responses but not
21      the intensity of subjective symptoms nor inflammatory responses. Dietary antioxidants  also
22      afforded partial protection to asthmatics by attenuating postexposure bronchial
23      hyperresponsiveness.
24
        August 2005                              6-45        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES

  2       Adams, W. C. (2000) Ozone dose-response effects of varied equivalent minute ventilation rates. J. Exposure Anal.
  3             Environ. Epidemiol. 10: 217-226.
  4       Adams, W. C. (2002) Comparison of chamber and face-mask 6.6-hour exposures to ozone on pulmonary function
  5             and symptoms responses. Inhalation Toxicol. 14: 745-764.
  6       Adams, W. C. (2003a) Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm ozone via
  7             square-wave and triangular profiles on pulmonary responses. Inhalation Toxicol. 15: 265-281.
  8       Adams, W. C. (2003b) Relation of pulmonary responses induced by 6.6-h exposures to 0.08 ppm ozone and 2-h
  9             exposures to 0.30 ppm ozone via chamber and face-mask inhalation. Inhalation Toxicol. 15: 745-759.
10       Adams, W. C.; Ollison, W. M.  (1997) Effects of prolonged simulated ambient ozone dosing patterns on human
11             pulmonary function and symptomatology. Presented at: 90th annual meeting of the Air & Waste Management
12             Association; June; Toronto, Ontario, Canada. Pittsburgh, PA: Air & Waste Management Association; paper
13             no. 97-MP9.02.
14       Adams, W. C.; Schelegle, E. S. (1983) Ozone and high ventilation effects on pulmonary function and endurance
15             performance. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 55: 805-812.
16       Adams, W. C.; Savin, W. M.; Christo, A. E. (1981) Detection of ozone toxicity during continuous exercise via the
17             effective dose concept. J. Appl. Physiol.:  Respir. Environ. Exercise Physiol. 51: 415-422.
18       Alexis, N.; Urch, B.; Tarlo, S.;  Corey, P.; Pengelly, D.; O'Byrne, P.; Silverman, F. (2000) Cyclooxygenase
19             metabolites play a different role in ozone-induced pulmonary function decline in asthmatics compared to
20             normals. Inhalation Toxicol. 12: 1205-1224.
21       Alpert, S. E.; Walenga, R. W.; Jaspers, I.; Qu, Q.; Chen, L. C. (1997) Ozone inactivates Cyclooxygenase in human
22             tracheal epithelial cells without altering PGHS-2 mRNA or protein. Am. J. Physiol. 272: L879-L887.
23       Aris, R. M.; Christian, D.; Hearne, P. Q.; Kerr, K.; Finkbeiner,  W. E.; Balmes, J.  R. (1993) Ozone-induced airway
24             inflammation in human subjects as determined by airway lavage and biopsy. Am. Rev. Respir. Dis.
25             148: 1363-1372.
26       Aris, R. M.; Tager, L; Christian, D.; Kelly, T.; Balmes, J. R. (1995) Methacholine responsiveness is not associated
27             with O3-induced decreases in FEVj. Chest 107: 621-628.
28       Avissar, N. E.; Reed, C. K.; Cox, C.; Frampton,  M. W.; Finkelstein, J. N. (2000)  Ozone, but not nitrogen dioxide,
29             exposure  decreases glutathione peroxidases in epithelial lining fluid of human lung. Am. J. Respir. Crit. Care
30             Med. 162: 1342-1347.
31       Balmes, J. R.; Chen, L. L.; Scannell, C.; Tager, L; Christian, D.; Hearne, P. Q.; Kelly, T.; Aris, R. M. (1996)
32             Ozone-induced decrements in FEVl and FVC do not correlate with measures of inflammation. Am. J. Respir.
33             Crit. Care Med. 153: 904-909.
34       Balmes, J. R.; Aris, R. M.; Chen, L. L.; Scannell, C.; Tager, I. B.; Finkbeiner, W.; Christian, D.; Kelly, T.; Hearne,
35             P. Q.; Ferrando, R.; Welch, B. (1997) Effects of ozone on normal and potentially sensitive human subjects.
36             part I: airway inflammation and responsiveness to ozone in normal and asthmatic subjects. Cambridge, MA:
37             Health Effects Institute. Research report no. 78; pp  1-37, 81-99.
38       Bascom, R.; Naclerio, R.  M.; Fitzgerald, T. K.; Kagey-Sobotka, A.; Proud, D. (1990) Effect of ozone inhalation on
39             the response to nasal challenge with antigen of allergic subjects. Am. Rev. Respir. Dis. 142: 594-601.
40       Basha, M. A.; Gross, K. B.; Gwizdala, C. J.; Haidar, A. H.; Popovich, J., Jr. (1994) Bronchoalveolar lavage
41             neutrophilia in asthmatic and healthy volunteers after controlled exposure to ozone and filtered purified air.
42             Chest 106: 1757-1765.
43       Beckett, W. S.; McDonnell, W. F.; Horstman, D. H.; House, D. E. (1985) Role of the parasympathetic nervous
44             system in acute lung response to ozone. J. Appl. Physiol. 59: 1879-1885.
45       Bedi, J. F.; Drechsler-Parks, D. M.; Horvath, S.  M. (1985) Duration of increased  pulmonary function sensitivity to
46             an initial ozone exposure. Am. Ind. Hyg.  Assoc. J. 46: 731-734.
47       Bergamaschi, E.; De Palma, G.; Mozzoni, P.;  Vanni, S.; Vettori, M. V.; Broeckaert, F.; Bernard, A.; Mutti, A.
48             (2001) Polymorphism of quinone-metabolizing enzymes and susceptibility to ozone-induced acute effects.
49             Am. J. Respir. Crit. Care Med. 163: 1426-1431.
50       Blomberg, A.; Helleday, R.; Pourazar, J.; Stenfors, N.; Kelly, F. J.; Frew, A. J.; Holgate, S. T.; Sandstrom, T. (1997)
51             Early airway and peripheral blood cell  responses to 0.20 ppm ozone in healthy human subjects. Eur. Respir. J.
52             10(suppl 25): 274S.
53       Blomberg, A.; Mudway, I. S.; Nordenhall, C.; Hedenstrom, H.; Kelly, F. J.; Frew, A. J.; Holgate, S. T.;
54             Sandstrom,  T. (1999) Ozone-induced lung function decrements do not correlate with early airway
55             inflammatory or antioxidant responses. Eur. Respir. J. 13: 1418-1428.
         August 2005                                    6-46        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Bosson, I; Stenfors, N.; Bucht, A.; Helleday, R.; Pourazar, I; Holgate, S. T.; Kelly, F. I; Sandstrom, T.; Wilson, S.;
  2             Frew, A. J.; Blomberg, A. (2003) Ozone-induced bronchial epithelial cytokine expression differs between
  3             healthy and asthmatic subjects. Clin. Exp. Allergy 33: 777-782.
  4       Brook, R. D.; Brook, J. R.; Urch, B.; Vincent, R.; Rajagopalan,  S.; Silverman, F. (2002) Inhalation of fine paniculate
  5             air pollution and ozone causes acute arterial vasoconstriction in healthy adults. Circulation 105: 1534-1536.
  6       Bush, M. L.; Asplund, P. T.; Miles, K. A.; Ben-Jebria, A.; Ultman, J. S. (1996) Longitudinal distribution of O3
  7             absorption in the lung: gender differences and intersubject variability. J. Appl. Physiol. 81: 1651-1657.
  8       Christian, D. L.; Chen, L. L.; Scannell, C. H.; Ferrando, R. E.; Welch, B. S.; Balmes, J. R. (1998) Ozone-induced
  9             inflammation is attenuated with multiday exposure. Am.  J. Respir. Crit. Care Med. 158: 532-537.
10       Coleridge, J. C. G.; Coleridge, H. M.; Schelegle, E. S.; Green, J. F. (1993) Acute inhalation of ozone stimulates
11             bronchial C-fibers and rapidly adapting receptors in dogs. J. Appl. Physiol. 74: 2345-2352.
12       Corradi, M.; Alinovi, R.; Goldoni, M.; Vettori, M.; Folesani, G.; Mozzoni, P.; Cavazzini, S.; Bergamaschi, E.;
13             Rossi, L.; Mutti, A. (2002) Biomarkers of oxidative stress after controlled human exposure to ozone. Toxicol.
14             Lett. 134: 219-225.
15       David, G. L.; Romieu, I.; Sienra-Monge, J. J.; Collins, W. J.; Ramirez-Aguilar, M.; Del Rio-Navarro, B. E.;
16             Reyes-Ruiz, N. I.; Morris, R. W.; Marzec, J. M.; London, S. J. (2003) Nicotinamide adenine dinucleotide
17             (Phosphate) reduced:quinone oxidoreductase and glutathione s-transferase ml polymorphism and childhood
18             asthma.  Am. J. Respir. Crit. Care Med. 168: 1199-1204.
19       Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
20             Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the
21             lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
22       Devlin, R. B.; McDonnell, W. F.; Becker, S.; Madden, M. C.; McGee, M. P.; Perez, R.; Hatch, G.; House, D. E.;
23             Koren, H. S. (1996) Time-dependent changes of inflammatory mediators in the lungs of humans exposed to
24             0.4 ppm ozone for 2 hr: a comparison of mediators found in bronchoalveolar lavage fluid 1 and 18 hr after
25             exposure. Toxicol. Appl. Pharmacol. 138: 176-185.
26       Devlin, R. B.; Folinsbee, L. J.; Biscardi, F.; Hatch, G.; Becker, S.; Madden,  M. C.; Robbins, M.; Koren, H. S. (1997)
27             Inflammation and cell damage induced by repeated exposure of humans to ozone. Inhalation Toxicol.
28             9:211-235.
29       Dimeo, M. J.; Glenn, M.  G.; Holtzman, M. J.; Sheller, J. R.; Nadel, J. A.; Boushey, H. A. (1981) Threshold
3 0             concentration of ozone causing an increase in bronchial reactivity in humans and adaptation with repeated
31             exposures. Am. Rev. Respir. Dis. 124: 245-248.
32       Federal Aviation Administration. (1980) Airplane cabin ozone contamination. F. R. (January 21) 45: 3880-3883.
33       Folinsbee, L. J.; Hazucha, M. J. (1989) Persistence of ozone-induced changes in lung function and airway
34             responsiveness. In: Schneider, T.; Lee, S. D.;  Wolters,  G. J. R.; Grant, L. D., eds. Atmospheric ozone research
3 5             and its policy  implications: proceedings of the 3rd US-Dutch international symposium; May 1988; Nijmegen,
36             The Netherlands. Amsterdam, The Netherlands: Elsevier Science Publishers; pp. 483-492. (Studies in
37             environmental science 35).
38       Folinsbee, L. J.; Hazucha, M. J. (2000) Time course  of response to ozone exposure in healthy adult females.
39             Inhalation Toxicol. 12: 151-167.
40       Folinsbee, L. J.; Horvath, S.  M. (1986) Persistence of the acute effects of ozone exposure. Aviat. Space Environ.
41             Med. 57: 1136-1143.
42       Folinsbee, L. J.; Silverman, F.; Shephard, R. J. (1977) Decrease of maximum work performance following ozone
43             exposure. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 42: 531-536.
44       Folinsbee, L. J.; Drinkwater, B. L.; Bedi, J. F.; Horvath, S. M. (1978) The influence of exercise on the pulmonary
45             function changes due  to exposure to low concentrations of ozone. In:  Folinsbee, L. J.; Wagner, J. A.; Borgia,
46             J. F.; Drinkwater, B. L.; Gliner, J. A.; Bedi, J. F., eds. Environmental stress: individual human adaptations.
47             New York, NY: Academic Press; pp.  125-145.
48       Folinsbee, L. J.; Bedi, J. F.; Horvath, S. M. (1980) Respiratory responses in  humans repeatedly exposed to  low
49             concentrations of ozone. Am. Rev. Respir. Dis.  121: 431-439.
50       Folinsbee, L. J.; McDonnell, W. F.; Horstman, D. H. (1988) Pulmonary function and symptom responses after
51             6.6-hour exposure to 0.12  ppm ozone with moderate exercise. JAPCA 38: 28-35.
52       Folinsbee, L. J.; Horstman, D. H.; Kehrl, H. R.; Harder, S.; Abdul-Salaam, S.; Ives, P. J.  (1994) Respiratory
53             responses to repeated prolonged exposure to 0.12 ppm ozone. Am. J.  Respir. Crit. Care Med. 149: 98-105.
54       Folinsbee, L. J.; Devlin, R. B.; Robbins, M. K.; Biscardi, F. H.;  Abdul-Salaam, S.; Koren, H. S. (1998) Repeated
55             exposure of humans to ozone: pulmonary function and symptom responses. Research Triangle Park, NC:
56             U.S. Environmental Protection Agency; National Center for Environmental Assessment; unpublished data.
         August 2005                                    6-47         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Foster, W. M; Stetkiewicz, P. T. (1996) Regional clearance of solute from the respiratory epithelia: 18-20 h
  2             postexposure to ozone. J. Appl. Physiol. 81: 1143-1149.
  3       Foster, W. M.; Costa, D. L.; Langenback, E. G. (1987) Ozone exposure alters tracheobronchial mucociliary function
  4             in humans. J. Appl. Physiol. 63: 996-1002.
  5       Foster, W. M.; Silver, J. A.; Groth, M. L. (1993) Exposure to ozone alters regional function and particle dosimetry in
  6             the human lung. J.  Appl. Physiol. 75: 1938-1945.
  7       Foster, W. M.; Wills-Karp, M.; Tankersley, C. G.; Chen, X.; Paquette, N. C. (1996) Bloodborne markers in humans
  8             during multiday exposure to ozone. J. Appl. Physiol. 81: 794-800.
  9       Foster, W. M.; Weinmann, G. G.; Menkes, E.; Macri, K. (1997) Acute exposure of humans to ozone impairs small
10             airway function. Ann. Occup. Hyg. 41(suppl. 1): 659-666.
11       Foster, W. M.; Brown, R. H.; Macri, K.; Mitchell, C. S. (2000) Bronchial reactivity of healthy subjects: 18-20 h
12             postexposure to ozone. J. Appl. Physiol. 89: 1804-1810.
13       Fox, S. D.; Adams, W. C.; Brookes, K. A.; Lasley, B. L. (1993) Enhanced response to ozone exposure during the
14             follicular phase of the menstrual cycle.  Environ. Health Perspect. 101: 242-244.
15       Foxcroft, W. J.; Adams, W. C. (1986) Effects  of ozone exposure on four consecutive days on work performance and
16             VO2max. J. Appl. Physiol. 61:  960-966.
17       Frampton, M. W.; Morrow, P. E.; Torres, A.; Cox, C.; Voter, K. Z.; Utell, M. J.; Gibb, F. R.; Speers, D. M.  (1997)
18             Ozone responsiveness in smokers and nonsmokers. Am. J.  Respir. Crit. Care Med. 155:  116-121.
19       Frank, R.; Liu, M.  C.; Spannhake, E. W.; Mlynarek, S.; Macri, K.; Weinmann, G. G. (2001) Repetitive ozone
20             exposure of young adults: evidence of persistent small airway dysfunction. Am. J. Respir. Crit. Care Med.
21             164: 1253-1260.
22       Freed, A. N.; Chou, C. L.; Fuller, S. D.; Croxton, T. L. (1996) Ozone-induced vagal reflex modulates airways
23             reactivity in rabbits. Respir. Physiol.  105: 95-102.
24       Gabrielson, E. W.; Yu, X.-Y.; Spannhake, E. W. (1994) Comparison of the toxic effects of hydrogen peroxide and
25             ozone on cultured human bronchial epithelial cells. Environ. Health Perspect. 102: 972-974.
26       Gerrity, T. R.; Bennett, W. D.; Kehrl, H.; DeWitt, P. J. (1993) Mucociliary clearance of inhaled particles measured
27             at 2 h after ozone exposure in humans.  J. Appl. Physiol. 74: 2984-2989.
28       Gong, H., Jr.; Bradley, P. W.; Simmons, M. S.; Tashkin, D. P. (1986) Impaired exercise performance and pulmonary
29             function in elite cyclists during low-level ozone exposure in a hot environment. Am. Rev. Respir. Dis.
30             134:726-733.
31       Gong, H., Jr.; Bedi, J. F.; Horvath, S. M. (1988) Inhaled albuterol does not protect against ozone toxicity in
32             nonasthmatic athletes. Arch. Environ. Health43: 46-53.
33       Gong, H., Jr.; Shamoo, D. A.; Anderson, K. R.; Linn, W. S. (1997a) Responses of older men with and without
34             chronic obstructive pulmonary disease to prolonged ozone  exposure. Arch. Environ. Health 52: 18-25.
35       Gong, H., Jr.; McManus, M. S.; Linn,  W. S. (1997b) Attenuated response to repeated daily ozone exposures in
36             asthmatic subjects. Arch. Environ. Health 52: 34-41.
37       Gong, H., Jr.; Wong, R.; Sarma, R. J.; Linn, W. S.; Sullivan, E. D.; Shamoo, D. A.; Anderson, K. R.; Prasad, S. B.
3 8             (1998) Cardiovascular effects of ozone  exposure in human volunteers. Am. J. Respir. Crit. Care Med.
39             158: 538-546.
40       Graham, D. E.; Koren, H. S. (1990) Biomarkers of inflammation in ozone-exposed humans: comparison of the nasal
41             and bronchoalveolar lavage. Am. Rev. Respir. Dis. 142: 152-156.
42       Graham, D.; Henderson, F.; House, D. (1988)  Neutrophil influx measured in nasal lavages of humans exposed to
43             ozone. Arch. Environ. Health 43: 228-233.
44       Gwynn, R. C.; Thurston, G. D. (2001) The burden of air pollution: impacts among racial minorities. Environ. Health
45             Perspect. Suppl. 109(4): 501-506.
46       Hamilton, R. F.; Li, L.; Eschenbacher, W. L.;  Szweda, L.; Holian, A. (1998) Potential involvement of
47             4-hydroxynonenal in the response of human lung cells to ozone. Am. J. Physiol. 274: L8-L16.
48       Hasday, J. D.; Bascom, R.; Costa, J. J.; Fitzgerald, T.; Dubin,  W.  (1999) Bacterial endotoxin is an  active component
49             of cigarette  smoke. Chest 115: 829-835.
50       Hazbun, M. E.; Hamilton, R.; Holian, A.; Eschenbacher, W. L. (1993) Ozone-induced increases in substance P and
51             8-epi-prostaglandin F2a in the airways of human subjects. Am. J. Respir. Cell Mol. Biol. 9: 568-572.
52       Hazucha, M. J.; Sant'Ambrogio, G. (1993) Effects of ozone on the activity of slowly (SAR) and rapidly adapting
53             (RAR) receptors in cats. FASEB J. 7: 407A.
54       Hazucha, M. J.; Folinsbee, L. J.; Seal, E., Jr. (1992) Effects of steady-state and variable ozone concentration profiles
55             on pulmonary function. Am. Rev. Respir. Dis. 146:  1487-1493.
         August 2005                                    6-48         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Hazucha, M. I; Folinsbee, L. I; Seal, E.; Bromberg, P. A. (1994) Lung function response of healthy women after
  2             sequential exposures to NO2 and O3. Am. J. Respir. Crit. Care Med. 150: 642-647.
  3       Hazucha, M. J.; Madden, M.; Pape, G.; Becker, S.; Devlin, R.; Koren, H. S.; Kehrl, H.; Bromberg, P. A. (1996)
  4             Effects of cyclo-oxygenase inhibition on ozone-induced respiratory inflammation and lung function changes.
  5             Eur. J. Appl. Physiol. Occup. Med. 73: 17-27.
  6       Hazucha, M. J.; Folinsbee, L. J.; Bromberg, P. A. (2003) Distribution and reproducibility of spirometric response to
  7             ozone by gender and age. J. Appl. Physiol. 95: 1917-1925.
  8       Hetrick, S. M.; Gould, W. D.; Christensen, D. E. (2000) Inflight cabin ozone aboard long duration C-5 airlift
  9             missions: a historical issue  revisited. Aviat. Space Environ. Med. 71: 408-414.
10       Hiltermann, T. J. N.; Stolk, J.; Hiemstra, P. S.; Fokkens, P. H. B.; Rombout, P. J. A.; Sont, J. K.;  Sterk, P.  J.;
11             Dijkman, J. H. (1995) Effect of ozone exposure on maximal airway narrowing in non-asthmatic and asthmatic
12             subjects. Clin. Sci. 89: 619-624.
13       Hiltermann, T. J. N.; de Bruijne, C. R.;  Stolk, J.; Zwinderman, A. H.; Spieksma, F. Th. M.; Roemer, W.;
14             Steerenberg, P. A.; Fischer, P.  H.; van Bree, L.; Hiemstra, P. S. (1997) Effects of photochemical air pollution
15             and allergen exposure on upper respiratory tract inflammation in asthmatics. Am. J. Respir. Crit. Care Med.
16             156: 1765-1772.
17       Hiltermann, J. T. N.; Lapperre, T.  S.;  Van Bree, L.; Steerenberg, P. A.; Brahim, J. J.; Sont, J. K.;  Sterk, P.  J.;
18             Hiemstra, P. S.; Stolk, J.  (1999) Ozone-induced inflammation assessed in sputum and bronchial lavage fluid
19             from asthmatics: a new noninvasive tool in epidemiologic studies on air pollution and asthma. Free Radical
20             Biol. Med. 27: 1448-1454.
21       Holz, O.; Torres, R. A.; Timm, P.;  Mticke, M.; Richter, K.; Koschyk, S.; Magnussen, H. (1999) Ozone-induced
22             airway inflammatory changes differ between individuals and are reproducible. Am. J. Respir. Crit. Care Med.
23             159:776-784.
24       Holz, O.; Mticke, M.; Paasch, K.;  Bohme, S.; Timm, P.; Richter, K.; Magnussen, H.; Torres, R. A. (2002) Repeated
25             ozone exposures enhance bronchial allergen responses in subjects with rhinitis or asthma. Clin. Exp. Allergy.
26             32:681-689.
27       Holz, O.; Tal-Singer, R.; Kanniess, F.; Simpson, K. J.; Gibson, A.; Vessey, R. S. J.; Janicki, S.; Magnussen, H.;
28             Torres, R. A.; Richter, K. (2005)  Validation of the human ozone challenge model as a tool  for assessing
29             anti-inflammatory drugs in early development. J. Clin. Pharmacol. 45: 498-503.
30       Hoppe, P.; Peters, A.; Rabe,  G.; Praml,  G.; Lindner, J.; Jakobi, G.; Fruhmann, G.; Nowak, D. (2003) Environmental
31             ozone effects in different population subgroups. Int. J. Hyg. Environ. Health 206: 505-516.
32       Horstman, D. H.; Folinsbee,  L. J.; Ives, P. J.; Abdul-Salaam, S.;  McDonnell, W. F. (1990) Ozone concentration and
33             pulmonary response relationships for 6.6-hour exposures with five hours of moderate exercise to 0.08, 0.10,
34             and 0.12 ppm. Am. Rev.  Respir.  Dis. 142:  1158-1163.
35       Horstman, D. H.; Ball, B. A.; Brown, J.; Gerrity, T.; Folinsbee, L. J. (1995) Comparison of pulmonary responses of
3 6             asthmatic and nonasthmatic subjects performing light exercise while exposed to a low level of ozone.
37             Toxicol. Ind. Health 11: 369-385.
38       Horvath, S. M.; Gliner, J. A.; Matsen-Twisdale, J. A. (1979) Pulmonary function and maximum exercise responses
39             following acute ozone exposure. Aviat. Space Environ. Med. 50: 901-905.
40       Horvath, S. M.; Gliner, J. A.; Folinsbee, L. J. (1981) Adaptation to ozone: duration of effect. Am. Rev. Respir. Dis.
41             123:496-499.
42       Housley, D. G.; Eccles, R.; Richards, R. J. (1996) Gender difference in the concentration of the antioxidant uric acid
43             in human nasal lavage. Acta Oto-Laryngol. 116: 751-754.
44       Jenkins, H. S.; Devalia, J. L.; Mister,  R. L.; Bevan, A. M.; Rusznak, C.; Davies, R. J. (1999) The effect of exposure
45             to ozone and nitrogen dioxide on the airway response of atopic asthmatics to inhaled allergen: dose- and
46             time-dependent effects. Am. J. Respir. Crit. Care Med. 160: 33-39.
47       Joad, J. P.; Kott, K. S.; Brie, J. M. (1996) The local C-fiber contribution to ozone-induced effects on the isolated
48             guinea pig lung. Toxicol. Appl. Pharmacol. 141: 561-567.
49       Torres, R.; Nowak, D.; Magnussen, H.;  Speckin, P.; Koschyk, S. (1996) The effect of ozone exposure on allergen
50             responsiveness in subjects with asthma or rhinitis. Am. J.  Respir. Crit. Care Med. 153: 56-64.
51       Torres, R. A.; Holz, O.; Zachgo, W.; Timm, P.; Koschyk, S.; Mtiller, B.; Grimminger, F.; Seeger, W.; Kelly, F. J.;
52             Dunster, C.; Frischer, T.; Lubec, G.; Waschewski, M.; Niendorf, A.; Magnussen, H. (2000) The effect of
53             repeated ozone exposures on inflammatory markers inbronchoalveolar lavage fluid and mucosal biopsies.
54             Am. J. Respir. Crit. Care Med. 161:  1855-1861.
55       Kehrl, H. R.; Hazucha, M. J.; Solic, J. J.; Bromberg, P. A. (1985) Responses of subjects with chronic obstructive
56             pulmonary disease after exposures to 0.3 ppm ozone. Am. Rev. Respir. Dis. 131: 719-724.
         August 2005                                     6-49         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Kehrl, H. R.; Vincent, L. M; Kowalsky, R. J.; Horstman, D. H.; O'Neil, J. I; McCartney, W. H.; Bromberg, P. A.
  2             (1987) Ozone exposure increases respiratory epithelial permeability in humans. Am. Rev. Respir. Dis.
  3             135: 1124-1128.
  4       Kehrl, H. R.; Peden, D. B.; Ball, B. A.; Folinsbee, L. J.; Horstman, D. H. (1999) Increased specific airway reactivity
  5             of persons with mild allergic asthma after 7.6 hours of exposure to 0.16 ppm ozone. J. Allergy. Clin.
  6             Immunol. 104: 1198-1204.
  7       Kleinman, M. T.; Mautz, W. J.; Bjarnason, S. (1999) Adaptive and non-adaptive responses in rats exposed to ozone,
  8             alone and in mixtures, with acidic aerosols. Inhalation Toxicol. 11:  249-264.
  9       Koren, H.  S.; Devlin, R. B.; Graham, D. E.; Mann, R.; McGee, M. P.; Horstman, D. H.; Kozumbo, W. J.; Becker, S.;
10             House, D. E.; McDonnell, W. F.; Bromberg, P. A. (1989) Ozone-induced inflammation in the lower airways
11             of human subjects. Am. Rev. Respir. Dis. 139:  407-415.
12       Koren, H.  S.; Devlin, R. B.; Becker, S.; Perez, R.; McDonnell, W. F. (1991) Time-dependent changes of markers
13             associated with inflammation in the lungs of humans exposed to ambient levels of ozone. Toxicol. Pathol.
14             19:406-411.
15       Kozumbo, W. J.; Hanley, N. M.; Agarwal, S.; Thomas, M. J.; Madden, M. C. (1996) Products of ozonized
16             arachidonic acid potentiate the formation of DNA single strand breaks in cultured human lung cells. Environ.
17             Mol. Mutagen. 27: 185-195.
18       Kreit, J. W.; Gross, K. B.; Moore,  T. B.; Lorenzen, T. J.; D'Arcy, J.; Eschenbacher, W. L. (1989) Ozone-induced
19             changes in pulmonary function and bronchial responsiveness in asthmatics. J. Appl. Physiol. 66: 217-222.
20       Krishna, M. T.; Springall, D.; Meng, Q.-H.; Withers, N.; Macleod, D.; Biscione, G.; Frew, A.; Polak, J.; Holgate, S.
21             (1997) Effects of ozone on epithelium and sensory nerves in the bronchial mucosa of healthy humans. Am. J.
22             Respir. Crit. Care Med. 156: 943-950.
23       Kulle, T. J.; Sauder, L. R.; Kerr, H. D.; Farrell, B. P.; Bermel, M. S.; Smith, D. M. (1982) Duration of pulmonary
24             function adaptation to ozone in humans. Am. Ind. Hyg. Assoc. J. 43: 832-837.
25       Kulle, T. J.; Milman, J. H.; Sauder, L. R.; Kerr, H. D.; Farrell, B. P.; Miller, W. R. (1984) Pulmonary function
26             adaptation to ozone in subjects with chronic bronchitis. Environ. Res. 34: 55-63.
27       Kulle, T. J.; Sauder, L. R.; Hebel, J. R.; Chatham, M. D. (1985) Ozone response relationships in healthy
28             nonsmokers. Am. Rev. Respir. Dis. 132: 36-41.
29       Linder, J.; Herren, D.; Monn, C.; Wanner, H.-U. (1988) Die  Wirkung von Ozon auf die korperliche
30             Leistungsfahigkeit [The effect of ozone on physical activity]. Schweiz Z. Sportmed.  36: 5-10.
31       Linn, W. S.; Jones, M. P.; Bachmayer, E. A.; Spier, C. E.; Mazur, S. F.; Avol, E. L.; Hackney, J. D. (1980)
32             Short-term respiratory effects of polluted ambient air: a laboratory study of volunteers in a high-oxidant
33             community. Am.  Rev. Respir. Dis. 121: 243-252.
34       Linn, W. S.; Fischer, D. A.; Medway, D. A.; Anzar, U. T.; Spier, C. E.; Valencia, L. M.; Venet, T. G.;  Hackney,
35             J. D. (1982a) Short-term respiratory effects of 0.12 ppm ozone exposure in volunteers with chronic
36             obstructive pulmonary disease. Am. Rev. Respir. Dis. 125: 658-663.
37       Linn, W. S.; Medway, D. A.; Anzar, U. T.; Valencia, L. M.;  Spier, C. E.; Tsao, F.  S.-D.; Fischer, D. A.; Hackney,
38             J. D. (1982b) Persistence of adaptation to ozone in volunteers exposed repeatedly for six weeks. Am.  Rev.
39             Respir. Dis. 125:491-495.
40       Linn, W. S.; Shamoo, D. A.; Venet, T. G.; Spier, C. E.; Valencia, L. M.; Anzar, U. T.; Hackney, J. D. (1983a)
41             Response to ozone in volunteers with chronic obstructive pulmonary disease. Arch. Environ. Health
42             38:278-283.
43       Linn, W. S.; Avol, E. L.; Hackney, J. D. (1983b) Effects of ambient oxidant pollutants on humans: a movable
44             environmental chamber study. In: Lee, S. D.; Mustafa, M.  G.; Mehlman, M. A., eds. International symposium
45             on the biomedical effects of ozone and related photochemical oxidants; March 1982; Pinehurst,  NC.
46             Princeton, NJ: Princeton Scientific Publishers, Inc.; pp. 125-137. (Advances in modern environmental
47             toxicology: v. 5).
48       Linn, W. S.; Avol, E. L.; Shamoo, D. A.; Spier, C. E.; Valencia, L. M.; Venet, T. G.; Fischer, D. A.; Hackney, J. D.
49             (1986) A dose-response study of healthy, heavily exercising men exposed to ozone at concentrations near the
50             ambient air quality standard. Toxicol. Ind. Health 2: 99-112.
51       Linn, W. S.; Gong, H., Jr.; Shamoo, D. A.; Anderson, K. R.; Avol, E. L. (1997) Chamber exposures of children to
52             mixed ozone, sulfur dioxide, and sulfuric acid.  Arch. Environ. Health 52: 179-187.
53       Liu, L.; Leech, J.  A.; Urch, R. B.; Silverman, F. S. (1997) In vivo salicylate hyroxylation: a potential biomarker for
54             assessing acute ozone exposure and effects in humans. Am. J. Respir. Crit. Care Med. 156:  1405-1412.
55       Liu, L.; Leech, J.  A.; Urch, R. B.; Poon, R.; Zimmerman, B.; Kubay, J. M.; Silverman, F. S. (1999) A comparison of
56             biomarkers of ozone exposure in human plasma, nasal lavage, and sputum. Inhalation Toxicol. 11: 657-674.
         August 2005                                    6-50         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Marthan, R.; Roux, E.; Savineau, J.-P. (1996) Human bronchial smooth muscle responsiveness after in vitro
  2             exposure to oxidizing pollutants.  Cell Biol. Toxicol. 12: 245-249.
  3       McBride, D. E.; Koenig, J. Q.; Luchtel,  D. L.; Williams, P. V.; Henderson, W. R., Jr. (1994) Inflammatory effects of
  4             ozone in the upper airways of subjects with asthma. Am. J. Respir. Crit. Care Med. 149: 1192-1197.
  5       McDonnell, W. F. (1996) Individual variability in human lung function responses to ozone exposure. Environ.
  6             Toxicol. Pharmacol. 2: 171-175.
  7       McDonnell, W. F.; Horstman, D. H.; Hazucha, M. J.; Seal, E., Jr.; Haak, E. D.; Salaam, S. A.; House, D. E. (1983)
  8             Pulmonary effects of ozone exposure during exercise: dose-response characteristics. J. Appl. Physiol.: Respir.
  9             Environ. Exercise Physiol. 54: 1345-1352.
10       McDonnell, W. F., Ill; Chapman, R. S.; Leigh, M. W.; Strope, G. L.; Collier, A. M. (1985a) Respiratory responses
11             of vigorously exercising children to 0.12 ppm ozone exposure. Am. Rev. Respir. Dis. 132: 875-879.
12       McDonnell, W. F., Ill; Horstman, D. H.; Abdul-Salaam, S.; House, D. E. (1985b) Reproducibility of individual
13             responses to ozone exposure. Am. Rev. Respir. Dis. 131: 36-40.
14       McDonnell, W. F.; Kehrl, H. R.; Abdul-Salaam, S.; Ives, P. J.; Folinsbee, L. J.; Devlin, R. B.; O'Neil, J. J.;
15             Horstman, D. H. (1991) Respiratory response of humans exposed to low levels of ozone for 6.6 hours. Arch.
16             Environ. Health 46: 145-150.
17       McDonnell, W. F.; Muller, K. E.; Bromberg, P. A.; Shy, C. M. (1993) Predictors of individual differences in acute
18             response to ozone exposure. Am. Rev. Respir. Dis.  147: 818-825.
19       McDonnell, W. F.; Stewart, P. W.;  Andreoni, S.; Seal, E.,  Jr.; Kehrl, H. R.; Horstman, D. H.; Folinsbee, L. J.; Smith,
20             M. V. (1997) Prediction of ozone-induced FEVj changes: effects of concentration, duration, and ventilation.
21             Am. J. Respir. Crit. Care Med. 156: 715-722.
22       McDonnell, W. F.; Stewart, P. W.;  Smith, M. V.; Pan, W.  K.; Pan, J. (1999) Ozone-induced respiratory symptoms:
23             exposure-response models and association with lung function. Eur. Respir. J. 14: 845-853.
24       McKenzie, D. C.; Stirling, D. R.; Fadl, S.; Allen, M. (1987) The effects of salbutamol on pulmonary function in
25             cyclists exposed to ozone: a pilot study. Can. J.  Sport Sci. 12: 46-48.
26       Melton, C. E. (1990) Airliner cabin ozone: an updated review. Washington, DC: Federal Aviation Administration,
27             Office of Aviation Medicine; report no. DOT/FAA/AM-89/13. Available from: NTIS, Springfield, VA;
28             AD-A219  264.
29       Messineo, T. D.;  Adams, W. C. (1990) Ozone inhalation effects in females varying widely in lung  size: comparison
30             with males. J. Appl. Physiol. 69:  96-103.
31       Michelson, P. H.; Dailey, L.; Devlin, R. B.; Peden, D. B. (1999) Ozone effects on the immediate-phase response to
32             allergen in the nasal airways of allergic asthmatic subjects. Otolaryngol. Head Neck Surg. 120: 225-232.
33       Mohammed, S. P.; Higenbottam, T. W.; Adcock, J. J. (1993) Effects of aerosol-applied capsaicin, histamine and
34             prostaglandin-E2 on airway sensory  receptors of anaesthetized cats. J. Physiol. Lond. 469: 51-66.
35       Molfino, N. A.; Wright, S. C.;  Katz, L; Tarlo, S.; Silverman, F.; McClean, P. A.; Szalai, J. P.; Raizenne, M.; Slutsky,
36             A. S.; Zamel, N. (1991) Effect of low concentrations of ozone on inhaled allergen responses in asthmatic
37             subjects. Lancet 338(8761):  199-203.
38       Mudway, I. S.; Kelly, F. J. (2000) Ozone and the lung: a sensitive issue. Mol. Aspects. Med. 21:  1-48.
39       Mudway, I. S.; Kelly, F. J. (2004) An investigation of inhaled ozone dose and the magnitude of airway inflammation
40             in healthy  adults. Am. J. Respir. Crit. Care Med. 169: 1089-1095.
41       Mudway, I. S.; Blomberg, A.; Frew, A.  J.; Holgate, S. T.;  Sandstrom, T.; Kelly, F. J. (1999) Antioxidant
42             consumption and repletion kinetics in nasal lavage fluid following exposure of healthy human volunteers to
43             ozone. Eur. Respir. J.  13: 1429-1438.
44       Mudway, I. S.; Stenfors, N.; Blomberg,  A.; Helleday, R.; Dunster, C.; Marklund, S. L.; Frew, A. J.; Sandstrom, T.;
45             Kelly, F. J. (2001) Differences in basal airway antioxidant concentrations are not predictive  of individual
46             responsiveness to ozone: a comparison of healthy and mild asthmatic subjects. Free Radical Biol. Med.
47             31:962-974.
48       Nagda, N. L.; Fortmann, R. C.; Koontz, M. D.; Baker, S. R.; Ginevan, M. E. (1989) Airliner cabin environment:
49             contaminant measurements, health risks, and mitigation options. Washington, DC: U.S. Department of
5 0             Transportation, Office of the Secretary. Available from: NTIS, Springfield, VA; PB91 -1593 84.
51       Nayak, A. S. (2003) The asthma and allergic rhinitis link.  Allergy Asthma Proc. 24: 395-402.
52       Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
53             pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol.
54             141: 111-122.
         August 2005                                    6-51         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Newson, E. I; Krishna, M. T.; Lau, L. C. K.; Howarth, P. H.; Holgate, S. T.; Frew, A. J. (2000) Effects of
  2             short-term exposure to 0.2 ppm ozone on biomarkers of inflammation in sputum, exhaled nitric oxide, and
  3             lung function in subjects with mild atopic asthma. J. Occup. Environ. Med. 42: 270-277.
  4       Nightingale, J. A.; Rogers, D. F.; Chung, K. F.; Barnes, P. J. (2000) No effect of inhaled budesonide on the response
  5             to inhaled ozone in normal subjects. Am. J. Respir. Crit. Care Med. 161: 479-486.
  6       Nikasinovic, L.; Momas, L; Seta, N. (2003) Nasal epithelial and inflammatory response to ozone exposure: a review
  7             of laboratory-based studies published since 1985. J. Toxicol. Environ. Health B 6: 521-568.
  8       Otto-Knapp, R.; Jurgovsky, K.; Schierhorn, K.; Kunkel, G. (2003) Antioxidative enzymes in human nasal mucosa
  9             after exposure to ozone. Possible role of GSTM1 deficiency. Inflamm. Res. 52: 51-55.
10       Passannante, A.  N.; Hazucha, M. J.; Bromberg, P. A.; Seal, E.; Folinsbee, L.; Koch, G. (1998) Nociceptive
11             mechanisms modulate ozone-induced human lung function decrements. J. Appl. Physiol.  85: 1863-1870.
12       Peden, D. B. (2001a) Air pollution in asthma: effect of pollutants on airway inflammation. Ann. Allergy Asthma
13             Immunol. 87(suppl. 3): 12-17.
14       Peden, D. B. (200 Ib) Effect of pollutants in rhinitis. Curr. Allergy Asthma Rep. 1: 242-246.
15       Peden, D. B.;  Setzer, R. W., Jr.; Devlin, R. B. (1995) Ozone exposure has both a priming effect on allergen-induced
16             responses and an intrinsic inflammatory action in the nasal airways of perennially allergic asthmatics. Am. J.
17             Respir. Crit. Care Med. 151: 1336-1345.
18       Peden, D. B.;  Boehlecke, B.; Horstman, D.; Devlin, R. (1997) Prolonged acute exposure to 0.16 ppm ozone induces
19             eosinophilic airway inflammation in asthmatic  subjects with allergies. J. Allergy Clin. Immunol. 100:
20             802-808.
21       Riediker, M.;  Monn, C.; Koller, T.; Stahel, W. A.; Wuthrich, B. (2001) Air pollutants enhance rhinoconjunctivitis
22             symptoms in pollen-allergic individuals. Ann. Allergy Asthma Immunol. 87: 311-318.
23       Rigas, M. L.; Ben-Jebria, A.; Ultman, J. S. (1997) Longitudinal distribution of ozone absorption in the lung: effects
24             of nitrogen dioxide, sulfur dioxide, and ozone exposures. Arch. Environ. Health 52: 173-178.
25       Romieu, L; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. L; Estela del
26             Rio-Navarro, B.; Hernandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and
27             antioxidant supplementation influence lung function in relation to ozone exposure in asthmatic children in
28             Mexico City. Thorax 59: 8-10.
29       Samet, J. M.; Hatch, G. E.; Horstman, D.; Steck-Scott, S.; Arab, L.; Bromberg, P. A.; Levine, M.; McDonnell, W.
30             F.; Devlin, R. B. (2001) Effect of antioxidant supplementation on ozone-induced lung injury in human
31             subjects.  Am. J. Respir. Crit. Care Med.  164: 819-825.
32       Scannell, C.; Chen, L.; Aris, R. M.; Tager, I.; Christian, D.; Ferrando, R.; Welch, B.; Kelly, T.; Balmes, J. R. (1996)
33             Greater ozone-induced inflammatory responses in subjects with asthma. Am. J. Respir. Crit.  Care Med.
34             154:24-29.
35       Schelegle, E. S.; Adams, W. C. (1986) Reduced exercise time in competitive simulations consequent to low level
36             ozone exposure. Med. Sci. Sports Exercise 18:  408-414.
37       Schelegle, E. S.; Adams, W. C.;  Siefkin, A. D. (1987) Indomethacin pretreatment reduces ozone-induced pulmonary
38             function decrements in human subjects. Am. Rev. Respir. Dis. 136: 1350-1354.
39       Schelegle, E. S.; Siefkin, A. D.; McDonald, R. J. (1991) Time course of ozone-induced neutrophilia in normal
40             humans. Am. Rev. Respir. Dis. 143: 1353-1358.
41       Schelegle, E. S.; Carl, M. L.; Coleridge, H. M.; Coleridge, J. C. G.; Green, J. F. (1993) Contribution of vagal
42             afferents  to respiratory reflexes evoked by acute inhalation of ozone in dogs. J. Appl. Physiol. 74: 2338-2344.
43       Schelegle, E. S.; Eldridge, M. W.; Cross, C. E.; Walby, W. F.; Adams, W. C. (2001) Differential effects of airway
44             anesthesia on ozone-induced pulmonary responses in human subjects. Am. J. Respir. Crit. Care Med.
45             163: 1121-1127.
46       Schonfeld, B.  R.; Adams, W. C.;  Schelegle, E. S. (1989) Duration of enhanced responsiveness upon re-exposure to
47             ozone.  Arch. Environ. Health 44: 229-236.
48       Schwartz, L. W.; Dungworth, D. L.; Mustafa, M. G.; Tarkington, B. K.; Tyler, W. S. (1976) Pulmonary responses of
49             rats to ambient levels of ozone: effects of 7-day intermittent or continuous exposure. Lab. Invest.
50       Seal, E., Jr.; McDonnell, W. F.; House, D. E.; Salaam, S. A.; Dewitt, P. J.; Butler, S. O.; Green, J.; Raggio, L.
51             (1993)  The pulmonary response of white and black adults to six concentrations of ozone. Am. Rev. Respir.
52             Dis.  147:  804-810.
53       Seal, E., Jr.; McDonnell, W. F.; House, D. E. (1996) Effects of age, socioeconomic status, and menstrual cycle on
54             pulmonary response to ozone. Arch. Environ. Health 51: 132-137.
55       Solic, J. J.; Hazucha, M. J.; Bromberg, P. A. (1982) The acute effects of 0.2 ppm ozone in patients with chronic
56             obstructive pulmonary disease. Am. Rev. Respir. Dis. 125: 664-669.
         August 2005                                    6-52        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Solway, I; Leff, A. R. (1991) Sensory neuropeptides and airway function. J. Appl. Physiol. 71: 2077-2087.
  2       Spannhake, E. W.; Reddy, S. P. M; Jacoby, D. B.; Yu, X.-Y.; Saatian, B.; Tian, J. (2002) Synergismbetween
  3             rhinovirus infection and oxidant pollutant exposure enhances airway epithelial cell cytokine production.
  4             Environ.  Health Perspect. 110: 665-670.
  5       Stenfors, N.; Pourazar, J.; Blomberg, A.; Krishna, M. T.; Mudway, I.; Helleday, R.; Kelly, F. J.; Frew, A. J.;
  6             Sandstrom, T. (2002) Effect of ozone on bronchial mucosal inflammation in asthmatic and healthy subjects.
  7             Respir. Med. 96: 352-358.
  8       Tepper, J.  S.; Costa, D. L.; Lehmann, J. R.; Weber, M. F.; Hatch, G. E. (1989) Unattenuated structural and
  9             biochemical alterations in the rat lung during functional adaptation to ozone. Am. Rev. Respir. Dis.
10             140:493-501.
11       Tepper, J.  S.; Wiester, M. J.; Weber, M. F.; Menache, M. G. (1990) Measurements of cardiopulmonary response in
12             awake rats during acute exposure to near-ambient concentrations of ozone. J. Appl. Toxicol.  10: 7-15.
13       Torres, A.; Utell, M. J.; Morow, P. E.; Voter, K.  Z.; Whitin, J. C.; Cox, C.; Looney, R. J.; Speers, D. M.; Tsai, Y.;
14             Frampton, M. W. (1997) Airway inflammation in smokers and nonsmokers with varying responsiveness to
15             ozone. Am. J. Respir. Crit.  Care Med. 156: 728-736.
16       Trenga, C. A.; Koenig, J. Q.; Williams, P. V. (2001) Dietary antioxidants and ozone-induced bronchial
17             hyperresponsiveness in adults with asthma. Arch. Environ. Health 56: 242-249.
18       U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other photochemical oxidants.
19             Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
20             Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA;
21             PB87-142949.
22       U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
23             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
24             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
25             http://cfpub2.epa.gov/ncea/.
26       Ultman, J. S.; Ben-Jebria, A.; Arnold,  S. F. (2004) Uptake distribution of ozone in human lungs: intersubject
27             variability in physiologic response. Boston, MA: Health Effects Institute; research report no. 125. Available:
28             http://www.healtheffects.org/Pubs/Ultman.pdf [29 July, 2005].
29       Vagaggini, B.; Carnevali, S.; Macchioni, P.; Taccola, M.; Fornai, E.; Bacci, E.; Bartoli, M. L.; Cianchetti, S.; Dente,
30             F. L.; Di  Franco, A.; Giannini, D.; Paggiaro. P. L. (1999) Airway inflammatory response to ozone in subjects
31             with different asthma severity. Eur. Respir. J. 13: 274-280.
32       Vagaggini, B.; Taccola, M.; Conti, L; Carnevali, S.; Cianchetti, S.; Bartoli, M. L.; Bacci, E.; Dente, F. L.;
33             Di Franco, A.; Giannini,  D.; Paggiaro, P. L. (2001) Budesonide reduces neutrophilic but not functional
34             airway response to ozone in mild asthmatics. Am. J. Respir. Crit.  Care Med.  164: 2172-2176.
35       Vagaggini, B.; Taccola, M.; Cianchetti, S.; Carnevali, S.; Bartoli, M. L.; Bacci, E.; Dente, F. L.; Di Franco, A.;
36             Giannini, D.; Paggiaro, P. L. (2002) Ozone exposure increases eosinophilic airway response  induced by
37             previous  allergen challenge. Am. J. Respir. Crit. Care Med. 166: 1073-1077.
38       Voter, K. Z.; Whitin, J. C.; Torres, A.; Morrow, P. E.; Cox, C.; Tsai, Y.; Utell, M. J.; Frampton, M. W. (2001)
3 9             Ozone exposure and the production of reactive oxygen species by bronchoalveolar cells in humans. Inhalation
40             Toxicol.  13:465-483.
41       Wang, G.; Umstead, T. M.; Phelps, D. S.; Al-Mondhiry, H.; Floras, J. (2002) The effect of ozone exposure on the
42             ability of human surfactant protein A variants to stimulate cytokine production. Environ. Health Perspect.
43             110:79-84.
44       Weinmann, G. G.; Weidenbach-Gerbase, M.; Foster, W. M.; Zacur, H.; Frank, R. (1995) Evidence for
45             ozone-induced small-airway dysfunction:  lack of menstrual-cycle and gender effects. Am. J.  Respir. Crit.
46             Care Med. 152: 988-996.
47       Yang, I. A.; Holz, O.; Torres, R. A.; Magnussen,  H.; Barton, S. J.; Rodriguez, S.; Cakebread, J. A.; Holloway, J. W.;
48             Holgate,  S. T. (2005) Association of tumor necrosis factor-a polymorphisms and ozone-induced change in
49             lung function. Am. J. Respir. Crit. Care Med. 171: 171-176.
50       Yeadon, M.; Wilkinson, D.; Darley-Usmar, V.; O'Leary, V. J.; Payne, A. N. (1992) Mechanisms contributing to
51             ozone-induced bronchial hyperreactivity in guinea-pigs. Pulm. Pharmacol. 5: 39-50.
52       Ying, R. L.; Gross, K. B.; Terzo, T. S.; Eschenbacher, W. L. (1990) Indomethacin does not inhibit the
53             ozone-induced increase in bronchial responsiveness in human subjects. Am. Rev. Respir. Dis. 142: 817-821.
54       Yu, M.; Pinkerton, K. E.; Witschi, H. (2002) Short-term exposure to aged and diluted sidestream cigarette smoke
55             enhances ozone-induced lung injury inB6C3Fl mice. Toxicol. Sci. 65: 99-106.
56       Zhang, L.-Y.; Levitt, R. C.; Kleeberger, S. R. (1995) Differential susceptibility  to ozone-induced airways
57             hyperreactivity in inbred strains of mice. Exp. Lung Res. 21: 503-518.


         August 2005                                    6-53         DRAFT-DO NOT QUOTE OR CITE

-------
 i                      7.  EPIDEMIOLOGIC STUDIES OF
 2               HUMAN HEALTH EFFECTS ASSOCIATED
 3                   WITH AMBIENT OZONE EXPOSURE
 4
 5
 6     7.1    INTRODUCTION
 7          This chapter evaluates current epidemiologic literature on health and physiological effects
 8     of ambient O3 exposure. Epidemiologic studies linking community ambient O3 concentrations to
 9     health effects were reported in the 1996 Ozone Air Quality Criteria Document (O3 AQCD; U.S.
10     Environmental Protection Agency, 1996a). Many of those studies reported that pulmonary
11     function decrements, respiratory symptoms, and hospital and emergency department admissions
12     in human populations were associated with ambient levels of O3.  Numerous more recent
13     epidemiologic studies discussed in this chapter evaluate the relationship of ambient O3 to
14     morbidity and mortality, and thereby provide an expanded basis for assessment of health effects
15     associated with exposures to O3 at concentrations currently encountered in the United States.
16          As discussed elsewhere in this document (Chapters 5 and 6), a substantial amount of
17     experimental evidence links O3 exposure unequivocally with respiratory effects in laboratory
18     animals and humans. These include structural changes in the bronchiolar-alveolar transition
19     (centriacinar) region of the lung, biochemical evidence of acute cellular/tissue injury,
20     inflammation, increased frequency and severity of experimental bacterial infection,  and
21     temporary reductions in mechanical lung function. These effects have been observed with
22     exposure to O3 at ambient or near-ambient concentrations.  Thus, many of the reported
23     epidemiologic associations of ambient O3 with respiratory health effects have considerable
24     biological credibility. Accordingly, the new epidemiologic studies of ambient O3 assessed here
25     are best considered in combination with information from the other chapters on ambient O3
26     concentration and exposure (Chapter 3), and toxicological effects of O3 in animals and humans
27     (Chapters 5 and 6, respectively).  The epidemiologic studies constitute important information on
28     associations between health effects and exposures of human populations to "real-world" O3 and
29     also help to identify susceptible subgroups and associated risk factors. A wide variety of
30     oxidants in both the gaseous and particulate phases have not been examined in relation to health
       August 2005                              7-1         DRAFT-DO NOT QUOTE OR CITE

-------
 1      outcomes in the literature.  Therefore, discussion in this chapter is limited to studies of human
 2      health effects associated with ambient O3 exposure.
 3
 4      7.1.1  Approach to Identifying Ozone Epidemiologic Studies
 5           Numerous O3 epidemiologic papers have been published since completion of the 1996 O3
 6      AQCD. The U.S. Environmental Protection Agency (NCEA-RTP) has implemented a
 7      systematic approach to identify relevant epidemiologic studies for consideration in this chapter.
 8      In general, an ongoing search has been employed in conjunction with other strategies to identify
 9      O3 epidemiologic literature pertinent to developing criteria for the O3 National Ambient Air
10      Quality Standards (NAAQS).  A publication base was established using Medline, Pascal,
11      BIOSIS, NTIS, and Embase, and a set of search terms proven by prior use to identify pertinent
12      literature. The search strategy was reexamined and modified to enhance identification of
13      published papers. PubMed was added to the search regime.
14           While the above search regime provided good coverage of the relevant literature,
15      additional approaches augmented the traditional search methods.  First, a Federal Register
16      Notice was issued requesting information and published papers from the public at large. Next,
17      non-EPA chapter authors, expert in this field, identified literature on their own. NCEA-RTP
18      staff also identified publications as an element of their assessment and interpretation of the
19      literature. Finally, additional potentially relevant publications were included following external
20      review as a result of comments from both the public and CASAC. The combination of these
21      approaches is believed to produce a comprehensive collection of studies appropriate for review
22      and  assessment here. The principal objective criteria used for selecting literature for the present
23      assessment is to include all identified studies that evaluated the relationship between measured
24      ambient O3 levels and a human health outcome.  New studies accepted for publication through
25      December 2004, as identified using the approaches above, have been included in this AQCD and
26      additional efforts have been made to assess more recent studies.
27
28      7.1.2  Approach to Assessing Epidemiologic Evidence
29           Definitions of the various types of epidemiologic studies assessed have been provided in an
30      earlier PM AQCD (U.S. Environmental Protection Agency, 1996b). Briefly, epidemiologic
31      studies are generally divided into two groups, morbidity studies and mortality studies. Morbidity

        August 2005                              7-2        DRAFT-DO NOT QUOTE OR CITE

-------
 1      studies evaluate O3 effects on a wide range of health endpoints, including the following:
 2      changes in pulmonary function, respiratory symptoms, self-medication in asthmatics, and airway
 3      inflammation; changes in cardiovascular physiology/functions; and cardiopulmonary emergency
 4      department visits and hospital admissions.  Mortality studies investigate O3 effects on total
 5      (nonaccidental) mortality and cause-specific mortality, providing evidence related to a clearly
 6      adverse endpoint. The epidemiologic strategies most commonly used in O3 health studies are
 7      prospective cohort studies, ecologic studies, time-series semi-ecologic studies, and case-
 8      crossover studies. All of these are observational studies rather than experimental studies.
 9           The approach to assessing epidemiologic evidence has been stated most recently in the
10      2004 PM AQCD (U.S. Environmental Protection Agency, 2004) and is summarized here. The
11      critical assessment of epidemiologic evidence presented in this chapter is conceptually based
12      upon consideration of salient aspects of the evidence of associations so as to reach fundamental
13      judgments as to the likely causal significance of the observed associations (see Hill,  1965).  The
14      general evaluation of the strength of the epidemiologic evidence reflects consideration not only
15      of the magnitude and precision of reported O3 effect estimates and their statistical significance,
16      but also of the robustness of the effects associations. Statistical significance corresponds to the
17      allowable rate of error (Type I error) in the decision problem constructed from assuming that a
18      simple null hypothesis of no association is true. It is a conditional probability; for statistical
19      significance, typically there is a less than 0.05 chance of rejecting the null hypothesis given that
20      it is true. Robustness of the associations is defined as stability in the effect estimates after
21      considering a number of factors, including alternative models and model specifications, potential
22      confounding by copollutants, as well as issues related to the consequences of measurement error.
23           Consideration of the consistency of the effects associations, as discussed in the following
24      sections, involves looking across the results of multiple- and single-city studies conducted by
25      different investigators in different places and times.  Relevant factors are known to exhibit much
26      variation across studies,  including, for example, the presence and levels of copollutants, the
27      relationships between central measures of O3 and exposure-related factors, relevant demographic
28      factors related to sensitive subpopulations, and climatic and meteorological conditions. Thus, in
29      this case, consideration of consistency and the related heterogeneity of effects are appropriately
30      understood as an evaluation of the similarity or general concordance of results, rather than an
31      expectation of finding quantitative results within a very narrow range.

        August 2005                                 7-3         DRAFT-DO NOT QUOTE OR  CITE

-------
 1           Looking beyond the epidemiologic evidence, evaluation of the biological plausibility of the
 2      O3-health effects associations observed in epidemiologic studies reflects consideration of both
 3      exposure-related factors and dosimetric/toxicologic evidence relevant to identification of
 4      potential biological mechanisms.  Similarly, coherence of health effects associations reported in
 5      the epidemiologic literature reflects consideration of information pertaining to the nature of the
 6      various respiratory- and cardiac-related mortality and morbidity effects and biological markers
 7      evaluated in toxicologic and human clinical studies.  These broader aspects of the assessment are
 8      only touched upon in this chapter but are more fully integrated in the discussion presented in
 9      Chapter 8.
10           In assessing the relative scientific quality of epidemiologic studies reviewed here and to
11      assist in interpreting their findings, the following considerations were taken into account:

12          (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 pollutant concentrations evaluated?
13          (2)  Were the study populations well defined and adequately selected so as to allow for
                 meaningful comparisons between study groups or meaningful temporal analyses of
                 health effects results?
14          (3)  Were the health endpoint measurements meaningful and reliable, including clear
                 definition of diagnostic criteria utilized and consistency in obtaining dependent
                 variable measurements?
15          (4)  Were the statistical analyses used appropriate, and properly performed and
                 interpreted?
16          (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?
17          (6)  Were the reported findings internally consistent, biologically plausible, and coherent
                 in terms of consistency with other known facts?
18           These guidelines provide benchmarks for judging the relative quality of various studies and
19      in assessing the body of epidemiologic evidence.  Detailed critical analysis of all epidemiologic
20      studies on O3 health effects, especially in relation to all of the above questions, is beyond the
21      scope of this document. Of most importance for present purposes are those studies which
        August 2005                                7-4         DRAFT-DO NOT QUOTE OR CITE

-------
 1      provide useful qualitative or quantitative information on concentration-response relationships for
 2      health effects associated with ambient air levels of O3 likely to be encountered in the U.S. among
 3      healthy and susceptible populations.
 4
 5      7.1.3   Considerations in the Interpretation of Epidemiologic Studies of
 6              Ozone Health Effects
 7           Prior to discussing results from recent O3 epidemiologic studies, issues and questions
 8      arising from the study designs and analysis methods used in the assessment of O3 effect
 9      estimates will be briefly presented. Study design can restrict the health effect parameters that
10      can be estimated. Separate considerations need to be made for acute versus chronic effect
11      studies as well as individual versus aggregate-level analyses.  Time-series studies and panel
12      studies are most frequently conducted in air pollution epidemiologic research.  Aggregate-level
13      exposure and/or outcome data are often used in these types of studies.  Analyses using
14      administrative health outcome data (e.g., numbers of deaths and emergency hospital admissions)
15      have inherent limitations as well  as strengths (Virnig and McBean, 2001). The impact of study
16      design or the loss of information  due to aggregation largely depends upon exposure variation
17      (Sheppard et al., 2005).
18           This section mainly focuses on the topics of exposure assessment and model specification
19      in air pollution epidemiologic studies. Potential biases that may result from O3 exposure
20      measurement error, and choice of exposure index and lag period are first presented.
21      A discussion of model specification issues and potential confounding by temporal factors,
22      meteorological effects, seasonal trends, and copollutants follow.
23
24      7.1.3.1  Exposure Assessment  and Measurement Error in Epidemiologic  Studies
25           In many air pollution epidemiologic studies, especially time-series studies with
26      administrative data on mortality and hospitalization outcomes, data from central ambient
27      monitoring sites generally are used as the estimate of exposure. Personal exposures of individual
28      study participants generally are not directly observed in epidemiologic studies. The use of O3
29      concentrations from ambient monitors as surrogate measures for personal O3 exposures was
30      discussed previously in Section 3.9.  Routinely collected ambient monitor data, though readily
        August 2005                               7-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      available and convenient, may not represent true personal exposure, which includes both
 2      ambient and non-ambient source exposures.
 3           In several studies focused on evaluating exposure to O3, measurements were made in a
 4      variety of indoor environments, including homes (Lee et al., 2004), schools (Linn et al., 1996),
 5      and the workplace (Liu et al., 1995).  Indoor O3 concentrations were, in general, approximately
 6      one-tenth of the outdoor concentrations in these studies.  Few indoor sources of O3 exist,
 7      possible sources being office equipment (e.g., photocopiers, laser printers) and air cleaners.
 8      As described in Section 3.8 of this document,  O3 in the indoor environment is largely dependent
 9      on the outdoor ambient O3 concentration. Other factors that influence the O3 concentration
10      indoors include the air exchange rate, outdoor infiltration, indoor circulation rate, and O3
11      removal process.
12           Sheppard (2005) states that non-ambient exposures typically vary across individuals but
13      are not likely to have strong temporal correlations.  In contrast, ambient concentrations for
14      individuals should be highly correlated as they vary over time similarly for everyone because of
15      changes in source generation, weather, and season.  The independence of ambient and
16      non-ambient exposure sources has important implications for selection of study designs that are
17      most effective for estimating health effects (Sheppard, 2005). In an ideal situation,  studies of air
18      pollution health effects would be conducted at the individual level, with information on personal
19      exposure to the various pollutants. However,  determining accurate personal exposure
20      information is difficult and generally impractical. A simulation study by Sheppard et al. (2005)
21      examining non-reactive pollutants observed that there was no noticeable difference  between
22      effect estimates using either total personal exposure or ambient concentration data when
23      non-ambient source exposures were independent of ambient source exposures in time-series
24      studies.  Sheppard (2005) concludes that for estimating acute effects, ambient concentration
25      measurements are adequate in time-series studies.  In the case of O3, there are limited
26      non-ambient sources; thus, ambient concentrations  of O3 are also likely to be adequate in the
27      analysis of O3 health effects in time-series studies.  Even with the lower exposure variation when
28      using only ambient concentration data, the large sample sizes and longer study duration make
29      time-series studies quite powerful.
30           As discussed thoroughly in the  2004 PM AQCD (Section 8.4.5), the resulting  exposure
31      measurement error  and its effect on the estimates of relative risk must be considered. In theory,

        August 2005                               7-6         DRAFT-DO NOT QUOTE OR CITE

-------
 1      there are three components to exposure measurement error in time-series studies as described by
 2      Zeger et al. (2000): (1) the use of average population rather than individual exposure data;
 3      (2) the difference between average personal ambient exposure and ambient concentrations at
 4      central monitoring sites; and (3) the difference between true and measured ambient
 5      concentrations. Zeger et al. indicated that the first and  third error components were largely
 6      Berksonian errors; although they would increase the standard errors, they would not bias the
 7      risk estimate.  However, the second error component resulting from the difference between
 8      average personal ambient exposure and outdoor ambient concentration levels might attenuate
 9      the risk estimate.
10           The impact of exposure measurement error on O3 effect estimates was demonstrated in a
11      study by Navidi et al. (1999). In this study,  a simulation was conducted using data from the
12      University of Southern California Children's Health Study of the long-term effects of air
13      pollutants on children. The effect estimate from computed "true" O3 exposure was compared to
14      effect estimates from exposure determined using several methods: (1) ambient stationary
15      monitors; (2) the microenvironmental approach (multiply concentrations in various
16      microenvironments by time present in each microenvironment); and (3) personal sampling.
17      Effect estimates based on all three exposure measures were biased towards the null. The bias
18      that results when using the microenvironmental and personal sampling approach is due to
19      nondifferential measurement error. Use of ambient monitors to determine exposure will
20      generally overestimate true personal O3 exposure (assumes that subjects are outdoors 100% of
21      their time and not in close proximity to sources that reduce O3 levels such as NO emissions from
22      mobile sources), thus generally their use will result in effect estimates that are biased towards the
23      null.
24           Zidek (1997) notes that a statistical analysis must balance bias and imprecision (error
25      variance).  Ignoring measurement error in air pollution epidemiologic studies often will result in
26      underestimated risk estimates. In a reanalysis of the study by Burnett et al. (1994) on the acute
27      respiratory effects of ambient air pollution, Zidek et al. (1998) observed that accounting for
28      measurement error, as well as a few additional changes to the analysis, resulted in qualitatively
29      similar conclusions. However, while the original analysis by Burnett et al. found that 5% of
30      daily respiratory admissions in the summer months was attributable to O3, Zidek et al. calculated
31      that O3 was associated with a 14% increase in respiratory admissions.  Available data and

        August 2005                               7-7        DRAFT-DO NOT QUOTE OR CITE

-------
 1      analysis limit our ability to weigh the importance of uncertainty due to measurement error
 2      relative to other sources in the studies reviewed.
 3           As discussed in Section 3.9, there is suggestive evidence that ambient O3 concentrations
 4      from central monitors may serve as valid surrogate measures for aggregate personal O3
 5      exposures in time-series studies. However, using ambient concentrations to determine exposure
 6      generally overestimates true personal O3 exposures, resulting in biased descriptions of
 7      underlying concentration-response relationships.  These effect estimates, though conservative
 8      from a testing perspective, must be evaluated and used with caution as they may lead to an
 9      underestimation of the overall impact of air pollution on health effects. A better understanding
10      of the relationship between ambient concentrations and personal exposures, and the factors that
11      affect the relationship will improve the interpretation of ambient concentration-population health
12      response associations observed in epidemiologic studies.
13
14      7.1.3.2  Ozone Exposure Indices Used
15           The O3-related effect estimates for mortality and morbidity health outcomes are usually
16      presented in this document as a relative risk, or risk rate relative to a baseline mortality or
17      morbidity rate. These relative risks are based on an incremental change in exposure.
18      To enhance comparability between studies, presenting these relative risks by a uniform exposure
19      increment is needed. However, determining a standard increment is complicated by the use of
20      different O3 exposure indices in the existing health studies. The three daily O3 exposure indices
21      that most often appear in the literature are  1-h average maximum (1-h max), 8-h average
22      maximum (8-h max), and 24-h average (24-h avg) concentrations.  As levels are lower and  less
23      variable for the longer averaging times, relative risks of adverse health outcomes for a specific
24      numeric concentration range are not directly comparable across metrics.  Using the nationwide
25      distributional data for O3 monitors in U.S.  Metropolitan Statistical Areas, increments
26      representative of a low-to-high change in O3 concentrations were approximated based on annual
27      mean to 95th percentile differences (Langstaff, 2003), as follows:
28
29
30
31
32
Daily Exposure Index     Exposure Increment (ppb)
     1-h max O3                    40
     8-h max O3                    30
     24-h avg O3                    20
        August 2005                               7-8         DRAFT-DO NOT QUOTE OR CITE

-------
 1      In the following discussion sections, efforts were made to standardize the O3 excess risks using
 2      these increments, except as noted, so that risk estimates could be compared across studies. Note
 3      that in the Annex Tables, effect estimates are not standardized; results are presented in the tables
 4      as they are reported in the papers.
 5
 6      7.1.3.3  Lag Time:  Period between Ozone Exposure and Observed Health Effect
 7           Lags of exposure may reflect the distribution of effects across time in a population and the
 8      potential mechanisms of effects. The choice of lag days for the relationship between exposure
 9      and health effects depends on the hypothesis being tested and the mechanism involved in the
10      expression of the outcome. Effects can occur acutely with exposure on the same or previous
11      day, cumulatively over several days, or after a delayed period of a few days. With knowledge
12      of the mechanism of effect, the choice of lag days can be determined prior to analysis.
13      For example, one can expect cough to occur acutely after exposure with a lag of 0 or 1 day, as
14      O3  can act as a short-term irritant. However, an O3-related inflammatory response may not lead
15      to asthma exacerbation until several days later.  An asthmatic may be impacted by O3 on the first
16      day of exposure, have effects triggered further on the second day, then report to the emergency
17      room for an asthmatic attack three days after exposure. Further, within a population of
18      asthmatics, exacerbation of asthma symptoms may be observed over a period of several days,
19      since each asthmatic has individual aspects of the disease and may be affected by the exposure
20      differently depending on his/her sensitivity and disease severity.  The results from controlled
21      human studies may be useful in assessing the adequacy of lags for some respiratory health
22      outcomes.
23           Some studies attempted to examine the overall  impact of O3 through distributed lag
24      models.  Schildcrout and Heagerty (2005) compared  regression analyses using single-day versus
25      distributed lag models.  The single-day lag model calculates a risk estimate that assumes
26      dependence only on exposure from the specified day. In contrast, the distributed lag model
27      provides an estimate that is a summary measure of the cumulative distributed lag effect from all
28      included lag days. The  standard error of the cumulative sum of the individual  distributed lag
29      coefficients takes into consideration the variance-covariance of the multiple lags,  and is therefore
30      larger than the standard error of the single-day lag coefficient.  Thus, if the underlying O3-health
31      outcome relationship was  a single-day effect, then modeling the relationship with a distributed

        August 2005                               7-9        DRAFT-DO NOT QUOTE OR CITE

-------
 1      lag model would make the estimate less significant. On the other hand, if the effect of O3 on
 2      health outcomes persisted over several days, then applying a single-day lag model would result
 3      in an underestimation of the multiday effects.  The correct choice requires balancing variance
 4      and bias.
 5           As the parameters estimated from single-day lag versus multiday lag models are not the
 6      same, interpretation and comparison of these results may be difficult. When comparing the
 7      impacts of these different models, the nuance of increments used in calculating the estimates is
 8      different depending on the model. For example, an excess percent mortality risk "per 20 ppb
 9      increase in 24-h avg O3" in a distributed lag model including lag 0- through 6-days tacitly means
10      a 20 ppb increase in each of the seven days. The difference in the exposure scenarios in the
11      single-day versus multiday lag model (i.e., 20 ppb increase in one day versus several consecutive
12      days) complicates a simple comparison of risk estimates from two different models using "the
13      same increment."
14           Only a limited number of studies have hypothesized a priori the lag structure to be
15      examined. Most of the O3 time-series studies examined relatively small numbers of single-day
16      lag models, typically lags of 0 through 3 days.  Sheppard et al. (1999) notes that when
17      considering single-day lag estimates it is important to consider the effect estimate in the context
18      of the pattern of adjacent lags as these estimates contain information from the adjacent days
19      owing to serial correlation of the pollutant series.  In many cases, a  pattern of positive
20      associations across several lag days were reported. For the respiratory and cardiovascular
21      outcomes investigated, the "most significant" lags were generally 0- or 1-day lags, suggesting
22      that the majority of the single-day associations are immediate, not a random pattern in which
23      associations can be observed on any of the lags examined with equal probabilities. For example,
24      two recent meta-analyses of O3-mortality effects observed that the combined estimate from
25      0-day lag models was larger than the estimate from longer lag days (Bell et al., 2005; Levy
26      et al., 2005).
27           Bias resulting from the selection of lags has not been examined specifically for O3 effects.
28      However, the issue of lags has been investigated for PM and the results of this analysis are most
29      likely of relevance for O3. Lumley and Sheppard (2000) performed a simulation study to
30      examine model selection bias in air pollution epidemiology using PM2 5 as an example.
31      Sheppard et al. (1999; reanalysis Sheppard, 2003) had investigated  the association between

        August 2005                              7-10        DRAFT-DO NOT QUOTE OR CITE

-------
 1      asthma hospital admissions and ambient PM2 5 concentrations over an eight-year period in
 2      Seattle, WA. Note that the results from Lumley and Sheppard (2000) and Sheppard et al. (1999)
 3      were based on GAM using default convergence criteria (see Section 7.1.3.7). A negative control
 4      analysis, using simulated data with no association between PM exposure and the health outcome,
 5      and a positive control analysis, in which a specified non-zero  excess risk is added to the
 6      simulation, were performed for comparison. The bias from selection of best of seven lags
 7      (0 to 6 days) and residual seasonal confounding in the negative control analysis (median log
 8      relative risk of 0.0013) was approximately half the log relative risk estimated from the observed
 9      data (0.0027), after adjusting for season and temperature.  In the positive control model (true log
10      relative risk of 0.0083), the bias was small (median log relative risk of 0.0080). Results from
11      these simulations indicate that bias from selection of lags may be small, but of the same
12      magnitude as the estimated health impacts.
13           Selection of lag periods should depend on the hypothesis of the study and the potential
14      mechanism of the effect.  When the mechanism of the health effect is unknown, investigating the
15      association between outcome and exposure using cumulative distributed lag models may be
16      informative.  Analyzing a large number of lags and simply choosing the largest and most
17      significant results may bias the air pollution risk estimates away from the null.  Most studies
18      have shown that O3 has a fairly consistent, immediate effect on health outcomes, including
19      respiratory hospitalizations and mortality. Several studies also observed significant O3 effects
20      over longer cumulative lag periods, suggesting that in addition to single-day lags, multiday lags
21      should be investigated to fully capture a delayed O3 effect on health outcomes.  In this document,
22      discussion largely focuses on effect estimates from 0- and 1-day lags, with some consideration of
23      cumulative, multiday lag effects. It is not straightforward to compare and contrast results from
24      single-day versus multiday lag models because the parameters estimated from these models are
25      not the  same. These complications need to be taken into consideration when interpreting results
26      from various lag models.
27
28      7.1.3.4   Model Specification to Adjust for Temporal Trends and Meteorologic Effects
29           Several challenges present themselves with respect to designing and interpreting
30      time-series studies. The principal challenge facing the analyst in the daily time-series context is
31      avoiding bias due to confounding by short-term temporal factors operating over time scales from

        August 2005                               7-11         DRAFT-DO NOT QUOTE OR CITE

-------
 1      days to seasons. In the current regression models used to estimate short-term effects of air
 2      pollution, two major potential confounders need to be considered: (1) seasonal trend and other
 3      "long-wave" temporal trends; and (2) weather effects.  Both of these variables tend to predict a
 4      significant fraction of fluctuations in time-series. Unfortunately, as  O3 has strong seasonal
 5      cycles and is formed more at higher temperatures, both terms are also highly correlated with O3.
 6      The correlation of O3 with these confounding terms tends to be higher than that for PM or other
 7      gaseous pollutants. In the U.S., the mass concentration of PM25 generally does not have strong
 8      seasonal cycles like O3 because PM2 5 tends to reflect both primary emissions (throughout the
 9      year, but often higher in winter in most U.S. cities) and secondary aerosols (higher in summer).
10      Therefore, PM25 and O3 effect estimates from studies primarily designed to examine PM25 health
11      effects may not be comparable as model specifications that may be appropriate for PM2 5 may
12      not necessarily be adequate for O3.
13           An examination of recent time-series studies indicates that several types of fitting
14      approaches have been used to adjust for temporal trends and weather effects. The use of
15      parametric and nonparametric smoothers with varying  degrees of freedom per year has emerged
16      as the prevailing approach. The use of larger degrees of freedom to  adjust for potential
17      confounding by time-varying factors may inadvertently result in ascribing more effects to these
18      unmeasured potential confounders and mask the air pollution effect. Often smaller pollution
19      effect estimates are observed when more degrees of freedom are used. Currently, the degrees of
20      freedom used to adjust for temporal trends in time-series studies generally range from 4 to
21      12 degrees of freedom per year using either nonparametric or parametric smoothers.  Statistical
22      diagnostics such as Akaike's Information Criteria, residual autocorrelation, or dispersion of the
23      regression model often are used to choose or evaluate the adequacy  of the degrees of freedom for
24      temporal trend. However, these diagnostics do not guarantee "adequate"  control for temporal
25      confounding, as choosing the appropriate extent of smoothing requires prior knowledge of the
26      nature of the confounding (e.g., shape and duration of influenza epidemics).
27           The issue of model specifications to adjust for temporal trends and weather variables in
28      time-series studies was a consideration of several researchers that conducted sensitivity analyses
29      of PM data (Health Effects Institute, 2003). The sensitivity  of O3 coefficients to model
30      specifications for temporal trend adjustment has not been as well-studied. Recent multicity
31      studies examined the sensitivity of O3  coefficients to the extent of smoothing for adjustment of

        August 2005                               7-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1      temporal trends and meteorologic factors (Bell et al., 2004; Huang et al., 2005; Ito et al., 2005).
 2      Most, if not all, O3 studies used the same model specifications to estimate the excess risks for
 3      PM and other gaseous pollutants. The model specification designed to control  confounding by
 4      meteorological and temporal factors for PM may not be necessarily adequate for O3. As noted
 5      above, O3 is expected to have the strongest correlation with both temporal (seasonal) trend and
 6      weather effects. The strong annual cycle in O3 concentrations presents a unique problem in
 7      time-series analyses where time trends are fitted simultaneously with pollution and other model
 8      terms (i.e., co-adjustment). In this setting, the annual O3 cycle itself may compete with the
 9      smooth function of time to explain some of the annual, cyclic behavior in the health outcome,
10      which can result in biased effect estimates for O3 when data for all seasons are  analyzed
11      together.
12           Current weather models used in time-series analyses can be classified into: (1) quantile
13      (e.g., quartile, quintile) indicators; (2) parametric functional forms such as V- or U-shape
14      functions; and (3) parametric (e.g., natural splines) or nonparametric (e.g., locally estimated
15      smoothing splines [LOESS]) smoothing functions. More recent studies tend to use smoothing
16      functions.  While these methods provide flexible ways to fit health outcomes as a function of
17      temperature and other weather variables, there are two major issues that need further
18      examination to enable more meaningful interpretation of O3 morbidity and mortality effects.
19           The first issue is the interpretation of weather or temperature effects. Most researchers
20      agree about the morbidity and mortality effects of extreme temperatures (i.e., heat waves or cold
21      spells).  However, as extreme hot or cold temperatures, by definition, happen rarely, much of the
22      health effects occur in the mild or moderate temperature range.  Given the significant correlation
23      between O3 and temperature, ascribing the association between temperature and health outcomes
24      solely to temperature effects may underestimate the effect of O3.  The second issue is that
25      weather model specifications are fitted for year-round data in most studies.  Such models will
26      ignore the correlation structure that can change across seasons, resulting in inefficiency
27      and model mis-specification. This is particularly important for O3, which appears to change
28      its relationship with temperature as well as with other pollutants across seasons.
29           This changing relationship between O3 and temperature, as well as O3 and other pollutants
30      across seasons, and its potential implications to health effects modeling have not been examined
31      thoroughly in the time-series literature. Even the flexible smoother-based adjustments for

        August 2005                               7-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1      seasonal and other time-varying variables cannot fully take into account these complex
 2      relationships. One obvious way to alleviate or avoid this complication is to analyze data by
 3      season. While this practice reduces sample size, its extent would not be as serious as PM (which
 4      is collected only every sixth day in most locations) because O3 is collected daily, though only in
 5      warm seasons in some states. An alternative approach is to include separate O3 concentration
 6      variables for each season (by multiplying O3 concentrations by a season indicator variable).
 7           In locations where seasonal variability may be a factor, O3 effect estimates calculated using
 8      year-round data can be misleading, as the changing relationship between O3, temperature, and
 9      other pollutants across seasons may have a significant influence on the estimates. Analyses have
10      indicated that confounding from seasonal variability may be controlled effectively by stratifying
11      the data by season.
12
13      7.1.3.5  Confounding Effects of Copollutants
14           Extensive discussions on the issues related to confounding effects among air pollutants in
15      time-series studies are provided in Section 8.4.3 of the 2004 PM AQCD.  Since the general
16      issues discussed in that document are applicable to all pollutants, such discussions are not
17      repeated here. What was not discussed in the 2004 PM AQCD was the issue of changing
18      relationships among air pollutants across seasons.  Compared to other pollutants, O3 has strong
19      seasonal cycles. Ambient O3 levels are typically higher in the summer or warm season, often
20      referred to as the O3 season.  In the winter or colder months, O3 levels tend to be much lower
21      compared to the summer months.  During the winter in some urban locations, O3 mainly comes
22      from the free troposphere and can be considered a tracer for relatively clean air (i.e., cold, clear
23      air coming down from the upper atmosphere), as discussed in Chapter 3 of this AQCD. The
24      clean air is associated with the passage of cold fronts and the onset of high-pressure conditions,
25      which occur with colder temperatures.  Thus, sunny clear winter days following a high-pressure
26      system are the days when air pollution levels from primary emissions (e.g., NO2, SO2, and PM
27      from local sources) tend to be lower and O3 is relatively higher. This can lead to negative
28      correlations between O3 and the primary pollutants in the winter. As shown in Figure 3-24 in the
29      Chapter 3 Annex, the relationship between O3 and PM2 5 was U-shaped for the year-round data in
30      Fort Meade, MD. The negative PM2 5/O3 slope was in the range of O3 concentrations less than
31      30 ppb, providing supporting evidence of the aforementioned winter phenomenon. Thus, the

        August 2005                               7-14         DRAFT-DO NOT QUOTE OR CITE

-------
 1      correlation between O3 and PM for year-round data may be misleading.  The high reactivity of
 2      O3 with certain copollutants further complicates the analysis.  For example, the reaction between
 3      NO (emitted from motor vehicles) and O3 results in reduced O3 levels but increased NO2 levels
 4      during high traffic periods.
 5           Multipollutant regression models often are used to assess potential confounding by
 6      copollutants; however, there are limitations to these models. Zidek et al. (1996) examined,
 7      through simulation, the joint effects of multicollinearity and measurement error in a Poisson
 8      regression model. The results illustrated the transfer of effects from the "causal" variable to the
 9      confounder.  However, in order for the confounder to have a larger effect size than the true
10      predictor, the correlation between the two covariates had to be very high (r > 0.9), with moderate
11      error (a > 0.5) for the true predictor and no error for the confounder in their scenarios. The
12      transfer-of-causality effect was lessened when the confounder also became subject to error.
13      Another interesting finding was the behavior of the standard errors of the coefficients. When the
14      correlation between the covariates was high (r = 0.9) and both covariates had no error, the
15      standard errors for both coefficients were inflated by a factor of two; however, this phenomenon
16      disappeared when the confounder had error. The effect of multicollinearity is generally even
17      more complex when analyzing real data.  For further discussion, see the 2004 PM AQCD
18      (Sections 8.4.3 and 8.4.5).
19           Uncertainty remains as to the use of multipollutant regression models to assess the
20      independent health effects of pollutants that are correlated.  Particularly in the case of O3,
21      concern remains as to whether multipollutant regression models for year-round data can adjust
22      for potential confounding adequately  due to the changing relationship between O3 and other
23      pollutants. Despite these limitations,  multipollutant models are still  the prevailing approach in
24      most, if not all, studies of O3 health effects and serve as an important tool in addressing the issue
25      of confounding by copollutants, especially in season-stratified analyses.
26
27      7.1.3.6  Model  Uncertainty from Multiple Hypothesis Testing
28           Epidemiologic studies that investigated the association between various measures of O3
29      (multiple lags, different metrics, etc.)  and various health outcomes often found significant
30      effects. A major question is: Are these significant associations an artifact of model selection
31      due to multiple testing and does this lead to overestimation  of the  effect estimates?

        August 2005                               7-15         DRAFT-DO NOT QUOTE OR CITE

-------
 1           Multiple testing occurs when multiple health outcomes are examined, several lags are
 2      tested, different metrics of O3 exposure are used, and many sub-populations are tested.
 3      Statistically testing a null hypothesis (i.e., there is no effect of O3) requires one to calculate the
 4      value of a test statistic (i.e., t-value). If the observed test statistic exceeds a critical value
 5      (oftentimes the 95th percentile) or is outside a range of values, the null hypothesis is rejected.
 6      However, when multiple testing is done using a critical value determined for a single test, the
 7      chance that at least one of the hypotheses is significant may be greater than the expected 5%
 8      error rate. This uncertainty clouds the interpretation and weakens the evidence against any of
 9      the null hypotheses.  Still, multiple hypotheses testing may be of great value.  For example,
10      developing a few hypotheses a priori allows researchers to explore more throughly potential
11      associations for an O3-related health effect. Sensitivity analyses, which are critical for model
12      validation, also involve multiple testing.  There are two types of sensitivity testing. One tests for
13      the consistency of the effect when different adjustments are made  for seasonal effects, or other
14      covariates. Another tests for sensitive subpopulations and other specific conditions.  In the
15      former case, one should guard against a multiple testing error by restricting the inferences to
16      consistency of the effect and not treat the hypotheses generated for sensitivity analyses as being
17      confirmatory.
18           Recent attention has focused on Bayesian model averaging as a method to address model
19      uncertainty from multiple hypothesis testing.  In Bayesian model averaging, predictions and
20      inferences are based on a set of models rather than a single model, and each model contributes
21      proportionally to the  support it receives from the observed data (Clyde, 1999). In addition to the
22      uncertainty of effect estimation, Bayesian model averaging can incorporate uncertainty regarding
23      the choice of confounding variables, pollutants, and lags. Koop and Tole (2004) used Bayesian
24      model averaging to analyze the effect of various air pollutants, including O3, SO2, CO, NO, NO2,
25      PM10_2 5, and PM2 5, on mortality in Toronto, Canada. The 50+ explanatory variables required the
26      fitting of an enormous number of potential models. Clyde et al. (2000) and Clyde  (2000) also
27      used Bayesian model averaging to analyze the relationship between PM  and mortality. Clyde
28      (2000) noted that Bayesian model averaging did not take into consideration factors that might
29      bias the estimated effect toward the null. For example, measurement error in the exposure
30      variables was not considered. In addition, the Poisson model (similar to many other regression
31      models) assumed that all individuals in a population had equal risks, including potentially

        August 2005                                7-16        DRAFT-DO NOT QUOTE OR CITE

-------
 1      susceptible populations such as those with respiratory illnesses and outdoor workers. While
 2      Bayesian model averaging can theoretically be used to take into account uncertainty, claims of
 3      causality based on observational studies may be highly sensitive to the choice of prior
 4      distributions and class of models under consideration (Clyde et al., 2000). Another limitation of
 5      Bayesian model averaging is that the estimated posterior effects may be diluted (i.e., result in
 6      smaller coefficients) when variables are highly correlated, as may be the case for air pollution
 7      studies (George,  1999 in comments to Hoeting et al., 1999).
 8           Additional methods to control for model uncertainty resulting from multiple hypothesis
 9      testing are by a priori deciding hypotheses that are confirmatory and exploratory, and limiting
10      the number of confirmatory tests. For example, Dominici et al. (2003) used a minimum number
11      of tests in the U.S. 90 cities study, which reduced the uncertainty associated with multiple
12      testing.  In addition, they performed sensitivity analyses to examine the consistency and
13      robustness of the effects.  Another approach is to partition the data into two sets, one for model
14      identification and a second for model confirmation.
15           The summary of health effects in this chapter is vulnerable to the errors of publication bias
16      and multiple testing. Recent studies (Bell et al., 2005; Ito et al., 2005; Lumley and Sheppard,
17      2000) have found indications of what the magnitudes of these errors might be in some instances.
18      Some researchers have used methods to protect their estimates against these errors.  Efforts have
19      been made to reduce the impact of multiple testing errors on the conclusions in this document.
20      To address multiple hypothesis testing in this chapter, emphasis will be on a priori hypotheses.
21      As identifying a priori hypotheses is difficult in the majority of the studies, the most common
22      hypotheses will be considered. For example, although many studies examined multiple single-
23      day lag models, priority would be given to the effects observed at 0- or 1-day lags rather than at
24      longer lags.  Both single- and multiple-pollutant models that include O3 will be considered and
25      examined for robustness of results.  Analyses of multiple model specifications for adjustment of
26      temporal or meteorological trends will be considered sensitivity analyses.  Sensitivity analyses
27      shall not be granted the same inferential weight as the original hypothesis-driven analysis;
28      however, these analyses will be discussed in this chapter as appropriate given their valuable
29      insights  that  may lead  scientific knowledge in new directions.
30
        August 2005                               7-17        DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.1.3.7  Impact of GAM Convergence Issue on Ozone Risk Estimates
 2           Generalized Additive Models (GAM) have been widely utilized for epidemiologic analysis
 3      of the health effects attributable to air pollution.  The impact of the GAM convergence issue was
 4      thoroughly discussed in Section 8.4.2 of the 2004 PM AQCD. Reports have indicated that using
 5      the default convergence criteria in the Splus software package for the GAM function can lead to
 6      biased regression estimates for PM and an underestimation of the standard error of the effect
 7      estimate (Dominici et al., 2002; Ramsay et al., 2003). GAM default convergence criteria has a
 8      convergence precision of 10~3 and a maximum number of 10 iterations. The more stringent
 9      convergence criteria refers to increased  stringency of both the convergence precision and
10      number of iterations. The default convergence criteria was found to be a problem when the
11      estimated relative risks were small and two or more nonparametric smoothing curves were
12      included in the GAM (Dominici et al., 2002). The magnitude and direction  of the bias depend in
13      part on the concurvity of the independent variables in the GAM and the magnitude of the risk
14      estimate.  Recent focus has been on the influence of the GAM function on effect estimates for
15      PM.  However, because O3 covaries more strongly with both weather and time factors than does
16      PM, the issue of GAM convergence criteria for O3 also needs to be considered.
17           A meta-analysis by Stieb et al. (2003) found some difference in O3-mortality risk estimates
18      between the GAM studies and non-GAM studies. GAM studies were defined as studies that
19      analyzed effect estimates using nonparametric smoothing functions of time or weather.
20      Non-GAM studies were all other studies, including those using Generalized Linear Models
21      (GLM) and Generalized Estimating Equations (GEE) in their analysis. In the single-pollutant
22      models, the O3-mortality risk estimates for the non-GAM studies (10 estimates) and GAM
23      studies (15 estimates) were 1.8% (95% CI:  0.5, 3.1) and 2.2% (95% CI:  1.4, 2.8), respectively,
24      per 40 ppb daily 1-h max O3. In the multipollutant models, the pooled risk estimate was 1.0%
25      (95% CI: -0.5, 2.6) for non-GAM studies (7 estimates) and 0.5% (95% CI: -1.0, 1.9) for GAM
26      studies (4 estimates).
27           Results from recent meta-analyses of O3-mortality effects suggest that there are no
28      substantial differences between GAM-affected estimates and non-GAM-affected estimates (Bell
29      et al., 2005; Ito et al., 2005; Levy et al.,  2005). GAM-affected studies included those that used
30      default convergence criteria. Non-GAM-affected studies included GAM  studies that used
31      stringent convergence criteria or those that used other modeling techniques.  Ito et al. (2005)

        August 2005                              7-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1      found that the single-pollutant combined estimate for the GAM-affected studies (15 estimates)
 2      and non-GAM-affected studies (28 estimates) were 1.92% (95% CI:  1.02, 2.81) and 1.40%
 3      (95% CI: 0.78, 2.02), respectively, per 20 ppb increase in 24-h avg O3. In the analysis by Levy
 4      et al. (2005), the single-pollutant combined estimate for the GAM-affected studies (29 estimates)
 5      and non-GAM-affected studies (17 estimates) were 1.56% (95% CI:  1.01,2.11) and 1.80%
 6      (95% CI: 1.17, 2.43), respectively, per 40 ppb increase in 1-h max O3. Bell et al. (2005) also
 7      reported that the pooled estimate was larger for the studies that were not GAM-affected.
 8           A few GAM studies reanalyzed O3 risk estimates using more stringent convergence criteria
 9      or GLM. Reanalysis of an asthma hospital admissions study in Seattle, WA (Sheppard et al.,
10      1999; reanalysis Sheppard, 2003) indicated that there were only slight changes in the risk
11      estimates when using more stringent convergence precision (10~8) in GAM. The original GAM
12      analysis indicated an excess risk of 9% (95% CI:  3, 17) whereas the stringent GAM analysis
13      found an excess risk of 11% (95% CI: 3, 19) per 30 ppb increase in 8-h max O3 at a 2-day lag.
14      Similar results were found using GLM with natural splines, 11% (95% CI: 2, 20). In the
15      reanalysis of Santa Clara County, CA data, Fairley (1999; reanalysis Fairley, 2003) used the
16      same methods as  the original analysis except the convergence precision (e) was increased from
17      10~4 to 10~12 and the maximum number of iterations (M) were increased from 10 to 107.  The
18      O3-mortality risk  estimate slightly increased from 2.8% (95% CI not provided) using default
19      GAM parameters to 2.9% (95% CI:  -0.3, 6.0) using stringent GAM parameters per 30 ppb
20      increase in 8-h max O3 at a 0-day lag.  The O3-mortality risk  estimates further increased to 3.0%
21      (95% CI: -0.3, 6.3) using GLM with natural cubic splines.  In the reanalysis of the Netherlands
22      data by Hoek et al. (2000; reanalysis Hoek, 2003), the O3 nonaccidental mortality risk estimates
23      increased from 1.3% (95% CI: 0.8, 1.9) using default GAM to 1.5% (95% CI:  1.0, 2.1) using
24      stringent GAM (e = 10~8, M = 103) and 1.6% (95% CI: 0.9, 2.4) using GLM with natural splines
25      per 30 ppb increase in 8-h avg O3 (12 p.m. - 8 p.m.) at a 1-day lag.
26           In the limited number of studies that have reanalyzed O3 risk estimates, there is little
27      evidence that default GAM analyses resulted in positively biased estimates as observed for PM.
28      Generally it appears that the use of default convergence criteria in GAM tends to bias risk
29      estimates towards the null, in addition to underestimating the standard errors.  However, one
30      study by Cifuentes et al. (2000) in Santiago, Chile observed a large difference in the O3-
31      mortality excess risks calculated using default GAM (0.9% [95%CI:  0.2, 1.6] per 40 ppb

        August 2005                              7-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1      increase in 1-h max O3) and GLM (0.1% [95% CI: -0.6, 0.8]). The GAM convergence problem
 2      appears to vary depending on data sets, and likely depends upon the intercorrelation among
 3      covariates and the magnitude of the risk estimate; thus, its impact on the results of individual
 4      studies cannot be known without a reanalysis.  In uniformity with the approach used in the 2004
 5      PM AQCD, the results from studies that analyzed data using GAM with default convergence
 6      criteria and at least two nonparametric smoothing terms are generally not considered in this
 7      chapter, with some exceptions as noted.
 8
 9      7.1.4  Approach to Presenting Ozone Epidemiologic Evidence
10           To produce a thorough appraisal of the evidence, key information (including study design,
11      analysis, mean O3 concentrations, and health outcome results) from important new studies is
12      presented in summary tables in Chapter 7 of the Annex. Each section of the chapter starts by
13      concisely highlighting important  points derived from the 1996 O3 AQCD assessment. In the
14      main body of the chapter, particular emphasis is focused on  studies and analyses that provide
15      pertinent information for the critical assessment of health risks from O3 exposure. Not all studies
16      should be accorded equal weight in the overall interpretive assessment of evidence regarding
17      O3-associated health effects. Among well-conducted studies with adequate control for
18      confounding, increasing scientific weight should be accorded in proportion to the precision of
19      their effect estimates. Small-scale studies without a wide range of exposures generally produce
20      less precise estimates compared to larger studies with a broad exposure gradient.  The size of the
21      study, as indicated by the length of the study period and total number of events, and the
22      variability of O3 exposures are important components of the precision of the health effect
23      estimates. More weight should be accorded to estimates from studies with narrow confidence
24      bands.
25           Emphasis is placed on text discussion of (1) new multicity studies that employ
26      standardized methodological analyses for evaluating O3 effects across several or numerous cities
27      and often provide overall effect estimates based on combined analyses of information pooled
28      across multiple cities; (2) studies that consider O3 as a component of a complex mixture of air
29      pollutants including PM and other gaseous criteria pollutants (CO, NO2, SO2); and (3) North
30      American studies conducted in the U.S. or Canada. Multicity studies are of particular interest
31      and value due to their evaluation  of a wider range of O3 exposures and large numbers of

        August 2005                              7-20        DRAFT-DO NOT QUOTE OR CITE

-------
 1      observations. They generally provide more precise effect estimates than most smaller scale
 2      studies of single cities. Compared to meta-analyses of multiple "independent" studies, a
 3      potential advantage of multicity studies is consistency in data handling and model specifications
 4      which eliminates variation due to analysis approach. Also, unlike meta-analyses, they do not
 5      suffer from potential omission of nonsignificant results due to "publication bias." Furthermore,
 6      geographic patterns of air pollution effects have the potential to provide especially valuable
 7      evidence regarding relative homogeneity and/or heterogeneity of O3 health effects relationships
 8      across geographic locations.  Due to the potential for confounding by copollutants, preference is
 9      given to studies with effect estimates from multipollutant models, i.e., models with both O3 and
10      PM rather than O3-only models. The potential impacts of different health care systems and the
11      underlying health status of populations also need to be accounted for in the assessment
12      (Hubbell et al., 2005; Levy et al., 2001); thus, U.S.  studies are emphasized over non-U.S.
13      studies. In accordance to the emphasis placed on the O3 epidemiologic studies in this chapter,
14      the tables in the Chapter 7  Annex were organized by region with multicity studies in each region
15      presented first.
16           In the coming sections, field/panel studies and studies of emergency department visits and
17      hospital admissions, which contributed to the establishment of the revised 1997 NAAQS for O3,
18      are presented first. This is followed by a discussion of O3-related mortality and effects of
19      chronic exposures to O3. The chapter ends with an  integrative discussion providing a summary
20      and conclusions.
21
22
23      7.2   FIELD STUDIES ADDRESSING ACUTE  EFFECTS OF OZONE
24      7.2.1  Summary of Key Findings on Field Studies of Acute Ozone Effects
25             From  the 1996 O3 AQCD
26           In the 1996 O3 AQCD, individual-level camp  and exercise studies provided useful
27      quantitative information on the concentration-response relationships linking human lung
28      function declines with ambient O3 concentrations. The available body of evidence supported a
29      dominant role of O3 in the  observed lung function decrements. Extensive epidemiologic
30      evidence of pulmonary function responses to ambient O3 came from camp studies.  Six studies
31      from three separate research groups provided a combined database on individual

        August 2005                              7-21        DRAFT-DO NOT QUOTE OR CITE

-------
 1      concentration-response relationships for 616 children (mostly healthy, nonasthmatic) ranging in
 2      age from 7 to 17 years, each with at least six sequential measurements of FEVj (forced
 3      expiratory volume in 1 second) while attending summer camps (Avol et al., 1990; Higgins et al.,
 4      1990; Raizenne et al., 1987, 1989; Spektor et al., 1988a, 1991). In the combined reanalysis by
 5      Kinney et al. (1996a) using consistent analytical methods, these data yielded an average
 6      relationship between afternoon FEVj  and concurrent-hour O3 concentration of -0.50 mL/ppb
 7      (95% CI:  -0.63, -0.36), with study-specific slopes ranging from -1.29 to -0.19 mL/ppb.
 8      Exposure in camp studies usually extended for multiple hours. Although the regression results
 9      noted above were based on one-hour O3 levels, single- and multiple-hour averages were
10      observed to be highly correlated; thus, these results might represent, to some extent, the
11      influence of multihour exposures.  In  addition to the camp study results, two studies involving
12      lung function measurements before and after well-defined exercise events in adults yielded
13      concentration-response slopes of-0.4 mL/ppb (95% CI:  -0.7, -0.1) (Selwyn et al., 1985) and
14      -1.35 mL/ppb (95% CI: -2.04, -0.66) (Spektor et al., 1988b). Ozone concentrations during
15      exercise events of approximately !/2-hour duration ranged from 4 to 135 ppb in these studies.
16          Results from other field panel studies also supported a consistent relationship between
17      ambient O3 exposure and acute respiratory morbidity in the population. Respiratory symptoms
18      (or exacerbation of asthma) and decrements in peak expiratory flow (PEF) occurred with
19      increased ambient O3 concentrations,  especially in asthmatic children (Lebowitz et al.,  1991;
20      Krzyzanowski et al., 1992). The results showed greater responses in asthmatic individuals than
21      in nonasthmatics (Lebowitz et al.,  1991; Krzyzanowski et al., 1992), indicating that asthmatics
22      might constitute a sensitive group in epidemiologic studies of oxidant air pollution.  Since the
23      1996 O3 AQCD, new research has examined a broad scope of field studies which are presented
24      next.
25
26      7.2.2   Introduction to Recent Field Studies of Acute Ozone Effects
27          Numerous field studies carried out over the past decade have tested for and, in many cases,
28      observed acute associations between measures of respiratory ill-health and O3 concentrations in
29      groups of subjects (Table AX7-1 in Chapter 7 Annex). Acute  field studies are distinguished
30      from time-series study designs in that they recruit and collect data from individual human
31      subjects instead of utilizing administrative data on aggregate health outcomes such as daily

        August 2005                              7-22        DRAFT-DO NOT QUOTE OR CITE

-------
 1      mortality, hospital admissions, or emergency department visits. Although individual-level health
 2      outcome data are collected in field studies, ambient O3 concentrations from centrally located
 3      monitoring stations are generally used to assess exposure. Because of the logistical burden
 4      associated with direct data collection from individual subjects, field/panel studies tend to be
 5      small in both numbers of subjects and in duration of follow-up. While this may limit the
 6      statistical power of field studies as compared with time-series studies, the ability to determine
 7      individual-level information on health outcomes and potentially confounding factors adds
 8      scientific value.
 9           The most common outcomes measured in acute field studies on the effects of air pollution
10      exposure are lung function and various respiratory symptoms.  Other respiratory outcomes
11      examined on a limited basis include inflammation and generation of hydroxyl radicals in the
12      upper airways, and school absences.  Several studies examined cardiovascular outcomes
13      including heart rate variability (HRV) and risk of myocardial infarctions (MI). The first group
14      of studies provided varying degrees of evidence supporting the conclusion that elevated O3 levels
15      could have negative impacts on lung function and symptoms, confirming and adding to the body
16      of knowledge that was presented in the 1996 O3 AQCD.  Some emphasis has been placed in
17      examining the independent role of O3 in the presence of PM and other pollutants.  The other new
18      studies contribute information on cardiopulmonary outcomes which have not been as well
19      documented previously.
20
21      7.2.3   Acute Ozone Exposure and Lung Function
22           As discussed in the 1996 O3 AQCD and in the earlier chapter of this document on
23      controlled human exposure studies (Chapter 6), a large body of literature from clinical and field
24      studies has clearly and consistently demonstrated reversible decrements in pulmonary function
25      following acute O3 exposure. Significant O3-induced spirometric and symptom responses have
26      been observed in clinical studies of exercising healthy young adults (see Section 6.2) and in
27      some potentially susceptible subpopulations, namely asthmatics and children (see Sections 6.3.2
28      and 6.5.1). Field studies of acute O3 exposure that examine pulmonary function fall into two
29      distinct groupings, those that conduct spirometry (measuring FEVl3 FVC [forced vital capacity],
30      and other spirometric indices) and those that measure PEF using peak flow meters.  Results from
31      the previous O3 AQCD and Chapter 6 of this document support the conclusion that the

        August 2005                              7-23        DRAFT-DO NOT QUOTE OR CITE

-------
 1      spirometric parameter FEVj is a strong and consistent measure of lung function and may be used
 2      in the assessment of asthma (Fuhlbrigge et al., 2001).  PEF is a closely related but different
 3      metric of lung function. PEF measurements have been shown to be more variable than FEVj in
 4      some studies (Vaughan et al., 1989; Cross and Nelson, 1991), and can have an element of
 5      uncertain reliability when self-administered by study subjects.  However, Lippmann and Spektor
 6      (1998) state that PEF measurements from small, inexpensive flow meters, which are more
 7      feasible to use in field studies, have been shown to produce similar results to PEF measured
 8      spirometrically.
 9           Studies of FEVj will be presented first, followed by a discussion of PEF studies. Other
10      dividing aspects within these two major types of lung function studies include health status of
11      subjects  (e.g., healthy,  mildly asthmatic, severely asthmatic), age group, time spent outdoors,
12      and exertion levels. Several studies brought these factors together to produce informative data.
13      Some FEVj studies involved both increased outdoor O3 exposure and higher exertion levels.
14      The results from this group of subjects may be comparable to those from exercising subjects in
15      the clinical studies discussed in Chapter 6.
16
17      7.2.3.1   Acute Ozone Studies with Spirometry (FEVj)
18           Studies published over the past decade have provided some new insights on the acute
19      effects of O3 on FEVj.  The results  of all studies that investigated quantitative O3-related effects
20      on FEVj are summarized in the following tables. Tables 7-la,b,c present changes in FEVj
21      associated with O3 exposure in adults while Tables 7-2a,b,c present effects in children.  Tables
22      7-lb and 7-2b present the effect of O3 on FEVj measured either in the morning or afternoon;
23      Tables 7-lc and 7-2c present O3 effects on changes in FEVj across the day (afternoon
24      FEVj - morning FEVj).  Studies that did not provide quantitative O3 data were not included in
25      the tables (Cuijpers et al., 1994; Delfino et al., 2004; Frischer et al., 1997). The data presented in
26      Hoppe et al. (1995a) were further analyzed in a subsequent paper (Hoppe et al., 2003); results
27      from the latter paper are included in the tables.  In general, the O3 effect estimates showed
28      decrements for FEVj across studies, especially in children. The studies presented in the tables
29      are discussed in further detail, starting with the O3 effect on individuals with elevated exertion
30      levels and  increased exposure due to time spent outdoors, followed by its effect on other
31      potential risk groups.

        August 2005                              7-24        DRAFT-DO NOT QUOTE OR CITE

-------
          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) a
65.9-70.4b
123 (40)
O3 Index
8-h avg
1-hmax
8-h avg
!/2-h max
1-hmax
        "Range of 8-h avg concentrations is presented by Schindler et al. (2001).
        b Range of mean i/2-h max O3 concentrations on high O3 days is presented for Hoppe et al. (2003).
 1     Exercise and outdoor w orker panels
 2          The current 8-hour NAAQS for O3 was originally based on results from controlled human
 3     exposure studies, as discussed in Chapter 6. These field studies with subjects at elevated
 4     exertion levels are of particular interest due to their similarities to the human chamber studies.
 5     The majority of human chamber studies have examined the effects of O3 exposure in subjects
 6     performing continuous or intermittent exercise for variable periods of time (see Chapter 6 of
 7     this O3 AQCD).
 8          A study by Brauer and colleagues (1996) reported unusually large O3 effects on lung
 9     function among outdoor workers. This study presented O3 effects during an extended outdoor
10     exposure period combined with elevated levels of exertion. The investigators repeatedly
11     measured spirometric lung function before and after outdoor summer work shifts over 59 days
12     on a group of 58 berry pickers in Fraser Valley, British Columbia, Canada. The subjects, both
13     male and female native Punjabi-speakers, ranged in age from 10 to 69 years old, with a mean age
14     of 44 years.  Outdoor work shifts averaged 11 hours in duration. The mean 1-h max O3
15     concentration was 40.3 ppb (SD 15.2). Exertion levels were estimated using portable heart rate
16     monitors carried over a period of four or more hours by a representative subset of subjects
17     during 16 work shifts. Heart rates over the work shift averaged 36% higher than resting levels.
       August 2005
7-25
DRAFT-DO NOT QUOTE OR CITE

-------
     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 Brauer et al.
2 Brauer et al.
3 Romieu et al
(1996)b
(1996)b
. (1998) c
4 Schindler et al. (2001)
5 Romieu et al
6 Hoppe et al.
7 Romieu et al
. (1998) c
(2003) b
. (1998) c
Berry
Berry
Street
pickers,
pickers,
workers
next morning
afternoon
on placebo,
Adults who never smoked
Street
workers
on placebo,

1st phase (lag 0-1)
(lag 0)
1st phase (lagO)
Athletes, afternoon (lag 0)
Street
workers
on supplement, 1st phase
58
58
19
3912
19
43
22
-6
-5
-3
-2
-2
-1
-1
.36
.40
.55
.96
.17
.26
.25
(-8.
(-6.
(-6.
(-5.
(-3.
(-2.
(-4.
.02,
.51,
.28,
.11,
.45,
.63,
.36,
-4.70)
-4.28)
-0.82)
-0.76)
-0.89)
0.10)
1.86)
  8   Romieu et al. (1998)c

  9   Romieu etal. (1998)c

 10   Romieu et al. (1998)c

 11   Hoppe et al. (2003)

 12   Romieu et al. (1998)c

 13   Hoppe et al. (2003)b

 14   Hoppe et al. (2003)b

 15   Hoppe et al. (2003)b

 16   Hoppe et al. (2003)b

 17   Hoppe et al. (2003)

 18   Romieu et al. (1998)c


 19   Hoppe et al. (2003)

 20   Hoppe et al. (2003)

 21   Hoppe et al. (2003)
(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

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 1)                         41

Elderly, afternoon (lag 2)                         41
                    -0.53 (-2.08, 1.01)

                    -0.40 (-1.94, 1.14)

                    -0.36 (-2.93, 2.20)

                    -0.22 (-3.86, 3.42)

                    0.18 (-0.72, 1.08)

                    0.24 (-0.64, 1.12)

                    0.48 (-0.97, 1.94)

                    0.62 (-0.45, 1.68)

                    0.71 (-0.65,2.07)

                    0.75 (-2.08, 3.58)

                    0.82 (-0.77, 2.42)


                    1.16 (-1.26, 3.58)

                    1.82 (-2.19, 5.84)

                    2.88 (-0.24, 6.00)
 aChange 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-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 FEVl for 40 year old Mexican-American
  males by Hankinson et al., 1999).
August 2005
                    7-26
DRAFT-DO NOT QUOTE OR CITE

-------
            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).
1     Post-shift FEVj and FVC decreased as a function of O3 concentration and the effects of O3
2     remained significant after adjusting for PM2 5 in the analysis.  Declines in lung function also
3     were observed on the morning following high O3 exposure. The effects seen in this study are
4     larger than have been reported previously in studies with briefer exposure durations. For
5     example, afternoon FEVj was 3.8 mL (95% CI:  -4.6, -3.0) lower per 1 ppb increase in O3
6     concentrations, compared to the decline  of 0.4 mL/ppb and 1.35 mL/ppb observed in the earlier
7     adult exercise studies (Spektor et al., 1988b; Selwyn et al., 1985).  These results are consistent
       August 2005
7-27
DRAFT-DO NOT QUOTE OR CITE

-------
          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
         a Range of mean i/i-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).
 1     with the interpretation that extended exposures to O3 produce more marked effects on lung
 2     function.  Further, when data were restricted to days with 1-h max O3 concentrations under
 3     40 ppb, the O3 effects on afternoon FEVj did not change in magnitude and remained significant.
 4     However, a possible role of copollutants cannot be completely excluded.
 5           In a Mexico City study of 47 outdoor street workers (Romieu et al., 1998), spirometry was
 6     performed repeatedly at the end of the work shift over a two-month period. Subjects were
 7     exposed to outdoor ambient O3 levels for a mean of 7.4 hours during the workday. Among those
 8     who had never taken an antioxidant supplement (subjects who received  a placebo during the
 9     first phase of the study), same day O3 concentrations were associated with decreases in FEVj.
10     A mean change of -71.6 mL (95% CI:  -113.9, -29.3) (approximately a 4% decline) was
11     observed per 40 ppb increase in 1-h max O3. The results from this study, in addition to those
12     from the Canadian study of berry pickers (Brauer et al., 1996), indicate that outdoor workers are
13     a potentially vulnerable population that may need protection from O3 exposures.
14
       August 2005
7-28
DRAFT-DO NOT QUOTE OR CITE

-------
     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    Romieu et al. (2002)b


  9    Romieu et al. (2002)b


  10   Chen etal. (1999)

  11   Ulmeretal. (1997)b

  12   Chen etal. (1999)

  13   Hoppe et al. (2003)c

  14   Linn et al. (1996)b

  15   Linn et al. (1996)b

  16   Romieu et al. (2002)b

  17   Hoppe et al. (2003)

  18   Hoppe et al. (2003)c

  19   Romieu et al. (2002)b


  20   Romieu et al. (2002)b


  21   Scarlett etal. (1996)d

  22   Romieu et al. (2002)b

  23   Hoppe et al. (2003)c

  24   Hoppe et al. (2003)
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)

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)

Children, 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)

            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)

            44     0.83 (-0.53, 2.20)
August 2005
                    7-29
DRAFT-DO NOT QUOTE OR CITE

-------
        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, afternoon (lag 1)                     44     0.93 (-0.80,2.66)

        26   Hoppe et al. (2003)     Children, morning (lag 2)                      44     1.17 (-0.36,2.70)

        27   Hoppe et al. (2003)     Children, afternoon (lag 2)                     44     1.20 (-0.12,2.52)

        28   Hoppe et al. (2003)c    Asthmatics, morning (lag 2)                    43     1.40 (-3.69,6.49)


        aChange 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).
        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)

        aChange 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 FEVl (mL). The data were transformed to percent change by dividing the
         estimates by 1,900 mL  (average FEVj among 8 to 10 year olds by Hankinson et al., 1999).
1           Hoppe et al. (1995a) examined forestry workers (n = 41) for changes in pulmonary

2      function attributable to O3 exposure in Munich, Germany. In addition, athletes (n = 43) were

3      monitored in the afternoon following a two-hour outdoor training period.  Pulmonary function

4      tests were conducted on days of both "high" (mean !/2-h max O3 of 64 to 74 ppb) and "low"

5      (mean !/2-h max O3 of 32 to 34 ppb) ambient O3 concentrations. From the average activity levels,

6      ventilation rates were estimated.  Athletes, who had a fairly high ventilation rate of 80 L/min,

7      experienced a significant decrease of 60.8 mL (95% CI: 6.4, 115.2) in FEVj per 40 ppb increase


       August 2005                                7-30        DRAFT-DO NOT QUOTE OR CITE

-------
 1      in l/2-h max O3. Among the forestry workers, a similar O3-related decline in FEVj also was
 2      observed (-56.0 mL [95% CI:  -118.4, 6.4]). In a subsequent study, Hoppe et al. (2003)
 3      reanalyzed the results of the athletes after stratifying the spirometric data by time of day
 4      (morning versus afternoon) and at different lag periods (lags of 0 to 2 days).  The reanalysis
 5      indicated that O3-related decrements were observed only with the afternoon FEVj at a 0-day lag,
 6      -1.26% (95% CI:  -2.63, 0.10) change in FEVj per 35 ppb increase in 3-h avg O3.
 7           One FEVj study clearly demonstrated small but measurable effects of multihour O3
 8      exposures on adults exercising outdoors. In Korrick et al. (1998), adult hikers (n = 530) of
 9      Mount Washington, NH performed spirometry before and after hiking for a mean of 8 hours
10      (range 2-12). The mean hourly O3  concentration ranged from 21 to 74 ppb.  After the hike, all
11      subjects combined experienced a small mean decline of 1.5% (95% CI:  0.2, 2.8) in FEVj and
12      1.3% (95% CI:  0.5, 2.1) in FVC per 30 ppb increase in the mean of the hourly O3 concentration
13      during the hike. In addition, Korrick et al. (1998) compared hikers who hiked 8 to 12 hours to
14      those who hiked 2 to 8 hours. Among those who hiked longer, the percent change in FEVj was
15      more than twofold greater per ppb exposure compared to those who hiked only for 2 to 8 hours.
16      Each hour hiked, which may reflect dose, was associated with a decline of 0.3% (p = 0.05) in
17      FEVb after adjusting for O3.
18           In a Mexico City study, the O3 effect attributable to exercise was determined using a group
19      of school children (n = 40) chronically exposed to moderate to high levels of O3 (Castillejos
20      et al., 1995).  Children were tested up to 8 times between August 1990 and October 1991.
21      Spirometry was performed by the children before and after a one-hour intermittent exercise
22      session outdoors. Outdoor O3 levels ranged up to 365 ppb, with a mean  of 112.3 ppb. Linear
23      trend analyses indicated a relationship between quintiles of O3 and percent change in lung
24      function. However, stratified analyses indicated that significant changes were observed only
25      with higher quintiles of O3 exposure (72-125 ppb and 183-365 ppb). Therefore, children
26      exercising  at higher O3 levels experienced declines in pulmonary function despite the repeated
27      daily exposure to moderate and high levels  of O3 in Mexico City.
28           Collectively, the above studies confirm  and extend clinical observations that prolonged
29      exposure periods, combined with elevated levels of exertion or exercise, may  magnify the effect
30      of O3 on lung function.  The most representative data come from the Korrick et al. (1998) hiker
31      study. This U.S. study provided outcome measures stratified by several  factors (e.g., gender,

        August 2005                              7-31        DRAFT-DO NOT QUOTE OR CITE

-------
 1      age, smoking status, presence of asthma) within a population capable of more than normal
 2      exertion.
 3
 4      Panel studies of children, elderly, and asthmatics
 5           Hoppe et al. (1995a,b) examined several potentially susceptible populations for changes in
 6      pulmonary function attributable to O3 exposure in Munich, Germany.  The forestry workers and
 7      athletes were discussed in the previous section.  Senior citizens (n = 41) and juvenile asthmatics
 8      (n = 43) were also monitored on "low" O3 and "high" O3 days.  Subjects were requested to stay
 9      outdoors for at least 2 hours just before the afternoon pulmonary function test. Clerks (n = 40)
10      were considered the nonrisk control group.  Although clerks spent the majority of their time
11      indoors, their outdoor exposures on high O3 days were similar to that of the four other risk
12      groups. The results showed no significant O3 effects on the senior citizens. Clinical studies also
13      have consistently shown that seniors are less responsive to O3 (Bedi et al., 1989;
14      Drechsler-Parks, 1995).  Asthmatics and clerks experienced slight reductions in FEVj on high O3
15      days. Among all risk groups, juvenile asthmatics experienced the largest O3-related decline in
16      FEVj, -84.0 mL (95% CI: -196.4, 28.4) per 40 ppb increase in Vi-h max O3.  To further
17      examine their hypotheses on  characteristics of O3 risk groups, Hoppe et al. (2003) conducted a
18      different analysis on a more expanded data base than utilized in the earlier study.  Children were
19      examined as an additional risk group. Hoppe et al.  (2003) presented both group mean values
20      and analyses on an individual basis.  For the group  mean values, consistent O3 effects were not
21      detectable.  On an individual basis, a potential pattern of O3 sensitivity was observed (see
22      Table AX7-1 in the Annex for details). About 20% of the children and asthmatics were regarded
23      as O3 responders (i.e., individuals with greater than 10% change in FEVj) compared to only 5%
24      of the elderly and athletes. These results indicated  that while the  majority of the population did
25      not react to O3 exposure, a small group of susceptible individuals experienced health effects
26      from O3.  The sample size limits quantitative extrapolation to larger populations, but may allow
27      cautious first estimates.
28           Several other panel studies performed spirometry in children, another potentially
29      susceptible group (Avol  et al., 1998; Chen et al.,  1999; Cuijpers et al., 1994; Frischer et al.,
30      1997; Linn et al., 1996; Romieu et al., 2002; Scarlett et al., 1996;  Ulmer et al., 1997).
31      All studies, with the exception of Avol et al. (1998) and Scarlett et al.  (1996), observed a

        August 2005                              7-32        DRAFT-DO NOT QUOTE OR CITE

-------
 1      decrease in FEVj associated with O3 exposure. One large study measured spirometric lung
 2      function in 895 school children in three towns in Taiwan (Chen et al., 1999).  Lung function was
 3      measured only once for each subject.  The authors reported significant associations between
 4      diminished FEVj and FVC with a 1-day lag of O3 concentrations.  Effect sizes were typical of
 5      those observed in past studies, i.e., 0.5 to 1.0 mL decline in FEVj per ppb increase in O3
 6      concentration. Ozone was the only air pollutant associated with changes in lung function in
 7      multipollutant models including SO2, CO, PM10, and NO2.
 8           Linn et al. (1996) repeatedly measured spirometric lung function among 269 school
 9      children in three southern California communities  (Rubidoux, Upland, and Torrance). Lung
10      function was measured over five consecutive days, once in  each of three seasons over two school
11      years. Between-week variability  was  controlled in the analysis by seasonal terms  in the model.
12      Statistical power was limited by the narrow range of exposures that were experienced within
13      each week.  In addition, the study was restricted to the school year, eliminating most of the
14      "high" O3 season from consideration.  During the study  period, 24-h avg O3 levels at the central
15      monitoring site ranged up to 53 ppb (mean 23 ppb) while personal measurements  ranged up to
16      16 ppb (mean 5 ppb).  A mean change of -11.6 mL (95% CI:  -20.6, -2.6) (approximately a 1%
17      decline) in FEVj was observed from morning to afternoon per 20 ppb increase in 24-h avg O3.
18      Other associations (involving individual morning or afternoon FVC and FEVj measurements)
19      went in the plausible direction but the O3 effect estimates were considerably smaller.
20           Ulmer et al. (1997) examined 135 children aged 8  to 11 years in two towns in Germany
21      from March to October 1994 for O3 effects on pulmonary function at four time periods. The
22      cross-sectional results at each of the four time points showed limited FVC and no  FEVj
23      associations. However, the longitudinal analysis, which combined data from all four periods
24      yielded a mean change of -87.5 mL (95% CI: -143.2, -31.7) (approximately a 5% decline)
25      in FEVj per 40 ppb increase in lA>-h max O3 for the town with the higher O3 levels  (median lA>-h
26      max of 50.6 ppb versus 32.1 ppb). In  the cross-sectional analysis, only between-person
27      variability was analyzed.  The longitudinal analysis, in which the subjects provided multiple
28      days of measurements, provided information on both between- and within-subject responses.
29           There are a limited number  of new epidemiologic  studies examining the effects of O3 on
30      FEVj; however, results from these studies indicate that acute exposure to O3 is associated with
31      declines in FEVj in children. These results further support the negative effects of O3  on lung

        August 2005                              7-33        DRAFT-DO NOT QUOTE OR CITE

-------
 1      function observed in the meta-analysis on children attending summer camp (Kinney et al.,
 2      1996a) and in the clinical literature.
 3
 4      7.2.3.2  Acute Ozone Studies of PEF
 5           Many studies of the acute effect of O3 on PEF examined self-administered PEF levels
 6      daily, both in the morning and afternoon.  PEF follows a circadian rhythm with the highest
 7      values found during the late afternoon and lowest values during the night and early morning.
 8      Due to the diurnal variation in PEF, most studies analyzed their data after stratifying by time of
 9      day.  The peak flow studies examined both asthmatic panels and healthy individuals. The
10      asthma panels are discussed first.
11
12      Asthma panels
13           The effects of acute O3 exposure on PEF in asthmatics were examined in several panel
14      studies.  Figures 7-1  and 7-2 present percent changes in morning and evening PEF outcomes
15      from seven panel studies of children, mostly asthmatic, ranging in age from 5 to 13 years.  The
16      effect estimates from all single-day and multiday lag models are presented. Only  single-city
17      results with analyses stratified by morning and afternoon are included in the figure.  Studies that
18      examined cross-day changes and daily variability in PEF (e.g., Just et al., 2002; Thurston et al.,
19      1997) are not included in the figure since such outcomes are not directly comparable.
20      Collectively, nearly all of the studies indicated decrements of peak flow but most  of the
21      individual estimates were not statistically significant.  The results from the individual studies are
22      further discussed below.
23           In Mexico City, two studies of asthmatic school  children were carried out simultaneously
24      in the northern (Romieu et al., 1996)  and southwestern sections of the city (Romieu et al., 1997).
25      In the northern study, 71 mildly asthmatic school children aged 5 to 13 years old,  were followed
26      over time for daily morning (before breakfast) and afternoon (bedtime) PEF.  In single-pollutant
27      models, O3 concentrations at 0-, 1-, and 2-day lags were associated with diminished morning and
28      afternoon PEF, but only the 0-day lag morning effect was significant. The  O3 effect became
29      nonsignificant when PM2 5 was added to the model. In the southwestern study, 65 mildly
30      asthmatic children aged 5 to 13 years old were followed during the summer and winter for daily
31      morning and afternoon PEF. Ozone concentrations at a 0- and 1-day lag were associated with

        August 2005                               7-34        DRAFT-DO NOT QUOTE OR CITE

-------
             HI
             o.
              a)
              U)
              C
             O

3 -
2-
1 -
0 -
-1 -

-2 -
-3 -
-5-
-
-7 -
ft
[Single-day Lags]



| J_
1 T
- I '
1



Y










i
i i
T




Asthmatic
Yes O
No X













| Multiday Lags]







V






















' <






<
>



(

)




)


C





) t




>










                       CD
                       -o
                       JS

                       £

                       3T
                       CO
                       o>
                       CO

                       "5
                             CD
                             II
E
CO
•g

CD
                             O
                                   CO

                                   cu
                                  o
CO
o
o
o
o
o
"x
0)
CD

c»
     CD
     II
o
o
o
"x
(U
                                                              CD  -t
                                                              10  co
      CO

      ID

      _CD


      O
      o:
o
o
o
"x
11)
           CD
           cn
           CD
      CO

      *&
      2
      o
      O
                                                         O =d  O
                            O
                            Q
                               o"
                               D)
                               co
                    O
in

csT
o
o
CM
       CO

       "CD

       CD


       '•E
       o
                                    Q
                                    •s.
                                    CD"
                                    o
                                    .E

                                    "co
                                    m
                                            _CD

                                            CO
                                            I


                                            CO
                                            LU
                                   CD
                                   O>
                                   CD
                                                                                    o
                                  CD
                                  cn
                                  CD
                                                                                    o
                                                                                    CD
      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).
1      afternoon PEF, with larger effects at a 1-day lag. Associations involving O3 were stronger than

2      those involving PM10. Several additional studies, both in the U.S. and in other countries,

3      reported significant associations between O3 exposure and decrements in PEF among asthmatics
      August 2005
         7-35
               DRAFT-DO NOT QUOTE OR CITE

-------
A

3 -
2 -
1 -
LL
HI 0 -
Q.
c -1 -
0) -2-
c
ns Q _
-C °
" -'-
0^
-5 -
-fi
w
-7 -

-8-
Single-day Lags


1
T












fO ^~
in CD
*~ II
« Z
z —
•— - CO
.>*>•.

™ O o
EDO
ci .y .a
jz ,c en X X
o- 1 ° » «
* ^~^ >*

o So
80
CO 0
'x i 'x
^.
CO , . O>
CO ^~~ O>
5 i ?
^3, O3 ^3-
•*—
— •— . —
"S ro "o>
•0 (D T3
O (fl O
                               TO
                               'ffi
                               c
                               JB
                               OJ
                               b
                                       CL
                                       Q.
                                       !O
                                       I
        '£
        o
        ce
                                                              O
             TO
             cu
                                                                             O
      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).
1     (Gielen et al., 1997; Jalaludin et al., 2000; Just et al., 2002; Ross et al., 2002; Thurston et al.,
2     1997).
      August 2005
7-36
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Other epidemiologic studies did not find a significant O3 effect on the lung function of
 2      asthmatics. Delfino et al. (1997a) examined morning and evening PEF among 22 asthmatics
 3      ranging in age from 9 to 46 years, living in Alpine, CA.  Daily ambient 12-h avg O3
 4      (8 a.m.-8 p.m.) concentrations ranged from 34 to 103 ppb, with a mean value of 64 ppb.
 5      Unique to this study, personal O3 exposures were measured using 12-h passive O3 samplers that
 6      were worn by the subjects. The personal 12-h avg O3 (8 a.m.-8 p.m.) concentrations, which had
 7      a mean value of 18 ppb, were much lower than the fixed-site ambient levels. Quantitative O3
 8      results were not reported but researchers stated that no O3 effects were observed on morning and
 9      evening PEF.  In Hiltermann et al. (1998), 60 nonsmoking adults aged 18 to 55 years in
10      Bilthoven,  the Netherlands, were followed between July and October 1995 with morning and
11      afternoon PEF measurements. Although negative associations were observed between O3  and
12      cross-day changes in PEF, the results were not significant.
13           Mortimer et al. (2002) examined 846 asthmatic children from the National Cooperative
14      Inner-City  Asthma Study (NCICAS) for O3-related changes in PEF. Children from eight urban
15      areas in the U.S. (St. Louis, MO; Chicago, IL; Detroit, MI; Cleveland, OH; Washington, DC;
16      Baltimore, MD; East Harlem, NY; and Bronx NY) were monitored from June through August
17      1993. This study provides representative data for the U.S. as children from multiple cities
18      throughout the East and Midwest were examined. Asthmatic children from urban areas are an
19      important subgroup of potentially at-risk populations. Study children either had physician-
20      diagnosed asthma and symptoms in the past 12 months or respiratory symptoms consistent with
21      asthma that lasted more than 6 weeks during the previous year.
22           Mortimer et al. (2002) examined O3-related changes in PEF for single-day lags from 1 to
23      6 days and a multiday lag period of 5 days.  Of all the pollutants examined, including O3, PM10,
24      NO2, and SO2, none were associated with evening PEF.  Only O3 was found to be associated
25      with morning PEF.  The effect estimates of the association between O3 and morning PEF for the
26      single-day  and multiday lags are depicted as error density curves in Figure 7-3 (for description of
27      error density  curves, see Annex Section AX7-2).  Small  morning effects were observed at  1- and
28      2-day lags. The effect of O3 on morning outcomes increased over several days.  The strongest
29      association between O3 and PEF was found with a multiday lag period (cumulative lag of 1 to 5
30      days). Unrestricted lag models suggested that the O3 exposure from 3 to 5 days prior had a
31      greater impact on morning % PEF than more immediate exposures. Mortimer et al. discussed

        August 2005                               7-37        DRAFT-DO NOT QUOTE OR CITE

-------
               2.0
               1.5
           g   1.0
               0.5
               0.0
                                      Single-day Lags
                                           2
                                         / xX
                      Multiday Lag 1-5    [ /   \
                                                         3/ /
      1
       \
      \\
       \  \
       \  \
       c\  \
       6 \  \
-3     -2.5      -2      -1.5     -1     -0.5      0
                           % Change in PEF
            0.5       1
                                                                                         1.5
       Figure 7-3.   Comparison of single-day lags (1-, 2-, 3-, 4-, 5-, and 6-day) to a cumulative
                     multiday lag (1- to 5-day) for percent changes in PEF per 30 ppb increase in
                     8-h avg O3 in urban children.
       Source: Derived from Mortimer et al. (2002).
 1     biological mechanisms for delayed effects on pulmonary function, which included increased
 2     bronchial reactivity secondary to airway inflammation associated with irritant exposure.  Animal
 3     toxicology and human chamber studies (see Chapters 5 and 6) provide further evidence that
 4     exposure to O3 may augment cellular infiltration and cellular activation, enhance release of
 5     cytotoxic inflammatory mediators, and alter membrane permeability.
 6           Figure 7-4 illustrates the probability density curves of the results from the city-stratified
 7     analysis and that from the pooled analysis of all eight cities. The error density curve for the
 8     all-cities analysis is a graphical  presentation of the all-cities regression analysis presented by
 9     Mortimer et al. (2002), a change in morning PEF of -1.18% (95% CI: -2.10, -0.26) per 30 ppb
10     increase in 8-h avg O3 (10 a.m.-6 p.m.) with a cumulative lag of  1 to 5 days.  The summary
11     density curve for the city-stratified analysis was calculated by summing together eight normal
12     distribution functions, one for each of the study cities, then taking the derivative of the summed
13     function (see Annex Section AX7-2 for further explanation of summary density curves).  The
14     area under the density curve and to the  left of a value on the x-axis is an estimate of the
15     probability that the effect estimate will  be less than or equal to that value. For example, the area
16     under the density curve to the left of 0% change in PEF is 99% in the all-cities analysis. A wider
       August 2005
                              7-38
DRAFT-DO NOT QUOTE OR CITE

-------
                       City-Stratified Analysis
                       All-Cities Analysis
              0.2
                Otlsi
                                                            -1
                                             % Change in PEF
       Figure 7-4. Density curves of the percent change in PEF per 30 ppb increase in 8-h avg O3
                   with a cumulative lag of 1 to 5 days for the individual eight NCICAS cities and
                   the pooled average of all 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).
 1     distribution was observed in the city-stratified analysis, with only 78% of the area less than zero.
 2     The all-cities analysis likely had a smaller standard error compared to the city-specific analysis
 3     as it was based upon more subjects and considered differences between cities to
 4     vary about the same mean effect. The regression analysis by Mortimer et al. (2002) suggested a
 5     lack of heterogeneity by city, as indicated by the nonsignificant interaction term between O3
 6     effect and city. As shown in Figure 7-4, the summary density curve of the city-stratified analysis
 7     has a peak at about the same value as the curve of the all-cities analysis,  suggesting a common
 8     O3 effect for all eight cities and small variation among them. The unimodal shape of the density
 9     curve of the city-stratified analysis also indicates the absence of outlying cities.
10          Mortimer et al. (2002) further noted that small declines in morning PEF may be of
11     uncertain clinical significance, thus they calculated the incidence of >W% declines in PEF.
12     A 5 to 15% change in FEVj has been expressed as having clinical importance to asthma
13     morbidity (American Thoracic Society, 1991; Lebowitz et al., 1987; Lippmann, 1988).
14     Although greater variability is expected in PEF measurements, a >W% change in PEF also may
15     have clinical significance.  In Mortimer et al. (2002),  O3 was associated with an increased
       August 2005
7-39
DRAFT-DO NOT QUOTE OR CITE

-------
 1      incidence of >10% declines in morning PEF (odds ratio of 1.30 [95% CI:  1.04, 1.61] per 30 ppb
 2      increase in 8-h avg O3 for a 5-day cumulative lag).  This finding suggests that exposure to O3
 3      might be related to clinically important changes in PEF in asthmatic children. This study also
 4      observed that excluding days when 8-h avg O3 levels were greater than 80 ppb provided effect
 5      estimates which were similar to those when all days were included in the analysis, indicating that
 6      the negative effect of O3 on morning PEF persisted at levels below 80 ppb. There is some
 7      concern, however, regarding the lack of an association between O3 and afternoon PEF.
 8           Results from the multicities study by Mortimer et al. (2002), as well as those from several
 9      regional studies provide evidence of a significant relationship between O3 concentrations and
10      PEF among asthmatics. Collectively, these studies indicate that O3 may be associated with
11      declines in lung function in this potentially susceptible population.
12
13      Panels of healthy subjects
14           The effect of O3 on PEF in healthy subjects also was investigated in several studies.
15      A study of 162 children (9 years of age) in England examined the relationship between O3 and
16      PEF in the winter and summer seasons (Ward et al., 2002). The O3 effect estimates were
17      generally positive in the winter and negative in the summer.  Single-day lags of 0- to 3-days
18      were examined; however, the strongest association was found with a multiday lag period.
19      During the summer, a decline of 11.10 L/min (95% CI:  0.18, 21.98) was observed in morning
20      PEF per 20 ppb increase in 24-h avg O3 with a 7-day cumulative lag.  Smaller O3 effects  were
21      observed on afternoon PEF.
22           During the summer of 1990, Neas et al. (1995) examined 83 children in Uniontown, PA
23      and reported twice daily PEF measurements. Researchers found that evening PEF was
24      associated with O3 levels weighted by hours spent outdoors.  Using a similar repeated measures
25      design, Neas et al. (1999) saw evidence for effects due to ambient O3 exposure among
26      156 children attending two summer day camps in the Philadelphia, PA area. Associations were
27      found between afternoon PEF (recorded before leaving camp) and same-day O3 concentrations,
28      and between morning PEF (recorded upon arrival at camp) and previous-day O3 concentrations.
29      However, as in the case of Ward et al. (2002), the relationship between PEF and O3 was
30      significant only when a multiday lag period was considered. Naeher et al. (1999), in a sample of
        August 2005                              7-40       DRAFT-DO NOT QUOTE OR CITE

-------
 1      473 nonsmoking women (age 19 to 43 years) living in Vinton, VA, also showed the strongest
 2      association between O3 and evening PEF with a 5-day cumulative lag exposure.
 3           Another study in southwestern Mexico City analyzed morning and afternoon PEF data
 4      collected from 40 school children aged 8 to 11 years (Gold et al., 1999).  Subjects provided
 5      measurements upon arriving and before departing from school each day. A negative effect of O3
 6      on PEF was observed, -1.60 mL/s (95% CI: -3.56, 0.36) and-1.80 mL/s (95% CI: -3.76,
 7      0.16) per 20 ppb increase in 24-h avg O3 on the same day afternoon and next day morning PEF,
 8      respectively. A greater effect was  observed for PEF regressed on O3 concentrations with a
 9      cumulative 10-day lag period (-3.50 mL/s [95% CI:  -5.52,  -1.49] on same day afternoon).
10      These results suggest a longer, cumulative effect of O3 on PEF.  Alternatively, the associations
11      observed at the 10-day lag period may reflect confounding by other time-varying factors or be a
12      chance finding from an exploratory analysis.
13           In a recent study of 43 mail carriers in Taichung City, Taiwan, PEF was monitored twice
14      daily during a six-week period (Chan and Wu, 2005). The mean 8-h avg O3 (9 a.m.-5 p.m.)
15      concentration during their work shift was 35.6 ppb (SD 12.1). Associations were observed
16      between evening PEF and O3 concentrations at lags of 0,  1 and 2 days. The greatest effect was
17      observed at a lag of 1 day, a 2.07% decline in PEF per 30 ppb increase in 8-h avg O3
18      (quantitative results for 95% CI not provided).  Similar O3 effects on morning PEF were
19      observed. The effect of O3 on PEF was robust to adjustment for copollutants; no association
20      with PEF was observed for PM10 and NO2 in multipollutant models.
21
22      7.2.4  Respiratory Symptoms
23           Studies published over the past decade represent an improved new body of data on the
24      symptom effects of O3. Respiratory symptoms in acute air pollution field studies are usually
25      measured using questionnaire forms or "daily diaries" that are filled out by study subjects,
26      usually without the direct supervision of research staff. Questions address the daily experience
27      of coughing, wheezing, shortness of breath (or difficulty breathing), production of phlegm, and
28      others.  While convenient and potentially useful in identifying acute episodes of morbidity,
29      measurements of daily symptoms are prone to a variety of errors. These include
30      misunderstanding of the meaning of symptoms, variability in individual interpretation of
31      symptoms, inability to remember symptoms if not recorded soon after their occurrence, reporting

        August 2005                             7-41        DRAFT-DO NOT QUOTE OR CITE

-------
 1      bias if days of high air pollution levels are identifiable by subjects, and the possibility of falsified
 2      data. In spite of these potential problems, the ease of data collection has made daily symptom
 3      assessment a common feature of field studies. Many of the studies reviewed above for lung
 4      function results also included measurements of daily symptoms.  Pearce et al. (1998) reports that
 5      one advantage in the case of asthma panels is that the population is usually already familiar with
 6      symptom terms such as wheezing and cough.  Delfino et al. (1998a) further states that the use of
 7      repeated daily symptom diaries has additional advantages of reducing recall bias given the
 8      proximity of events and allowing health effects to be modeled with each subject serving as their
 9      own control over time. Also, study design can blind the participants from the air pollution
10      aspect of the study. Careful efforts by study staff can help ensure that the symptom diaries will
11      provide information that is less affected by the potential problems noted.
12           Similar to studies of lung function, respiratory symptom studies can be divided into two
13      groups, asthma panels or healthy subjects.  Asthma panel studies are presented first.
14
15      Asthma panels
16           Most studies examining respiratory symptoms related to O3 exposure focused on asthmatic
17      children.  Among the health outcomes, of particular interest were those associated with asthma,
18      including cough, wheeze, shortness of breath, and increased medication use. Figures 7-5 and 7-6
19      present the odds ratios for O3-related cough and medication use among asthmatic children from
20      six studies (Gielen et al., 1997; Jalaludin et al., 2004;  Just et al., 2002; Ostro et al., 2001; Romieu
21      et al., 1996, 1997).  Only single city/region studies that present odds ratios are included in the
22      figure for consistency.  Studies that present change in severity of symptoms, another informative
23      health outcome, are not included in the figure since this symptom outcome differs from
24      indicating simple presence of symptoms.  The study by Gent et al. (2003) also is not included in
25      this figure as odds ratios for cough and mediation use were analyzed for quintiles of O3
26      concentrations using the lowest quintile as the reference.  These studies are discussed separately.
27           The various effect estimates for the association between O3 concentrations and cough are
28      depicted in Figure 7-5. Despite the variability in the individual effect estimates, there is some
29      consistency in the O3 effects. In general, the majority of the odds ratios appear to be greater than
30      one among the single-day lag models, suggesting an association between acute exposure to O3
31      and increased cough among asthmatic children.  Figure 7-6 presents the odds ratios for

        August 2005                               7-42        DRAFT-DO NOT QUOTE OR CITE

-------
 >^ >> "~^ 
                                       E
                                       o
                                       cc.
                   c
                   13
                   ^
                   ta
                   to
JB
§
b
                                ro
                                0.
            o
            o
            CM
                  0
                  C
                  T3
                  c
                  13
Figure 1-5. Odds ratios for the incidence of cough among asthmatic children per

           standardized increment (see Section 7.1.3.2). For single-day lag models,

           current day O3 effects are shown with the exception of Ostro et al. (2001)

           which only presented results from a 3-day lag. For multiday lag models, the

           cumulative effects of a 0- to 4-day lag are shown. *Note that Gielen et al.

           (1997) and Jalaludin et al. (2004) presented results for prevalence of cough.
August 2005
        7-43
DRAFT-DO NOT QUOTE OR CITE

-------
t.o
4.0 -
0)
U)
? 3.5^
_o
IS
=1 3.0-
Q)
2
§ 2.5-
LLI
%mm
g
0 2,0^
IS
DC
% 1.5-
T3
O
1.0 -
n ^ -
Single-day Lags












1












1 . . ,
1 » f
|Multiday Lags]
i
(












1












( \
                    <
                    O
o
o
      O
                          CO
                          Q-
                        ra
                        Q_
                    
                                      en
.2

o
CC
ra

"S
Z3
*
E
o
OL
            O
            o
            CM
                        o
                        o
                        CM
Figure 7-6. Odds ratios for extra medication use among asthmatic 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

           effects of a 0- to 4-day lag are shown.
August 2005
  7-44
       DRAFT-DO NOT QUOTE OR CITE

-------
 1      O3-associated bronchodilator use.  The results from medication use are less consistent than those
 2      from cough; one study by Just et al. (2002) observed strong positive associations, but had wide
 3      confidence intervals.
 4           Among the studies reporting results for daily symptoms and asthma medication use,
 5      several observed associations with O3 concentrations that appeared fairly robust (Delfino et al.,
 6      2003; Desqueyroux et al., 2002a,b; Gent et al., 2003; Hilterman et al., 1998; Just et al., 2002;
 7      Mortimer et al., 2000, 2002; Newhouse et al., 2004; Romieu et al., 1996, 1997; Ross et al., 2002;
 8      Thurston et al., 1997). Mortimer et al. (2002) reported morning symptoms in 846 asthmatic
 9      children from eight urban areas of the U.S. to be most strongly associated with a cumulative
10      4-day lag of O3 concentrations in the NCICAS.  The NCICAS used standard protocols which
11      included instructing caretakers of the subjects to record symptoms in the daily diary by
12      observing or asking the child (Mitchell et al., 1997). Symptoms reported included cough, chest
13      tightness, and wheeze. In the analysis pooling data from all eight cities, the odds ratio for the
14      incidence of symptoms was 1.35 (95% CI:  1.04, 1.69) per 30 ppb increase in 8-h avg O3
15      (10 a.m.-6 p.m.).  Excluding days when 8-h avg O3 was greater than 80 ppb, the odds ratio
16      was 1.37 (95% CI:  1.02, 1.82) for incidence of morning symptoms. Figure 7-7 presents the
17      probability density curves of the odds ratios for the incidence of symptoms from the city-
18      stratified analysis and that from  the all-cities analysis.  This figure confirms the regression
19      results that there is a significant increase in odds for incidence of symptoms, as the area under
20      the density curve with an odds ratio greater than  one is 99%. Mortimer et  al. (2002) did not
21      observe significant interactions among the eight cities, indicating that there was no heterogeneity
22      among the city-specific estimates. The unimodal distribution of the city-stratified summary
23      density curve also suggests a lack of significant heterogeneity in O3 effects among the cities.  It
24      should be noted that other pollutants, including PM10 (monitored in 3 cities), NO2 (in 7 cities),
25      and SO2 (in all 8 cities), also were associated with increased incidence of morning symptoms.
26      In multipollutant models, the O3 effect was shown to be slightly diminished.
27           Another one of the larger studies was that of Gent and colleagues (2003), where
28      271 asthmatic children under age 12 and living in southern New England were followed over
29      6 months (April through September) for daily symptoms. The data were analyzed for two
30      separate groups of subjects, 130 who used maintenance asthma medications during the follow-up
31      period and 141 who did not. The need for regular medication was considered to be a proxy for

        August 2005                              7-45        DRAFT-DO NOT QUOTE OR CITE

-------
               2
            o>
           Q
               1
               Obm
                                            98.8%
                                                                  City-Stratified Analysis
                                                                  All-Cities Analysis
                                                 75.7%
                  0.5
         1                      1.5
Odds Ratio for Incidence of Symptoms
       Figure 7-7.  Density curves of the odds ratios for the incidence of symptoms per 30 ppb
                   increase in 8-h avg O3 with a cumulative lag of 1 to 4 days for the individual
                   eight cities and the pooled average of all cities. Note that 99% and 76% of the
                   areas under the curves are greater than one for the pooled cities and
                   individual cities analyses, respectively.
       Source: Derived from Mortimer et al. (2002).
 1     more severe asthma. Not taking any medication on a regular basis and not needing to use a
 2     bronchodilator would suggest the presence of very mild asthma. Effects of 1-day lag O3 were
 3     observed on a variety of respiratory symptoms only in the medication user group.  Both daily
 4     1-h max and 8-h max O3 concentrations were similarly related to symptoms such as chest
 5     tightness and shortness of breath. Effects of O3, but not PM25, remained significant and even
 6     increased in magnitude in two-pollutant models. Some of the associations were noted at 1-h
 7     max O3 levels below 60 ppb. In contrast, no effects were observed among asthmatics not using
 8     maintenance medication. In terms of person-days of follow-up, this is one of the larger studies
 9     currently available that address symptom outcomes in relation to O3, and provides supportive
10     evidence for effects of O3 independent of PM25. Study limitations include limited control for
11     meteorological factors and the post-hoc nature of the population stratification by medication use.
12          Some international studies have reported significant symptoms associations with O3.
13     The incidence of asthma attacks was associated with O3 concentrations in a group  of 60 severe
14     asthmatics (mean age 55 years) followed over a 13-month period in Paris (Desqueyroux et al.,
       August 2005
                7-46
DRAFT-DO NOT QUOTE OR CITE

-------
 1      2002a). In a similar study, Desqueyroux et al. (2002b) observed O3-associated exacerbation of
 2      symptoms in 39 adult patients (mean age 67 years) with chronic obstructive pulmonary disease
 3      (COPD).  Interestingly, in contrast to the controlled human studies (see Section 6.3.1, Subjects
 4      with COPD), the O3 effect appeared larger among subjects who smoked and those with more
 5      severe COPD. However, the low O3 concentrations experienced during this study (summer
 6      mean 8-h max O3 of approximately 21 ppb [SD 9]) raise plausibility questions.  In a study of
 7      60 nonsmoking asthmatic adults (aged 18 to 55 years) in Bilthoven, the Netherlands, Hilterman
 8      and colleagues (1998) reported associations between O3 and daily symptoms of shortness of
 9      breath and pain upon deep inspiration.  The O3 associations were stronger than those of PM10,
10      NO2, SO2, and black smoke (BS). No differences in response were evident between subgroups
11      of subjects defined on the basis of steroid use or airway hyperresponsiveness. Daily use of
12      bronchodilators or steroid inhalers was not found to be associated with O3 in this study.
13           Other studies showed  only limited or a lack of evidence for symptom increases associated
14      with O3 exposure (Avol et al., 1998; Chen et al., 1998; Delfmo et al., 1996, 1997a,  1998a;
15      Gielen et  al., 1997; Jalaludin et al., 2004; Ostro et al., 2001; Taggart et al., 1996). Avol et al.
16      (1998) studied symptoms in asthmatic, wheezy, and healthy children aged 10 to 12 years in
17      southern California.  Some symptom associations were noted but they were inconsistent. For
18      example,  children with wheeze were at increased risk of difficulty breathing and wheezing at
19      low O3 concentrations, but not at higher O3 concentrations.  Authors noted that O3 concentrations
20      were relatively low and that children studied did not spend  substantial time outdoors engaged in
21      physical activities. Ostro et al. (2001) reported no associations between daily symptoms and
22      ambient O3 concentrations in a cohort of 138 African-American children with asthma followed
23      over 3 months (August to October) in Central Los Angeles  and Pasadena, CA.  However, the use
24      of extra asthma medication was associated with 1-h max O3 concentrations at a 1-day lag.
25      Delfmo and colleagues (1996) followed 12 asthmatic teens  living in  San Diego, CA for
26      respiratory symptoms over a two-month period and saw no relationship with central site ambient
27      O3. Personal O3 exposures measured with passive diffusion monitors were associated with the
28      composite symptom score and p2-agonist inhaler use, but the relationship with symptom score
29      disappeared when weekday/weekend differences  were controlled in the statistical analysis.
30      Study power was likely compromised by the small sample size.  This observation of stronger
31      associations with O3 levels from personal monitors implies  that gains in power may be achieved

        August 2005                              7-47         DRAFT-DO NOT QUOTE OR CITE

-------
 1      if exposure misclassification is reduced through the use of personal exposure measurements
 2      rather than central site ambient O3 concentrations.  A similar study of 22 asthmatics in Alpine,
 3      CA observed no effects of O3 on symptoms when personal O3 exposure was used as the exposure
 4      metric (Delfino et al., 1997a). However, a later study in the same location involving 24 subjects
 5      (Delfino et al., 1998a) did find an association between respiratory symptoms and ambient O3
 6      exposure, with stronger O3 effects experienced by asthmatics not on anti-inflammatory
 7      medication. In this study, a binary symptom score was used, whereas the earlier study used a
 8      linear symptom score of 0 through 6.
 9           In conclusion, the various studies seem to indicate that O3 concentrations are associated
10      with respiratory  symptoms and increased medication use in asthmatics. The multicities study by
11      Mortimer et al. (2002) provides an asthmatic population most representative of the U.S., and
12      several single-city studies also add to the knowledge base. However, there are a number  of well-
13      conducted, albeit relatively smaller studies that have not found these effects.
14
15      Panels of healthy subjects
16           Fewer studies examined the effect of O3 on respiratory symptoms in healthy individuals.
17      Neas et al. (1995) reported that in school children, evening cough was associated with O3 levels
18      weighted by hours spent outdoors. The study by Linn and colleagues (1996) of 269 school
19      children in southern California reported no associations between respiratory symptoms and O3,
20      but subjects were exposed to fairly low O3 concentrations as determined using personal
21      monitors. Gold et al. (1999) examined symptoms in 40 healthy children in southwest Mexico
22      City. Pollutant exposures were associated with increased production of phlegm in the morning,
23      although the effects of the air pollutants (PM25, PM10, and O3) could not be separated in
24      multipollutant models.  Hoek and Brunekreef (1995) did not find a consistent association
25      between ambient O3 levels and the prevalence or incidence of respiratory symptoms in children
26      living in two rural towns in the Netherlands. Collectively, these studies indicate that there is no
27      consistent evidence of an association between O3 and respiratory symptoms among healthy
28      children.
29
        August 2005                              7-48        DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.2.5  Acute Airway Inflammation
 2           Acute airway inflammation has been shown to occur among adults exposed to 80 ppb O3
 3      over 6.6 hours with exercise in controlled chamber studies (Devlin et al., 1991). Kopp and
 4      colleagues (1999) attempted to document inflammation of the upper airways in response to
 5      summer season O3 exposures by following a group of 170 school children in two towns in the
 6      German Black Forest from March to October of 1994.  To assess inflammation, the investigators
 7      collected nasal lavage samples at 11 time points spanning the follow-up period.  The nasal
 8      lavage samples were analyzed for markers of inflammation, including eosinophil cationic
 9      protein, albumin, and leukocyte counts.  Subjects who were sensitized to inhaled allergens were
10      excluded. When analyzed across the entire follow-up period, no association was detected
11      between upper airway inflammation and O3 concentrations. More detailed analysis showed that
12      the first significant O3 episode of the summer was followed by a rise in eosinophil cationic
13      protein levels, however, subsequent and even higher O3 episodes had no effect.  These findings
14      suggest an adaptive response of inflammation in the nasal airways that is consistent with
15      controlled human studies  (see Section 6.9, Effects of Inflammation and Host Defense).
16           Frischer and colleagues (1993) collected nasal lavage samples from 44 school children in
17      Umkirch, Germany the morning after "low" O3 days (<140 |ig/m3 or approximately 72 ppb) and
18      "high" O3 days (>180  |ig/m3 or approximately 93 ppb) to measure levels of biochemical markers
19      of inflammation. The researchers found that higher O3 levels were associated with increased
20      polymorphonuclear leukocyte counts in all children, and increases in myeloperoxydases and
21      eosinophilic cation proteins among children without symptoms of rhinitis (n = 30). These results
22      indicated that O3 was associated with inflammation in the upper airways. Frischer et al. (1997)
23      further investigated whether hydroxyl radical attacks played a role in mediating the O3-
24      associated inflammatory response of the airways. Ortho- and/>ara-tyrosine levels were
25      measured in the nasal  lavage samples and the ortholpara radical ratio was used to determine the
26      generation of hydroxyl radicals.  Significant increases in the ortholpara ratio were observed on
27      days following high ambient O3 levels. However, the ortholpara ratio was not related to
28      polymorphonuclear leukocyte counts, suggesting that there was  no detectable relationship
29      between hydroxyl radical  attacks and the inflammatory response seen in these children. Similar
30      to the study by Kopp et al. (1999), the ortholpara ratio decreased at the end of the summer
31      although O3 concentrations were still high, providing additional  evidence for a possible adaptive

        August 2005                             7-49        DRAFT-DO NOT QUOTE OR CITE

-------
 1     response. These findings, however, do not preclude the possibility that other unmeasured
 2     effects, including cell damage or lower airway responses, may have occurred with ongoing
 3     summer season exposures. In fact, a study of joggers repeatedly exposed to O3 while exercising
 4     over the summer in New York City suggested that cell damage may occur in the absence of
 5     ongoing inflammation (Kinney et al., 1996b).
 6           In two Mexico City studies by Romieu et al. (1998, 2002), the effect of antioxidant
 7     supplements on the association between O3 and lung function in outdoor workers and asthmatic
 8     children was investigated.  Romieu and colleagues (1998) observed significant inverse
 9     associations between O3 and lung function parameters, including FVC, FEVb and FEF25_75
10     (forced expiratory flow at 25 to 75% of FVC), among outdoor workers who were on the placebo,
11     but not among those taking the antioxidant supplement during the first phase of testing.
12     Likewise, O3 concentrations were associated with declines in lung function among children with
13     moderate-to-severe asthma who were on the placebo, but no associations were found among
14     those who were taking the vitamin C and E supplement (Romieu et al., 2002). These results
15     indicate that supplementation with antioxidants may modulate the impact of O3 exposure on the
16     small airways of two potentially at-risk populations, outdoor workers and children with
17     moderate-to-severe asthma. In a further analysis, genetic factors were found to contribute to the
18     variability between individuals in the effects of O3 on lung function (Romieu et al., 2004).
19     Individuals with polymorphism of the glutathione S-transferase gene (GSTM1 null genotype)
20     lack glutathione transferase enzyme activity, which plays an important role in protecting cells
21     against oxidative damage.  Results from this  analysis indicate that asthmatic children with
22     GSTM1 null genotype were found to be more susceptible to the impact of O3 exposure on small
23     airways function. Romieu et al. (2004) noted that supplementation with the antioxidant  vitamins
24     C and E above the minimum daily requirement might compensate for the genetic susceptibility.
25
26     7.2.6  Acute Ozone Exposure and School Absences
27           The association between school absenteeism and ambient air pollution was assessed  in  a
28     few studies (Chen et al., 2000; Gilliland et al., 2001; Park et al., 2002). In the study by Chen and
29     colleagues (2000), daily school absenteeism was examined in 27,793 students (kindergarten to
30     sixth grade) from 57 elementary school students in Washoe County, NV over  a two-year period.
31     One major limitation of this study was that the percent of total daily absences  was the outcome

       August 2005                             7-50        DRAFT-DO NOT QUOTE OR CITE

-------
 1      of interest, not illness-related absences, as reasons for absences were not noted in all schools.
 2      In models adjusting for PM10 and CO, ambient O3 levels were associated with school
 3      absenteeism.  With a distributed lag of 1 to 14 days, O3 concentrations were associated with a
 4      10.41% (95% CI:  2.73, 18.09) excess rate of school absences per 40 ppb increase in 1-h max O3.
 5      PM10 and CO concentrations also were associated with school absenteeism, however, the effect
 6      estimate for PM10 was negative. The inverse relationship between O3 and PM10 may have
 7      partially attributed to the negative association observed between PM10 and school absenteeism.
 8           Ozone-related school absences also were examined in a study of 1,933 fourth grade
 9      students from 12 southern California communities participating in the Children's Health Study
10      (Gilliland et al., 2001). Due to its comprehensive characterization of health outcomes, this study
11      is valuable in assessing the effect of O3 on illness-related school absenteeism in children. The
12      study spanned a period, January through June 1996, that captured a wide range of exposures
13      while staying mostly below the highest levels observed in the summer season.  All school
14      absences that occurred during this period were followed up with phone calls to determine
15      whether they were illness-related. For illness-related absences, further questions assessed
16      whether the illness was respiratory or gastrointestinal, with respiratory symptoms including
17      runny nose/sneeze, sore throat, cough, earache, wheezing, or asthma attacks. Multiple pollutants
18      were measured at a central site in each of the  12 communities.  The statistical analysis controlled
19      for temporal cycles, day of week, and temperature,  and expressed exposure as a distributed lag
20      out to 30 days. Associations were found  between the 30-day distributed lag of 8-h avg O3
21      (10 a.m.-6 p.m.) and all absence categories. Larger O3 effects were seen for respiratory causes
22      (147%  [95% CI:  6, 478] increase in absences per 30 ppb increase  in 8-h avg O3) than  for
23      nonrespiratory causes (61% [95% CI: 9,  138] increase). Among the respiratory absences, larger
24      effects  were seen for lower respiratory diseases than for upper respiratory diseases.
25      Multipollutant analyses were not performed; however, in single-pollutant models neither PM10 or
26      NO2 were associated with any respiratory or nonrespiratory illness-related absences. Some
27      concern exists regarding the possibility of residual  seasonal confounding given the six-month
28      time span of the monitoring period and the long lag periods of exposure, which  are likely to
29      capture seasonally changing factors such  as pollen  episodes. Further, the biological relevance of
30      O3 concentrations lagged 30 days present an interpretive challenge.
        August 2005                               7-51        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Park et al. (2002) examined the association between air pollution and school absenteeism
 2      in 1,264 students, first to sixth grade, attending school in Seoul, Korea.  The study period
 3      extended from March 1996 to December 1999, with 8-h avg O3 concentrations ranging from
 4      3.13 ppb to 69.15 ppb (mean 22.86 ppb). Note that analysis was performed using Poisson GAM
 5      with default convergence criteria. Same day O3 concentrations were positively associated with
 6      illness-related absences, but inversely associated with non-illness-related absences.  PM10
 7      concentrations also were positively associated with illness-related absences. In two-pollutant
 8      models containing O3 and PM10, both estimates were robust, with a slightly greater effect seen
 9      for O3.
10           Results from Chen et al. (2000), Gilliland et al. (2001), and Park et al. (2002) suggest that
11      ambient O3 concentrations, on the same day as well as accumulated over two to four weeks, may
12      be associated with school absenteeism, particularly illness-related absences. Further replication
13      is needed before firm conclusions can be reached regarding the effect of O3 on school absences.
14
15      7.2.7   Cardiovascular Endpoints
16           Several air pollution studies have examined various cardiovascular endpoints (Table
17      AX7-2 in Chapter 7 Annex).  The earlier studies focused on PM effects. For a more thorough
18      discussion of these PM studies and their health endpoints, refer to the 2004 PM AQCD (Section
19      8.3.1).  More recently studies have examined associations of O3 and other gaseous pollutants
20      with various measures of heart beat rhythms in panels of elderly subjects, as discussed below.
21      Other studies examined the increased risk of MI related to air pollutants exposures.
22
23      7.2.7.1  Cardiac Autonomic Control
24           Alterations in heart rate and/or rhythm are thought to reflect pathophysiologic changes that
25      may represent possible mechanisms by which ambient air pollutants such as O3 may exert acute
26      effects  on human health. Decreased HRV has been identified as a predictor of increased
27      cardiovascular morbidity and mortality. Brook et al. (2004) state that HRV, resting heart rate,
28      and blood pressure are modulated by a balance between the two determinants of autonomic tone
29      (the sympathetic and parasympathetic nervous systems).  They note that decreased HRV predicts
30      an increased risk of cardiovascular morbidity and mortality in the elderly and those with
31      significant heart disease, which is generally determined by analyses of time (e.g., standard

        August 2005                               7-52        DRAFT-DO NOT QUOTE OR CITE

-------
 1      deviation of normal R-R intervals) and frequency domains (e.g., low frequency/high frequency
 2      ratio by power spectral analysis, reflecting autonomic balance) measured during 24 hours of
 3      electrocardiography. Decreased parasympathetic input to the heart may provide an important
 4      mechanistic link between air pollution and cardiovascular mortality by promoting fatal
 5      tachy arrhythmias.
 6           The potentially adverse effects of air pollutants on cardiac autonomic control were
 7      examined in a large population-based study, among the first in this field. Liao et al. (2004)
 8      investigated short-term associations between ambient pollutants and cardiac autonomic control
 9      from the fourth cohort examination (1996-1998) of the population-based Atherosclerosis Risk in
10      Communities Study (ARIC). PM10, O3, and other gaseous air pollutants were examined in this
11      study. PM10 (24-h avg) and O3 exposures (8-h avg, 10 a.m.-6 p.m.) one day prior to the
12      randomly allocated examination date were used. They calculated 5-minute HRV indices
13      between 8:30 a.m. and  12:30 p.m., and used logarithmically-transformed data on high-frequency
14      (0.15 to 0.40 Hz) and low-frequency (0.04 to 0.15 Hz) power, standard deviation of normal R-R
15      intervals, and mean heart rate.  The effective sample sizes for O3 and PM10 were 5,431 and
16      4,899, respectively, from three U.S. study centers in North Carolina, Minnesota, and Mississippi.
17      PM10 concentrations measured one day prior to the HRV measurements were inversely
18      associated with both frequency- and time-domain HRV indices. Ambient O3 concentrations
19      were inversely associated with high-frequency power among whites. Consistently more
20      pronounced associations were suggested between PM10 and HRV among persons with a history
21      of hypertension.  Liao et al. note that these findings may represent potentially important
22      arrhythmogenic mechanisms of ambient air pollution. The acute adverse effect of air pollution
23      on cardiac autonomic control hypothesizes that increased air pollution levels may  stimulate the
24      autonomic nervous system and lead to an imbalance of cardiac autonomic control  characterized
25      by sympathetic activation unopposed by parasympathetic control.  Such an imbalance of cardiac
26      autonomic control may predispose susceptible people to greater risk of life-threatening
27      arrhythmias and acute cardiac events. The findings from Liao et al. were cross-sectionally
28      derived from a population-based sample and reflect the short-term effects  of air pollution on
29      HRV. When the regression coefficients for  each individual pollutant model were  compared, the
30      effects for PM10 were considerably larger than the effects for gaseous pollutants such as O3.
31      Because of the population-based sample, this study does have better generalizability than other

        August 2005                              7-53        DRAFT-DO NOT QUOTE OR CITE

-------
 1      smaller panel studies. The findings are suggestive of short-term effects of air pollutants,
 2      including O3, on HRV at the population level.
 3           Another population-based study of air pollutants and HRV was conducted in Boston on
 4      497 men from the VA Normative Aging Study (NAS) (Park et al., 2005).  Ozone showed several
 5      associations with HRV outcomes.  Stronger associations were reported with PM2 5.
 6      In two-pollutant models, the magnitude of the percent changes for both PM2 5 and O3 diminished
 7      slightly. In analyses by ischemic heart  disease, hypertension, and diabetes status, stronger
 8      associations of HRV with O3 and PM2 5 were observed for individuals with ischemic heart
 9      disease and hypertension.  These results are consistent with a Mexico City study (n = 34) by
10      Holguin et al. (2003) which reported an HRV effect for O3 in subjects with hypertension. The
11      association of O3 exposure with reduced low-frequency power in the full cohort seemed to be
12      driven by subjects not taking calcium-channel blockers (Park et al., 2005). This suggests that
13      this drug is blocking effects of O3 on the sympathetic pathway. This study cohort consists of all
14      males and almost all whites. This population-based study suggests that short-term exposures
15      to O3 are predictors of alteration in cardiac autonomic function as measured by HRV among
16      older male adults.
17           Two related studies in Boston, MA, examined the association between air pollution and the
18      incidence of ventricular arrhythmias (Dockery et al., 2005; Rich et al., 2005).  A total of 203
19      patients with implanted cardioverter defibrillators who lived within 25 miles of the ambient
20      monitoring site at the Harvard School of Public Health were monitored. They had a total of
21      635 person-years of follow-up or an average of 3.1 years per subject. In the analysis by Dockery
22      et al. (2005), positive associations were observed between ventricular arrhythmias within three
23      days of a prior event and a two-day mean of several air pollutants, including PM2 5, black carbon,
24      NO2, CO, and SO2.  No associations were observed with O3. There was, however, a suggestion
25      of increasing risk with increasing quintiles of O3 (p < 0.05).  The analysis by Rich et al. (2005)
26      observed stronger O3  effects on ventricular arrhythmias using a case-crossover study design.
27      Case periods were defined by the time each arrhythmic event began; for each case, three to four
28      control periods were selected by matching on weekday and hour of the day within the same
29      calender month. For a 20  ppb increase  in 24-hour moving average O3, a 27% (95% CI: 0, 60)
30      increased risk of ventricular arrhythmias was estimated.  Significant effects also were found
31      for PM2 5, NO2, and SO2. In two-pollutant models, the O3 effect was found to be generally

        August 2005                               7-54        DRAFT-DO NOT QUOTE OR CITE

-------
 1      robust. Stratified analysis by the presence of a recent ventricular arrhythmia within the previous
 2      three days indicated that O3 was associated with increased risk among subjects without a recent
 3      event (37% [95% CI: 6, 79]), but not among those with recent events (5% [95% CI: -27, 49]).
 4      Rich et al. explained that the use of the case-crossover study design and conditional analysis
 5      might have contributed to the stronger associations observed in their study compared to Dockery
 6      et al. In addition, the use of a 24-hour moving average instead of a calender-day air pollution
 7      concentration might have reduced exposure misclassification, resulting in larger effect estimates.
 8           Other studies do not provide evidence for an O3 effect on HRV  and cardiac arrhythmias
 9      (Peters et al., 2000a; Rich et al., 2004; Vedal et al., 2004). These studies, however, may have
10      had limited power to examine subtle effects. Gold et al. (2000; reanalysis Gold et al., 2003)
11      reported results that suggest that O3 exposure may decrease vagal tone, leading to reduced HRV.
12      Schwartz et al. (2005) reported  a weak association of O3 with the root mean squared differences
13      between adjacent R-R intervals in a study of 28 elderly subjects and noted that lack of personal
14      exposure measurements may render such studies less able to assess autonomic functions. This
15      study reported the strongest effects for black carbon.
16
17      7.2.7.2  Acute Myocardial Infarction
18           The effect of O3 on the incidence of MI was examined in a limited number of studies.
19      Acute MI was studied in relation to air pollution in Toulouse, France  based on the existence of
20      an acute MI registry (Monitoring Trend and Determinants in Cardiovascular Disease
21      [MONICA]) and an air quality network covering the same population (Ruidavets et al., 2005).
22      After adjustment for temperature, relative humidity, and influenza epidemics, the relative risk of
23      acute MI occurrence was 1.76 (95% CI:  1.12, 2.45) for current day O3 concentrations. The
24      increased risk of MI was more evident in the oldest group, 55 to 64 years of age. Further, the
25      oldest subjects without a personal history of ischemic heart disease were more susceptible to an
26      acute event when O3 levels increased.  No PM data was reported in this study.
27           In a case-crossover study  (n = 772) in Boston, MA, Peters et al. (2001) reported an odds
28      ratio of 1.27 (95% CI:  0.87, 1.88) per 40 ppb increase in 2-h avg O3 (1 hour before onset of
29      event). Stronger effects on the incidence of MI were observed for PM2 5 and PM10.
30
        August 2005                               7-55         DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.2.7.3  Cardiovascular Endpoints in Human Clinical Studies
 2           In a controlled human exposure study discussed in Chapter 6, Sections 6.3.4 and 6.10,
 3      Gong et al. (1998a) studied 10 nonmedicated hypertensive and 6 healthy male adults exposed to
 4      0.3 ppm O3 with intermittent exercise in relation to various cardiovascular effects.  The overall
 5      results did not indicate acute cardiovascular effects of O3 in either the hypertensive or control
 6      subjects.  The authors observed an increase in rate-pressure product and heart rate, a decrement
 7      for FEVl3 and a >10 mm Hg increase in the alveolar/arterial pressure difference for O2 following
 8      O3 exposure. These findings suggest that O3 can exert cardiovascular effects indirectly by
 9      impairing alveolar-arterial O2 transfer and potentially reducing O2 supply to the myocardium.
10      Ozone exposure may increase myocardial work and impair pulmonary gas exchange to a degree
11      that may be clinically important in persons with  significant pre-existing cardiovascular
12      impairment.
13
14      1.2.1 A  Summary of Field Studies with Cardiovascular Outcomes
15           A limited epidemiologic database examining cardiovascular outcomes in relation to O3
16      exposures is available. Among these studies, three were population-based and involved cohorts
17      such as the ARIC (Liao et al., 2004), MONICA (Ruidavets et al., 2005), and NAS (Park et al.,
18      2005).  Such studies may offer more informative results based on their large subject-pool  and
19      design. Results from these three studies were suggestive of an association between O3 exposure
20      and the cardiovascular endpoints studied.  As in  the case of respiratory disease outcomes, Brook
21      et al. (2004) state that the increase in relative risk for cardiovascular disease due to air pollution
22      is small compared with the impact of the established cardiovascular risk factors. However,
23      because of the enormous number of people affected, even conservative risk estimates can
24      translate into a substantial increase in mortality due to cardiovascular disease within the
25      population.  The impact of air pollution on cardiovascular disease therefore may represent a
26      serious public health problem.
27
28      7.2.8 Summary of Field Studies Assessing Acute Ozone Effects
29           •  Results from recent field/panel studies  support the evidence from clinical studies that
30              acute O3 exposure is associated with a significant effect on lung function, as indicated
31              by decrements in FEVl3 FVC, and PEF. The declines in lung function were noted
32              particularly in children and asthmatics.

        August 2005                              7-56        DRAFT-DO NOT QUOTE OR CITE

-------
 1          •   Limited evidence suggests that more time spent outdoors, higher levels of exertion,
 2              and the related increase in O3 exposure may potentiate the risk of respiratory effects.
 3              In addition to children and asthmatics, adults who work or exercise outdoors may be
 4              particularly vulnerable to O3-associated health effects.
 5
 6          •   Many new studies have examined the association between O3 concentrations and a
 7              wide variety of respiratory symptoms (e.g., cough, wheeze, production of phlegm,
 8              and shortness of breath). Collectively, the results  suggest that acute exposure to O3
 9              is associated with increased respiratory  symptoms and increased as-needed medication
10              use in asthmatic children.
11
12          •   Additional panel studies investigated the effect of O3 on other health outcomes,
13              including school absences, and markers of inflammation and oxidative damage.
14              Ozone exposure was associated with increases in respiratory-related school absences,
15              as well as increased inflammation and generation of hydroxyl radicals in the upper
16              airways.  Use of antioxidant supplements was found to diminish the O3 effect on lung
17              function.
18
19          •   Some field studies have examined the association  between O3 and cardiac physiologic
20              outcomes.  The current evidence is rather limited but supportive of a potential effect on
21              HRV, ventricular arrhythmias, and the incidence of MI.  Additional studies need to be
22              performed before any conclusions can be made regarding an O3 effect on
23              cardiovascular outcomes.

24

25

26     7.3    ACUTE EFFECTS OF OZONE ON DAILY EMERGENCY
27            DEPARTMENT VISITS AND HOSPITAL ADMISSIONS

28     7.3.1    Summary of Key Findings on Studies  of Emergency Department Visits
29              and Hospital Admissions from the 1996 O3 AQCD

30          In the 1996 O3 AQCD, aggregate population time-series studies of O3-related health effects

31     provided relevant evidence of acute responses, even below  a 1-h max O3  of 0.12 ppm.

32     Emergency room visits and hospital admissions were examined as possible  outcomes following

33     exposure to O3. In the case of emergency room visits, the evidence was limited (Bates et al.,

34     1990; Cody et al.,  1992; Weisel et al., 1995; White et al., 1994), but results  generally indicated

35     an O3 effect  on morbidity. The strongest and most consistent evidence of O3 effects, at levels

36     both above and below 1-h max O3 levels of 0.12 ppm, was provided by the multiple studies that

37     had been conducted on summertime daily hospital admissions for respiratory causes  in various

38     locales in eastern North America (Bates and Sizto, 1983, 1987, 1989; Burnett et al., 1994;
       August 2005                              7-57       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Lipfert and Hammerstrom, 1992; Thurston et al., 1992, 1994). These studies consistently
 2      demonstrated that O3 air pollution was associated with increased hospital admissions, accounting
 3      for roughly one to three excess respiratory hospital admissions per million persons with each 100
 4      ppb increase in 1-h max O3.  This association had been shown to remain even after statistically
 5      controlling for the possible confounding effects of temperature and copollutants (e.g., H+, SO4 2,
 6      PM10), as well as when considering only days with 1-h max O3 concentrations below 0.12 ppm.
 7      Overall, the aggregate population time-series studies considered in the 1996 O3 AQCD provided
 8      strong evidence that ambient exposures to O3 can cause significant exacerbations of preexisting
 9      respiratory disease in the general public.
10
11      7.3.2   Review of Recent Studies of Emergency Department Visits for
12              Respiratory Diseases
13           Emergency department visits represent an important acute outcome that may be affected by
14      O3 exposures.  Morbidities that result in emergency department visits are closely related to, but
15      are generally less severe than, those that result in unscheduled hospital admissions.  In many
16      cases, acute health problems are successfully treated in the emergency department; a subset of
17      more severe cases that present initially to the emergency department may require admission to
18      the hospital.
19           Several studies have been published in the past decade examining the temporal
20      associations between O3 exposures and emergency department visits for respiratory diseases
21      (Table AX7-3  in Chapter 7 Annex).  Total respiratory causes for emergency room visits may
22      include asthma, pneumonia, bronchitis, emphysema, upper and lower respiratory infections such
23      as influenza, and a few other minor categories.  Asthma visits typically dominate the daily
24      incidence counts. Chronic bronchitis and emphysema  often are combined to define COPD,
25      which is a prominent diagnosis among older adults with lung disease.  Figure 7-8 presents
26      percent changes in emergency department visits for asthma from single-pollutant models,
27      with results expressed in standardized increments.  The lags presented in the figure vary
28      depending on reported results. Most studies reported effect estimates from a short lag period
29      (0 to 2 days). Results from Weisel et al. (2002) are not included as comparable risks estimates
30      for O3 are not presented. Among the U.S. studies, there was one multeity study which examined
31      three cities in Ohio (Jaffe et al., 2003). Several presented Atlanta, GA data.  In general, O3

        August 2005                              7-58        DRAFT-DO NOT QUOTE OR CITE

-------
OQ
to
o
o
                                           % 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 20C
I I I I I I I 1 1 1 1 I I I I I 1 1 i II
U.S. and Canada | A s_ „
£ |ng i X Allypar

m 	 lag 2 or 3
O Cool


• lay 0-2
•*- lag 0-2
••- lag 0-2
-•- Iag1
• Inn ^

Europe I
{ Best of lag 0 to 3 by city
— X— Iag1
	 X— lagO
x lag (1-1


lag O-'l
C\ Ian 0
— >4 	 lag 0



        Figure 7-8.  Ozone-associated percent change (95% CI) in emergency department visits for asthma per

                   standardized increment (see Section 7.1.3.2).

-------
 1      effect estimates from warm season only analyses tended to be positive and larger compared to
 2      results from cool season or all year analyses.
 3           Among studies with adequate controls for seasonal patterns, many reported at least one
 4      positive association with O3. These studies examined emergency department visits for total
 5      respiratory complaints (Delfmo et al., 1997b,  1998b; Herfiandez-Garduno et al., 1997;
 6      Ilabaca et al., 1999; Jones et al., 1995; Lin et al., 1999), asthma (Friedman et al., 2001; Jaffe
 7      et al., 2003; Stieb et al.,  1996; Tenias et al., 1998; Tobias et al., 1999; Tolbert et al., 2000;
 8      Weisel  et al., 2002), and COPD (Tenias et al., 2002).
 9           One recent study examined emergency department visits for total and cause-specific
10      respiratory diseases in Atlanta, GA over an 8-year period (Peel et al., 2005). A distributed lag of
11      0 to 2 days was specified a priori. Ozone concentrations were associated with emergency
12      department visits for total respiratory diseases and upper respiratory infections in all ages.
13      A marginally significant association was observed with asthma visits (2.6% [95% CI:  -0.5, 5.9]
14      excess risk per 30 ppb increase in 8-h max O3), which became stronger when analysis was
15      restricted to the warm months (3.1% [95% CI: 0.2, 6.2] excess risk). In multipollutant models
16      adjusting for PM10, NO2 and CO, O3 was the only pollutant that remained significantly associated
17      with upper respiratory infections. Another large asthma emergency department study was
18      carried  out during the months of May through September from 1984 to 1992 in St. John, New
19      Brunswick, Canada (Stieb et al., 1996).  Effects were examined separately among children aged
20      less than 15 years and in persons aged 15 years and older. A significant effect of O3 on
21      emergency department visits was reported among persons 15 years and older.  There was
22      suggestion of a threshold somewhere in the range below a 1-h max O3 of 75 ppb. A study in
23      Valencia, Spain from 1994 to 1995 observed that emergency room visits for asthma among
24      persons over 14 years old were robustly associated with relatively low O3 levels (median 1-h
25      max O3 of 62.8 |ig/m3 or approximately 32.4 ppb) (Tenias et al., 1998). The excess risk of
26      asthma  emergency room visits was larger in the warm season (May to October), 85% (95%
27      CI:  20, 188) excess risk per 40 ppb increase in 1-h max O3, compared to the cool season
28      (November-April), 31% (95% CI: -24, 125) excess risk (Tenias et al., 1998).
29           Among the studies that observed a positive association between O3 and emergency
30      department visits for respiratory outcomes, O3 effects were found to be robust to adjustment for
31      PM10, NO2, SO2, and BS (Lin et al., 1999; Peel et al., 2005; Tenias et al., 1998).  One study by

        August 2005                              7-60        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Tolbert and colleagues (2000) observed that the significant univariate effects of both O3 and
 2      PM10 on pediatric asthma emergency department visits in Atlanta, GA became nonsignificant in
 3      two-pollutant regressions, reflecting the high correlation between the two pollutants (r = 0.75).
 4           For several other studies with total respiratory and asthma outcomes, inconsistencies
 5      confound an interpretation of likely causal effects. For example, in a Montreal, Canada study,
 6      O3 effects on total respiratory emergency department visits were seen in a short data series from
 7      the summer of 1993 but not in a similar data series from the summer of 1992 (Delfmo et al.,
 8      1997b).  The significant 1993 results were seen only for persons older than 64 years. A very
 9      similar analysis of two additional summers (1989 and 1990) revealed  an O3 association only for
10      1989 and again only in persons over 64 years old (Delfmo et al., 1998b).  An analysis of data on
11      respiratory emergency department visits from June to August of 1990 in Baton Rouge, LA
12      reported O3 effects in  adults, but not in children or among the elderly (Jones et al., 1995).
13           Tobias and colleagues (1999) showed that regression results for asthma  emergency
14      department visits could be quite sensitive to methods used to control for asthma epidemics.
15      Ozone was associated with the outcome variable in only one of eight models tested. An Atlanta,
16      GA study by Zhu et al. (2003) examined asthma emergency department visits in children during
17      three summers using Bayesian hierarchical modeling to address model variability.  Data were
18      analyzed at the zip code level to account for spatially misaligned longitudinal  data. Results
19      indicated a positive, but nonsignificant relationship between O3 and emergency room visits
20      for asthma.
21           Other studies also reported no association between O3 and emergency department visits for
22      respiratory causes (Atkinson et al., 1999a; Castellsague et al., 1995; Chew et al.,  1999; Hwang
23      and Chan, 2002; Sunyer et al., 1997). Using Bayesian hierarchical modeling,  Hwang and Chan
24      (2002) examined the effect of air pollutants on daily clinic visits for lower respiratory illnesses
25      across 50 cities in Taiwan. All pollutants except O3 were associated with daily clinic visits.  In a
26      pooled analysis of emergency admissions for asthma in four European cities as part of the Air
27      Pollution on Health: European Approach (APHEA) study, there was no overall effect  of O3
28      observed (Sunyer et al., 1997). Atkinson et al. (1999a) in London, England also did not find an
29      association between O3 and emergency department visits at a mean 8-h max O3 concentration of
30      17.5 ppb. One study by Thompson and colleagues (2001) in Belfast, Northern Ireland observed
31      a decreased risk of childhood asthma admissions (-21% [95% CI:  -33, -6] per 20 ppb increase

        August 2005                              7-61        DRAFT-DO NOT QUOTE OR CITE

-------
 1      in 24-h avg O3) in the cold season (November-April).  After adjusting for benzene levels, O3 was
 2      no longer associated with asthma emergency department visits. The inverse relationship of O3
 3      with benzene concentrations (r = -0.65), and perhaps with other pollutants, might have produced
 4      the apparent protective effect of O3.  No significant O3 effect was found in the warm season
 5      (May-October).  The O3 levels were low in both seasons, with a mean 24-h avg O3 concentration
 6      of 18.7 ppb in the warm season and  17.1 ppb in the cold season. A study by Hajat et al. (1999,
 7      2002) of physician consultations for asthma, lower respiratory diseases, and upper respiratory
 8      diseases in London reported negative associations with O3, which was suggestive of residual
 9      confounding by copollutants or weather factors (note that data were analyzed using Poisson
10      GAM with default convergence criteria).  Several other emergency department studies looking at
11      O3 are more difficult to interpret due to inadequate control for seasonal patterns, very low O3
12      levels, or because no quantitative results were  shown for O3 (Buchdahl et al., 1996, 2000; Garty
13      et al.,  1998; Holmen et al., 1997; Lierl and Hornung, 2003; Lipsett et al.,  1997; Nutman et al.,
14      1998).
15           Although several studies found a significant association between O3 concentrations and
16      emergency department visits for respiratory causes, some inconsistencies were observed. The
17      inconsistencies may be attributable,  at least partially, to differences in model specifications and
18      analysis approach among the various studies. For example, ambient O3 concentrations, length of
19      the study period,  and statistical methods used to control confounding by seasonal patterns and
20      copollutants appear to affect the observed O3 effect on emergency department visits. The body
21      of evidence remains inconclusive regarding effects of O3 on the risk of emergency department
22      visits.
23
24      7.3.3   Studies of Hospital Admissions for Respiratory Diseases
25           Hospital admissions represent  a medical  response to a serious degree of morbidity for a
26      particular disease. Scheduled hospitalizations  are planned in advance when a particular clinical
27      treatment is needed. However, unscheduled admissions are ones that occur in response to
28      unanticipated disease exacerbations  and are more likely to be affected by environmental factors,
29      such as air pollution. As such, the hospital admissions studies reviewed here focused
30      specifically on unscheduled admissions.  Study details and results from hospital admissions
31      studies published over the past decade are summarized in Table AX7-4 (in the  Chapter 7

        August 2005                              7-62        DRAFT-DO NOT  QUOTE OR CITE

-------
 1      Annex). As a group, these hospitalization studies tend to be larger in terms of geographic and
 2      temporal coverage, and indicate results that are generally more consistent than those reviewed
 3      above for emergency department visits.  As in the case for all studies that examine changes in
 4      aggregate measures of acute disease outcomes over time, the following should be considered in
 5      comparing results:  (1) difference in types of respiratory diseases for hospital admission; (2) age
 6      of study population; (3) mean level of O3 during study; (4) single-city versus multicity studies;
 7      (5) length of study (e.g., <5 years versus >5 years); (6) analysis by season versus all year;
 8      (7) O3-only versus multipollutant models; (8) number of exposure lag days; and (9) type of study
 9      (e.g., case-crossover versus time-series). These factors are considered in the sections below with
10      further discussion on potential confounding of the O3 effect estimate by seasonal factors and
11      copollutants.
12
13      7.3.3.1  All Year and Seasonal Effects of Ozone on Respiratory Hospitalizations
14           The effect of O3 on respiratory hospitalizations was examined in  various studies conducted
15      in the U.S. and abroad.  Figures 7-9 and 7-10 present risk estimates from all total respiratory
16      hospital admission studies. Burnett et al. (1995), which did not present quantitative results for
17      O3, and Yang et al. (2003), which only presented odd ratios, were not  included in the figures.
18      In cases where multiple lags were presented, the multiday lag was selected to represent the
19      cumulative effect from all days examined. If only single-day lags were analyzed, the effect
20      estimate of the shortest lag time, usually a lag of 0 or 1 day, was presented.  Figure 7-9 plots the
21      effect estimates and 95% CIs from 15 studies that analyzed all year data. The risk estimates are
22      arranged by age groups. The preponderance of positive risk estimates, with some that are
23      statistically significant, is readily apparent. Figure 7-10 presents the season-stratified effect
24      estimates by region. For studies that reported risk estimates from all four seasons, only the
25      summer and winter estimates are presented.  It appears that the warm season estimates,
26      collectively, tend to be larger, positive values compared to all year and cool season estimates.
27      All of the negative estimates were from  analyses using cool season data only, which might
28      reflect the inverse correlation between O3 and copollutants, namely PM, during that season.
29      These studies are discussed below in further detail.
30           Among the respiratory hospitalization studies, the most robust and informative results were
31      observed when a broad geographic area was examined using a consistent analytical methodology

        August 2005                               7-63         DRAFT-DO NOT QUOTE OR CITE

-------
                                                      % Change in Respiratory Hospitalization
               Burnett et al. (2001):Toronto, Canada (age <2)  -
                   Atkinson et al. (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)  -
                  Wong et al. (1999a): Hong Kong (age 0-4)  -
                 Linn et al. (2000): Los Angeles, CA (age 30+)  -
                   Atkinson et al. (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, Austrailia  -
                  Wongetal.(I999a): Hong Kong (age5-64)  -
                   Atkinson et al. (1999b): London, England  -
                Ponce de Leon et al. (1996): London, England  -
                  Prescott et al. (1998): Edinburgh, Scotland  -
           Schouten etal. (1996): Amsterdam, the Netherlands  -
              Petroeschevsky et al. (2001): Brisbane, Austrailia  -
                         Wongetal. (1999a): Hong Kong  -
        GwynnandThurston (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 et al. (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): Brisbane, Austrailia  -
                          Wongetal. (1999a): Hong Kong  -
                                                       -20
                                                         I
           -10
            I
                   10
                    I
20
 I
30
 I
                                                    | Age 0-14 Years
                 lag 0
                                                    [Age 15-64 Years
                                                    [ Age 65+Years
                                                    All Ages
                                                      lag 0-4
                                               - lag 0
                                               lag 0-3
                                |0
                            lag 3
                           	•-
                      lag 0-2
                     	 lag 0-3
                                    lag 0-2
                                                 lag 0-5
                                                       lag 2
                                                  lag 0-3
                             	 lag 0
                             -*	 lag 0-2
                                                     lag 0-3
                                      lag 0-1
                         Iag3-
                              lag 0-3
                             lag 1
                                lag 1
                                 lag 1
                                          Iag1
                                                      Iag1
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.
August 2005
7-64
DRAFT-DO NOT QUOTE OR CITE

-------
                                         % Change in Respiratory Hospitalizations
                              -20
0
20
40
60
                 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)
           Ponce de Leon et al. (1996):
           London, England (all ages)

              Schouten et al, (1996):
          Amsterdam and Rotterdam,
            the Netherlands (all ages)
               Wongetal.(1999a):
               Hong Kong (all ages)
1 j 1 1 1 1

U.S. and Canada
-X-
• lagO
• Inn -1 9
Q * lag 1

	 «H
0 •

Europe
— • — lag 0-2
	
-e, "*" la92
[Asia |
n " lap 0-3
\J
i

X All year
• Warm
O Cool



lag 0-4








       Figure 7-10.   Ozone-associated percent change (95% CI) in total respiratory
                      hospitalizations by season per standardized increment (see Section 7.1.3.2).
1      (Anderson et al., 1997; Burnett et al., 1995, 1997a). These studies have all reported an O3 effect

2      on respiratory hospital admissions.  The largest such study to-date was carried out using data on

3      all-age respiratory hospital admissions from 16 Canadian cities with populations exceeding

4      100,000 during the period 1981 to 1991 (Burnett et al., 1997a). In addition to O3, the authors

5      evaluated health effects of SO2, NO2, CO, and coefficient of haze (a surrogate for black carbon

6      particle concentrations).  Pooling the 16 cities, a positive association was observed between

7      respiratory hospital admissions and the 1-day lag O3 concentration in the spring (5.6% [95%CI:
       August 2005
     7-65
 DRAFT-DO NOT QUOTE OR CITE

-------
 1      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]).
 2      The results for fall were also positive, though of smaller magnitude (3.8% [95%CI:  -0.2, 7.9]).
 3      There was no evidence for an O3 effect in the winter season (-0.8% [95%CI: -4.8, 3.3]).
 4      Control outcomes related to blood, nervous system, digestive system, and genitourinary system
 5      disorders were not associated with O3. In a previous study focused mainly on evaluating health
 6      impacts of sulfate particles, Burnett and colleagues (1995) reported results from a time-series
 7      analysis of all-age respiratory hospital admissions to 168 hospitals in Ontario, Canada over a
 8      6-year period (1983 to 1988). The outcome data were prefiltered to remove seasonal variations
 9      using a weighted 19-day moving average.  The authors reported that O3 was associated with
10      respiratory hospital admissions; however, no quantitative results for O3 were presented.
11           Results from an analysis of five European cities indicated strong and consistent O3 effects
12      on unscheduled hospital admissions for COPD (Anderson et al., 1997). The five cities
13      examined — London, Paris, Amsterdam, Rotterdam, and Barcelona — were among those
14      included in the multeity APHEA study. The number of years of available data varied from 5 to
15      13 years among the cities. City-specific effect estimates were pooled across cities using
16      weighted means.  An association with O3 was observed in full year analyses.  Season-stratified
17      analyses indicated that the O3 effect was larger in the warm season (April-September), 4.7%
18      (95% CI: 1.6, 7.9) excess risk per 40 ppb increase in 1-h max O3, compared to the cool season
19      (October-March), 1.6% (95% CI: -3.1, 7.9) excess risk.  There was no significant heterogeneity
20      in O3 effects among the cities.
21           Several additional studies carried out in one or two cities over a span of five or more years
22      provided substantial  additional evidence regarding O3 effects on respiratory hospital admissions
23      (Anderson et al.,  1998; Burnett et al., 1999, 2001; Moolgavkar et al.,1997; Petroeschevsky et al.,
24      2001; Ponce de Leon et al., 1996; Sheppard et al., 1999 [reanalysis Sheppard, 2003]; Yang et al.,
25      2003). Moolgavkar and colleagues (1997) reported significant and robust O3 effects on
26      respiratory hospital admissions in adults 65 years and older in Minneapolis and St. Paul, MN,
27      but not in Birmingham, AL.  The absence of effects in the southern city may reflect less
28      penetration of O3 into the indoor environment due to greater use of air conditioning, and thus
29      less correlation between central site O3 monitoring and actual exposures of the urban populace.
30      Ozone effects on all-age and age-stratified asthma and total respiratory hospital admissions were
31      observed in Brisbane, Australia (Petroeschevsky et al., 2001). Effect sizes appeared consistent

        August 2005                              7-66        DRAFT-DO NOT QUOTE OR CITE

-------
 1      in the warm and cool seasons (data not provided). Petroeschevsky et al. commented that the
 2      year-round effect of O3 might reflect the relatively small degree of seasonal variation in O3 levels
 3      observed in Brisbane.  Although O3 levels were quite low year-round, they did not notably
 4      decline during the winter period.  The authors also noted that given the subtropical climate in
 5      Brisbane, characterized by warm, dry winters, perhaps the proportion of the population exposed
 6      to winter O3 concentrations was higher than in cities where inclement winter weather might force
 7      populations indoors.
 8           Another set of studies have examined associations between O3 and respiratory
 9      hospitalizations in single cities over shorter (<5 years) time spans.  Positive and significant O3
10      effects were reported in Cleveland, OH (Schwartz et al., 1996); New York City (Gwynn and
11      Thurston, 2001); Northern New Jersey (Weisel et al., 2002); Toronto, Canada (Burnett et al.,
12      1997b); Helsinki, Finland (Ponka and Virtanen, 1996); Sao Paulo, Brazil (Braga et al., 1999;
13      Gouveia and Fletcher,  2000a); and Hong Kong (Wong et al.,  1999a). The Helsinki study
14      reported significant effects of O3 on both asthma and on digestive disorders in a setting of very
15      low O3 concentrations  (Ponka and Virtanen, 1996), which raises questions of plausibility.
16           Less consistent effects of O3 were seen in other respiratory hospitalization studies
17      (Schouten et al., 1996; Lin et al., 2003; Lin et al., 2004; Morgan et al., 1998a; Oftedal et al.,
18      2003). In a study conducted in Amsterdam and Rotterdam, the Netherlands, associations
19      between O3 and respiratory admissions were observed; however, results were difficult to
20      interpret due to the large number of statistical tests performed (Schouten et al., 1996).  In a
21      California study by Neidell (2004), a negative association was observed between hospitalizations
22      for asthma and naturally occurring seasonal variations in O3 within zip codes in children aged
23      0 to 18 years. However, the O3 effect was found to be influenced by socioeconomic  status.
24      Among children of low socioeconomic status, O3 generally was associated with increased
25      hospitalizations, with statistical significance reached in certain age groups.  Neidell further stated
26      that avoidance behavior on high O3 days may have attributed to the negative relationship
27      observed in children of higher socioeconomic status.
28           No associations between respiratory hospital admissions and O3 were seen in studies from
29      Los Angeles, CA (Linn et al., 2000; Mann et al.,  2002; Nauenberg and Basu, 1999);  Vancouver,
30      Canada (Lin et al., 2004); London, England (Atkinson et al., 1999b); Edinburgh, Scotland
31      (Prescott et al., 1998);  and Drammen, Norway (Hagen et al., 2000). Several of these studies

        August 2005                               7-67        DRAFT-DO NOT QUOTE OR CITE

-------
 1      were carried out in locations with low O3 levels, suggestive of a nonlinear concentration-
 2      response relationship (Lin et al., 2004; Prescott et al., 1998).  The nonsignificant findings in the
 3      South Coast air basin, CA area are surprising given the elevated O3 concentrations observed
 4      there (Mann et al., 2002). Inadequate control of seasonal confounding may underlie some of the
 5      nonsignificant and negative findings.  An additional factor likely contributing to the variability
 6      of results is the relatively small sample sizes included in some of these studies.
 7           For respiratory hospitalization outcomes, the largest, most significant associations with O3
 8      concentrations were observed when using short lag periods, in particular a 0-day lag (exposure
 9      on same day) and a 1-day lag (exposure on previous day).  In a study of 16 Canadian cities by
10      Burnett et al. (1997a), the strongest association between O3 and respiratory hospitalizations was
11      found at a 1-day lag.  A decline in the magnitude and significance of the effect was seen with
12      increasing days lagged for O3. Anderson et al. (1997) investigated the association between O3
13      and daily hospital admissions for COPD in five European cities. Lags up to 5 days were
14      examined, and the largest risk estimates were found using 0- and 1-day lags. These results
15      suggest that O3 has a short-term effect on respiratory hospitalizations.
16           Burnett et al. (2001) investigated the association between respiratory hospitalizations and
17      O3 in children less than 2 years of age.  Lags up to five days were examined  after stratifying by
18      season (Figure 7-11). In the summer season, significant associations between O3 and daily
19      admissions  were found in several of the lags, with the largest risk estimate of 12.5% (95% CI:
20      5.7, 19.7) excess  risk per 40 ppb increase in 1-h max O3  at a 1-day lag. In comparison, the
21      O3-related risk estimate was 30.2% (95% CI:  18.0, 42.4) using a cumulative lag period of
22      5 days. The large effect  estimate for the cumulative lag period indicated that O3 exposure likely
23      had an immediate effect that persisted over several days.
24           Weisel et al. (2002) stated that a lag period of  1 to  3 days between exposure to O3 and
25      hospital admissions or emergency department visits  for asthma was plausible because it might
26      take time for the disease  to progress to the most serious responses following exposure.
27      In addition, taking medication could delay further the progression of the adverse effect.  Thus,
28      although strongest associations are found at lags of 0 and 1 days, examining longer single-day
29      lag periods  or multiday lag periods may further enhance understanding of the effect of O3 on
30      hospitalizations.
31

        August 2005                               7-68        DRAFT-DO NOT QUOTE OR CITE

-------
              0.20 R
                                0          10          20          30          40
                                  % Change in Respiratory Hospitalization
                                         50
       Figure 7-11.  Comparison of single-day lags (0-, 1-, 2-, 3-, 4-, and 5-day) to a cumulative
                     multiday lag (0- to 4-day) for percent changes in total respiratory
                     hospitalizations per 40 ppb increase in 1-h max O, in children less than
                     two years of age.
       Source: Derived from Burnett et al. (2001).
 1           In conclusion, while some inconsistencies are noted across studies, the evidence supports
 2     the findings of significant and robust effects of O3 on various respiratory disease hospitalization
 3     outcomes.  Large multicity studies, as well as many studies from individual cities have reported
 4     significant O3 associations with total respiratory hospitalizations, asthma, and COPD, especially
 5     in studies analyzing the O3 effect during the summer or warm season.
 6
 7     7.3.3.1   Potential Confounding of the Ozone Effect on Respiratory Hospitalizations
 8              by Copollutants
 9           As in the case for most air pollution studies, potential confounding of the association
10     between O3 and respiratory hospitalizations by copollutants generally was examined using
11     multipollutant regression models.  Figure  7-12 compares the risk estimates from models with
12     and without adjustment for PM indices. This figure indicates that O3 risk estimates are fairly
13     robust to PM adjustment in all  year and warm season only data. None of the studies examined
14     PM-adjusted O3 risk estimates  in cool season only data.
15           Several analyses of a large data set from Toronto, Canada spanning the years 1980 to 1994
16     reported O3 effects on respiratory hospitalizations for all ages (Burnett et al., 1997b, 1999) and
       August 2005
7-69
DRAFT-DO NOT QUOTE OR CITE

-------
                                   % 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 PM2 5  -|
                    with PM10.2.5
          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



J ^ O03only
u
• • CH with PM
lag 1-3 w 3 VVIIM rm



lin n A A


Q lag 1
PI
m lag 0
V
Latin America |
Q lagO
Australia

Asia

Pi
0 lag 2
V
PI
^ U lag 0-3
V
       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.
1      for persons under the age of 2 years (Burnett et al., 2001).  In the 1999 and 2001 studies,

2      analyses were performed using Poisson GAM (default convergence criteria) with a

3      nonparametric LOESS prefilter applied to the pollution and hospitalization data. All studies

4      demonstrated that O3 effects were robust when adjusting for PM indices, whereas PM effects
       August 2005
7-70
DRAFT-DO NOT QUOTE OR CITE

-------
 1      from single-pollutant models were markedly attenuated when O3 was added to the regression.
 2      These results imply more robust associations with respiratory hospitalizations for O3 than PM.
 3           Results from the APHEA study indicated strong and consistent O3 effects on unscheduled
 4      hospital admissions for COPD (Anderson et al., 1997). Significant effects also were seen for
 5      BS, TSP, and NO2. The authors reported that among all pollutants examined, the most consistent
 6      and significant findings were for O3.  No two-pollutant model results were reported. Several
 7      additional studies also observed that there was no substantial difference in the O3 effect after
 8      adjusting for PM in the regression model (Gouveia and Fletcher, 2000a; Petroeschevsky et al.,
 9      2001; Ponce de Leon et al., 1996).
10           Collectively, these results suggest that copollutants generally do not confound the
11      association between O3 and respiratory hospitalizations. Ozone risk estimates were robust to PM
12      adjustment in all year and warm season only data.
13
14      7.3.4   Association of Ozone with Hospital Admissions for
15              Cardiovascular Disease
16           A subset of hospital admissions studies have examined the association of O3 with
17      cardiovascular  outcomes (see Figure 7-13).  Several have found negative or inconsistent
18      associations (Ballester et al., 2001; Burnett et al.,  1999; Fung et al., 2005; Koken et al., 2003;
19      Linn et al., 2000; Mann et al., 2002; Morgan et al., 1998a; Petroeschevsky et al., 2001;
20      Poloniecki et al., 1997; Prescott et al., 1998). Other studies, especially those that examined
21      the relationship when O3 exposures were higher, have observed robust positive associations
22      between O3 and cardiovascular hospitalizations (Atkinson et al., 1999b; Burnett et al.,  1997b;
23      Chang et al., 2005; Tsai et al., 2003a; Wong et al., 1999a,b; Yang et al., 2004a).  In Toronto,
24      Canada, Burnett et al. (1997b)  reported a positive association between O3 and cardiovascular
25      hospital admissions in a summer-only analysis. The results were robust to adjustment for
26      various PM indices, while the PM effects diminished when adjusting for gaseous pollutants.
27      Other studies stratified their analysis by temperature, warms days (>20 °C) versus cool days
28      (<20 °C). The  analysis using warms days consistently produced positive associations  (Chang
29      et al., 2005; Tsai et al., 2003a;  Yang et al., 2004a). In two studies conducted in Hong Kong,
30      total cardiovascular as well as circulatory, ischemic heart disease, and heart failure were all
31      significantly associated with O3 in the cool but not the warm season (Wong et al., 1999a,b).

        August 2005                              7-71         DRAFT-DO NOT QUOTE OR CITE

-------
                           -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.
August 2005
               7-72
DRAFT-DO NOT QUOTE OR CITE

-------
 1      In Wong et al. (1999b), O3 concentrations were similar in both seasons, with warm season levels

 2      slightly lower, mean 8-h avg O3 concentrations of 31.2 |ig/m3 (or 16.1 ppb), compared to the

 3      cool season, mean 34.8 |ig/m3 (or 18.0 ppb).  The authors speculated that differing activity

 4      patterns and home ventilation factors may have contributed to the seasonal differences in O3

 5      effects. Weather in Hong Kong is mild throughout the year, but less humid and cloudy in the

 6      cool season.  Thus, during the cool season people are more likely to open windows or stay

 7      outdoors, resulting in higher personal exposures even with similar ambient concentrations.

 8           Among the growing group of hospitalization studies that examined the effect of O3 on

 9      cardiovascular admissions, several have found inconsistent associations, especially for all year

10      analyses.  However, in studies that stratified analyses by seasonal or meteorological factors,

11      evidence is suggestive  of an association between O3 and cardiovascular hospitalizations.

12

13      7.3.5   Summary of Acute  Ozone Effects on Daily Emergency Department
14              Visits and Hospital Admissions

15          • The vast majority of emergency room visits and hospitalization studies conducted
16            over the past decade have looked at effects of O3 on either total respiratory diseases
17            and/or asthma.  Among the hospitalization studies, O3 was found to be associated with
18            both outcomes in many cases.  Studies of emergency department visits for respiratory
19            conditions also reported O3 effects, but the results tended to be less consistent across
20            studies.
21
22          • Many of the daily emergency department visits and hospitalization studies analyzed
23            O3 risk estimates using year-round data.  Given the strong seasonal variations in
24            O3 concentrations and the changing relationship between O3 and other copollutants
25            by seasons, inadequate adjustment for seasonal effects might have masked or
26            underestimated the association between O3 and the respiratory disease outcomes.
27            Season-stratified analyses typically yielded more reliable O3 effect estimates.
28
29          • Several studies have examined the association between O3 and respiratory
30            hospitalizations while controlling for other pollutants in the analytical model.
31            In most cases, O3 effects have been reported to be robust to adjustment for copollutants,
32            particularly PM. Therefore, the evidence is supportive of independent O3 effects on
33            respiratory hospital admissions.
34
35          • A subset of hospital admission studies examined the effect of O3 on cardiovascular
36            outcomes. The evidence is inconclusive regarding the association between O3 exposure
37            and cardiovascular hospitalizations in year-round data analyses. However, in the limited
38            number of studies that accounted for seasonal or meteorological factors, results suggested
        August 2005                              7-73        DRAFT-DO NOT QUOTE OR CITE

-------
 1            that O3 was associated with increased risk of cardiovascular hospital admissions during
 2            the warm season.
 3
 4
 5     7.4   ACUTE EFFECTS OF OZONE ON MORTALITY
 6     7.4.1   Summary of Key Findings on Acute Effects of Ozone on Mortality
 7             From the 1996 O3 AQCD
 8          A limited number of studies examined O3-mortality associations at the time of the previous
 9     O3 AQCD, most of which were from the 1950s and 1960s.  The 1996 O3 AQCD considered these
10     historical studies to be flawed because of inadequate adjustment for seasonal trends or
11     temperature and the use of questionable exposure indices. There were only a few time-series
12     studies that examined O3-mortality associations between the 1980s and mid-1990s. These
13     studies used more sophisticated approaches in addressing seasonal confounding and weather
14     models. One of these studies (Shumway et al., 1988) focused on the associations with long-term
15     fluctuations in Los Angeles, CA but did not examine short-term associations. A study that
16     reanalyzed the Los Angeles, CA data with a focus on the short-term associations (Kinney and
17     Ozkaynak, 1991) did find that, of the PM and gaseous criteria pollutants, O3 (reported as total
18     oxidants) was most strongly associated with total nonaccidental mortality.  Then two studies, one
19     using Detroit, MI data (Schwartz, 1991) and the other using St. Louis, MO and Kingston-
20     Harriman, TN data (Dockery et al., 1992), reported that PM but not O3 was significantly
21     associated with mortality. However, the 1996 O3 AQCD discussed that without sufficient
22     presentation of model specifications, it was difficult to evaluate whether the lack of O3-mortality
23     associations was possibly due to mis-specification of the weather model. In summary, due to the
24     insufficient number of studies that examined O3-mortality associations and the uncertainties
25     regarding weather model specifications, the 1996 O3 AQCD was unable to quantitatively assess
26     O3-mortality excess risk estimates, or even provide qualitative assessment of the likelihood of
27     O3-mortality associations.
28
29     7.4.2   Introduction to Assessment of Current Ozone-Mortality  Studies
30          Introductory discussions of the PM-mortality effects often cite historical air pollution
31     incidents such as the 1952 London, England smog episode in which thousands of deaths were
32     attributed to the air pollution from coal burning. There is no counterpart "historical episode" for

       August 2005                             7-74        DRAFT-DO NOT QUOTE OR CITE

-------
 1      O3-mortality effects. Instead, the early recognition of the adverse health effects of summer
 2      oxidant air pollution, mainly from Los Angeles and other major cities with a high density of
 3      automobiles, were based on symptoms such as eye and throat irritations.  Thus, the focus of PM
 4      epidemiology  and that of O3 epidemiology have been historically different.
 5           As shown in Table AX7-5 in the Chapter 7 Annex, the number of short-term mortality
 6      studies that analyzed O3 has increased markedly since the last publication of the O3 AQCD in
 7      1996. The increased attention to PM-mortality associations in the early 1990s lead to the
 8      increase in studies that also examined O3, most often as a potential confounder for PM.
 9      Although many of these PM studies also reported O3 estimates, they often lacked specific
10      hypotheses regarding mortality effects of O3 as the focus of these studies was to examine the
11      PM-mortality  effect. This is in contrast to the O3-morbidity studies, most of which were
12      specifically designed to  examine effects of "summer haze" and O3 (or oxidants) on respiratory
13      and other symptoms, lung functions, and emergency department visits, etc. However, new
14      studies with hypotheses  developed specifically  for O3 effects on mortality have become
15      available, such as the large U.S. 95 communities study by Bell et al. (2004), the U.S. 14 cities
16      study by Schwartz (2005), and the 23 European cities study by Gryparis et al. (2004) discussed
17      in the next section.
18
19      7.4.3   Single-Pollutant Model Ozone-Mortality Risk Estimates
20           To facilitate a quantitative overview of the O3-mortality effect estimates and their
21      corresponding uncertainties, the percent excess risks of total nonaccidental mortality calculated
22      using all year  data are plotted in Figures 7-14 and 7-15.  Studies that only conducted seasonal
23      analyses will be presented in the next section.  These figures do not include studies that only
24      examined cause-specific mortality. Figure 7-14 only presents the results from single-day lag
25      models.  Results from multiday lag models are shown in Figure 7-15. All effect estimates are
26      from single-pollutant models and include all age groups unless noted otherwise.  The majority of
27      the estimates are positive with a few exceptions. Five multicity studies, three from the U.S.
28      (Bell et al., 2004; Samet et al., 2000 [reanalysis Dominici et al., 2003]; Schwartz, 2005) and two
29      from Europe (Gryparis et al.,  2004; Touloumi et al., 1997), also showed generally positive
30      associations.
31

        August 2005                              7-75        DRAFT-DO NOT QUOTE OR CITE

-------
                    Bell etal. (2004): U.S. 95 communities -
  Sametetal.(2000;reanalysisDominicietal.,2003): U.S.90cities -
                          Schwartz(2005): U.S.14cities-
        Kinney and Ozkaynak (1991): Los Angeles County, CA —
               Kinney etal. (1995): Los Angeles County, CA —
                    Fairley (2003): Santa Clara County,CA -
                      Dockeryetal.(1992): St.Louis,MO -
        Lippmannetal.(2000;reanalysislto,2004): Detroit,Ml -
               Chock etal. (2000): Pittsburgh, PA (age 0-74) -
                  Dockery et al. (1992):  Eastern Tennessee —
                   Klemm and Mason  (2000): Atlanta, GA -
  Touloumi et al. (1997): 4 European cities (Athens, Barcelona, London, Paris) -
                  Anderson etal. (1996): London, England —
                   Bremneretal.(1999): London,England —
                 Prescottetal.(1998): Edinburgh.Scotland -
                       Zmirouetal.(1996): Lyon,France —
     Hoek etal. (2000; reanalysis Hoek, 2003): The Netherlands -
   Roemerand van Wijnen (2001): Amsterdam, the Netherlands —
          Verhoeff etal.(1996): Amsterdam,the Netherlands -
           Peters et al. (2000b): Coal basin in Czech Republic -
                 Petersetal.(2000b): NEBavaria,Germany -
             Garda-Aymerich etal.(2000): Barcelona,Spain —
                    Sunyeretal.fi 996): Barcelona, Spain —
                   Borja-Aburto et al. (1997): Mexico City -
                 Borja-Aburto et al. (1998): SW Mexico City -
             Gouveia and Fletcher (2000b): Sao Paulo, Brazil -
                      Ostroetal.(1996): Santiago,Chile -
                  Morgan etal. (1998b): Sydney,Australia -
                 Simpson etal.(1997): Brisbane, Australia —
                          Kim etal.(2004): Seoul,Korea -
                          Leeetal.(1999): Seoul,Korea -
                   Lee and Schwartz (1999): Seoul, Korea -
                          Lee etal. (1999): Ulsan, Korea -
                                                                   % Change in Mortality
                                                           -10
                                                              i
                      0
                      I
                       10
                        I
20
 I
                                                   Latin America
                   -Iag1
                                                   Australia
                                      © Multicity combined
                                      • Single city
                          lagO
                          lagO
                          Iag1
                          Iag1
                                   lagO
                           — Iag1
                            lagO
                           - lagO
                                 lagO
                                      Best of lag 0 to 3 by city
                                                 lagO
                            Iag1
                                 • lagO
                          - lagO
                          -•	lagO
                                    lagO
                           lagO
                          -  lagO
                           lagO
                                 lagO
                                	•-
                                              lagO
                                                                                            • lag 1
                           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.
August 2005
7-76
DRAFT-DO NOT QUOTE OR CITE

-------
                             Bell etal. (2004): _
                          U.S.95 communities

                      Gamble(1998): Dallas, TX-

                       ltoandThurston(1996):_
                             Cook County, IL

                         Lippmann etal.(2000;
                  reanalysislto,2004): Detroit, Ml"

                          Lipfert et al. (2000a): _
                        7 counties in PA and NJ
                          Gryparisetal.(2004): _
                            23 European cities

                 Touloumi et al. (1997): 4 European cities
                     (Athens, Barcelona, London, Paris)"

                             Hoeketal.(2000;_
            reanalysis Hoek, 2003): The Netherlands

                  Roemerand van Wijnen (2001):
                    Amsterdam, the Netherlands
                      Borja-Aburto etal.(1998):
                              SW Mexico City

                  O'Neill etal. (2004): Mexico City _
                                                        % Change in Mortality
                                                  -10
                                                    i
                                           Latin America
             0
              I
               10
                I
20
 I
                  lag 0-3
            lag 0-1  • SE not given; significant at p = 0.055
                             Multicity combined

                             Single city
                          11-2
                        lag 0-1
                  lag 0-1
                        Best of cumulative lags
                        up to 5 days by city
                     lag 0-6
             «	lag 0-6
                  - lag 1-2


                   lag 0-1
       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.
1            The initial primary objective of the original NMMAPS (Samet et al., 2000; reanalysis

2      Dominici et al., 2003) was to investigate the effects of PM, but the study also comprehensively

3      examined mortality risk estimates from gaseous pollutants in 90 U.S. cities over the period of

4      1987 to 1994.  Among the 90 cities, 80 monitored O3 either year-round or during the warm

5      season.  The study illustrated that the mortality risk estimates for O3 varied by season.  The

6      estimate using all available data was about half of that for summer-only data at a lag of 1-day
       August 2005
7-77
DRAFT-DO NOT QUOTE OR CITE

-------
 1      (see Section 7.6.3.2 for further discussion).  Bell et al. (2004) extended the original NMMAPS
 2      by adding six more years (from 1987 to 2000) and 15 more communities (a total of
 3      95 communities), and examined the effects of O3 on mortality. The results of this study are
 4      discussed in detail here because of the study's emphasis on U.S. data and the inclusion of 95
 5      large communities across the country, making this mortality study most representative of the
 6      U.S. population. In addition, this study is one of the few that have focused specifically on O3
 7      hypotheses testing and investigated several important issues.  Among the 95 communities
 8      examined in this study, 55 monitored O3 throughout the year and 32 only monitored during the
 9      warm season. Eight additional  cities switched from warm season only to year-round monitoring
10      or year-round to warm season only monitoring at some point during the study period. The mean
11      24-h avg O3 concentration was approximately 26 ppb for the 95 communities.
12      Within-community results first  were calculated using single-day lags of 0,  1, 2, and 3 days, and
13      a 7-day distributed lag in O3 exposure. A two-stage Bayesian hierarchical model was used to
14      determine a national average effect estimate, taking into consideration city-to-city variation.
15      Figure 7-16 presents the Bayesian community-specific and national average O3 risk estimates for
16      total mortality per 20 ppb increase in 24-h avg O3 from a constrained 7-day distributed lag
17      model.  The Bayesian community-specific estimates were  shrunk to the national average
18      estimate by a factor that was inversely proportional to the  heterogeneity of the community -
19      specific relative rates.  The heterogeneity of the effect estimates from the individual cities is
20      partially attributable to differences in pollution characteristics, the use of air conditioning,  time-
21      activity patterns, and socioeconomic factors. Due  to the random variation as well as  the smaller
22      sample sizes within each city, emphasis is given to the national average effect estimate.
23           In the U.S. 95 communities study, the largest risk estimate for O3-mortality was obtained
24      with a 0-day lag, followed by diminishing risk estimates with 1-, 2-, and 3-day lags (Figure
25      7-17). Ozone exposure at a 0-day lag was associated with a 0.50% (95% PI:  0.24, 0.78) excess
26      risk in mortality per 20 ppb increase in 24-h avg O3.  The 7-day distributed lag model, which
27      examined the cumulative effect from the same day and six previous days, also is shown in
28      Figure 7-17. A cumulative excess mortality risk of 1.04% (95% PI: 0.54,  1.55) per 20 ppb
29      increase in 24-h avg O3 during the previous week was observed.  In a related U.S. study of the
30      19 largest cities by Huang et al. (2005), the  O3 estimate for the summer season was 1.47% (95%
31      PI: 0.54,  2.39) excess  risk of cardiopulmonary mortality with current-day exposure.  Smaller

        August 2005                              7-78        DRAFT-DO NOT QUOTE OR CITE

-------
                                     % Change in Mortality



_ uonvcr
aan Antonio


Loxingion
Dirmingrtam
St P t" I6"0
at. ctcr_Durg
«? h I


P ° .
A*l .*
Ok I h c^t
« , ~J P p H
5 rcaplu—
-jacrorncniQ
umaria
t fi /A h
3nI3 AnD i AnanOirn
1 j, ':3'T!Pa




D M ^^"e


fnaianapofi-


Lorpu- >.nri_ti
F^ileiah


Aicron


... P.
H ff I




Jefsev Citw
jcrocy VIW


Alb
AfDuqucrquc
P
M f~rc_no
















Lincoln
Wa^hrnciton






Houston
p. .


Honolulu
D i .
... , -


Swrarusp








NATIONAL
3-2-1012345
III I I I I



























*



*



^



~
: *
*
"

: *_
*

: *












•

»
; *
~

_

^

~.



: ~






; *
| ^


*
| ^



'• *.






Figure 7-16.  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 (32 of the 95 communities only had
             warm season data).

Source: Derived from Bell et al. (2004).
August 2005
7-79
DRAFT-DO NOT QUOTE OR CITE

-------
                                              0.5             1
                                           % Change in Mortality
                         1.5
       Figure 7-17.  Comparison of single-day lags (0-, 1-, 2-, and 3-day) to a cumulative multiday
                     lag (0- to 6-day) for percent changes in all cause mortality per 20 ppb
                     increase in 24-h avg O3 in all ages.
       Source: Derived from Bell et al. (2004).
 1     effects also were observed with 1- and 2-day lags of exposure.  The effect estimate for the 7-day
 2     distributed lag was 2.52% (95% PI: 0.94, 4.10) excess risk of cardiopulmonary mortality. These
 3     findings suggest that the effect of O3 on mortality is immediate, but also may persist over
 4     multiple days.
 5           The influence of higher O3 levels on the risk estimate also was evaluated in the U.S.
 6     95 communities study.  When the data were restricted to days with 24-h avg O3 levels less than
 7     60 ppb for the 1-day lag analysis, the national estimate did not substantially change (0.30%
 8     [95% PI:  0.06, 0.54] per 20 ppb increase for days with levels below 60 ppb versus 0.36% [95%
 9     PI: 0.12, 0.61] for all days).  These results  suggest that the O3-mortality associations occur at
10     24-h avg O3 levels below 60 ppb.
11           Schwartz (2005) examined O3-mortality associations using data from 14 U.S. cities.
12     A case-crossover study design was used to  compare the influence of adjustment methods for
13     temperature (regression splines of temperature versus matching case and control periods by
14     temperature).  The risk estimate obtained by matching (0.92% [95% CI: 0.06, 1.80] per 40 ppb
15     increase in 1-h max O3) was similar to that  obtained with regression splines (0.76% [95% CI:
       August 2005
7-80
DRAFT-DO NOT QUOTE OR CITE

-------
 1      0.13,1.40]), suggesting that the O3-mortality risk estimates were not sensitive to these
 2      adjustment methods for temperature.
 3           The APHEA 1 project (Touloumi et al., 1997) reported a pooled random effects estimate of
 4      4.5% (95%  CI:  1.6, 7.7) per 40 ppb increase in 1-h max O3 using the best single-day lag model
 5      results from four European cities (London, Athens, Barcelona, and Paris). As an extension of
 6      the four European cities study, researchers of the APHEA 2 project investigated the effect of O3
 7      on total, cardiovascular, and respiratory mortality in 23 cities throughout Europe (Gryparis et al.,
 8      2004). Ozone data was available year-round in all 23 cities.  A cumulative lag of 0 to 1 days was
 9      hypothesized a priori. A two-stage hierarchical model, which accounted for statistical variance
10      and heterogeneity among cities, was used to estimate the pooled regression coefficients. Due to
11      substantial heterogeneity among cities, random effects regression models were applied. The
12      pooled effect estimate for the 23 European cities (0.23% [95% CI:  -0.85, 1.95]  per 40 ppb
13      increase in 1-h max O3 for all seasons) was positive but considerably smaller compared to that
14      obtained in the APHEA 1 study. The researchers noted that there was a considerable seasonal
15      difference in the O3 effect on mortality, thus the small effect for the all year data might be
16      attributable to inadequate adjustment for confounding by season. This seasonal effect will be
17      discussed further in the next section.
18           Collectively, the single-pollutant model estimates from the single- and multiple-city
19      studies shown in Figures 7-14 and 7-15 suggest an excess risk of total nonaccidental mortality
20      associated with acute O3 exposure.  Despite the different analytical approaches and alternative
21      model specifications used in the various studies, overall, the range of estimates were relatively
22      narrow, with most of the positive estimates falling in the range from 0.5 to 5% excess risk in
23      mortality per standardized increment.
24
25      7.4.4   Meta-analyses  of O3-Mortality Risk Estimates
26           Several studies in recent years conducted meta-analyses of O3-mortality associations (Levy
27      et al., 2001; Stieb et al., 2002, 2003; Thurston and Ito, 2001; World Health Organization,  2004).
28      Figure 7-18 presents the combined O3  risk estimates from the various meta-analyses. Most of
29      these studies included GAM studies using default convergence criteria except Stieb et al.  (2003),
30      which compared effect estimates from GAM-affected studies to non-GAM studies.  All of these
31      meta-analyses reported fairly consistent and positive combined estimates, approximately 2%

        August 2005                               7-81        DRAFT-DO NOT QUOTE OR CITE

-------
                                                          % 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)
                                                         1
                                              4
                                              I
       Figure 7-18.  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.
1      excess total nonaccidental mortality per standardized increment (see Section 7.1.3.2). However,

2      most of these studies were not analytical in design in that they did not attempt to examine the

3      source of heterogeneity, although one suggested an influence of weather model specification

4      (Thurston and Ito, 2001) and another reported evidence of publication bias (World Health

5      Organization, 2004) in the past literature. None of these studies address the issue of season-

6      specific estimates, therefore, interpreting these combined estimates requires caution.

7           Most recently, three research groups conducted independent meta-analyses of O3-mortality

8      associations (Bell et al., 2005; Ito et al., 2005; Levy et al., 2005).  These analyses attempted to

9      evaluate the source of heterogeneity using the most up-to-date literature database.  These
       August 2005
7-82
DRAFT-DO NOT QUOTE OR CITE

-------
 1      analyses were also systematically compared and discussed (Bates, 2005; Goodman, 2005). The
 2      all-season combined point estimates per standardized increment from these three meta-analyses
 3      were remarkably consistent:  1.75% (95% PI: 1.10, 2.37), 1.6% (95% CI: 1.1, 2.0), and 1.64%
 4      (95% CI: 1.25, 2.03), for the Bell et al., Ito et al., and Levy et al.  studies, respectively.  All three
 5      studies also indicated that the estimates were higher in warm seasons. Each of these studies is
 6      briefly summarized below. Their findings related to specific issues are discussed later in the
 7      corresponding sections.
 8           Bell et al. (2005) conducted a meta-analysis of 144 effect estimates from 39 U.S. and
 9      non-U.S. studies and estimated pooled effects by lags, age groups, specific causes, and exposure
10      metrics. The results were also compared with their NMMAPS results (Bell et al., 2004).
11      A two-stage Bayesian hierarchical model was used to estimate the combined estimate by taking
12      into account the within-city variance (the statistical uncertainty) and between-study variance (the
13      heterogeneity across cities). They concluded that the results provided strong evidence of a short-
14      term association between O3 and mortality that was not sensitive to adjustment for PM or model
15      specifications (discussed in Section 7.4.6).  However, they suggested that, based on comparisons
16      between the meta-analysis results and NMMAPS results, there was evidence of publication bias
17      (1.75% [95% CI:  1.10, 2.37] per 20 ppb increase in 24-h  avg O3 for meta-analysis versus 0.50%
18      [95% CI: 0.24, 0.78] for NMMAPS 0-day lag results).
19           Ito et al. (2005) conducted a meta-analysis of 43 U.S. and non-U.S. studies but also
20      analyzed data from 7 U.S. cities to further examine the issues identified in their meta-analysis.
21      Adjusting for PM did not substantially influence the O3-mortality effect estimates in either the
22      meta-analysis or 7 U.S. cities analysis. The multicity analysis further indicated that the
23      difference in the weather adjustment model could result in a twofold difference in risk estimates
24      (e.g., 1.96% versus 0.96% in multicity combined estimates across alternative weather models for
25      the O3-only, all year case). In the meta-analysis, they found suggestive evidence of publication
26      bias (a significant asymmetry in the funnel plot),  but adjusting for the asymmetry reduced the
27      combined estimate only slightly (from 1.6% [95% CI: 1.1, 2.0] to 1.4% [95% CI: 0.9,  1.9] per
28      20 ppb increase in 24-h avg O3). The  extent of potential bias implicated in this study differed
29      compared to that in Bell et al. (2005).  The source of this difference is not clear, but Ito et al.
30      state that sensitivity analyses comparing estimates from commonly used weather model
        August 2005                               7-83        DRAFT-DO NOT QUOTE OR CITE

-------
 1      specifications suggest that the stringent weather model used in NMMAPS may tend to yield
 2      smaller risk estimates than those used in other studies.
 3           Levy et al. (2005) analyzed 48 estimates from 28 studies from the U.S., Canada, and
 4      Europe using an empiric Bayesian meta-regression with covariates including the relationship
 5      between O3 and other pollutants, proxies for the relationship between personal exposure and
 6      ambient concentration such as air conditioning prevalence,  and statistical methods used. They
 7      found that the air conditioning prevalence (a greater effect in cities with less air conditioning)
 8      and lag time (same-day  effects larger than lagged effects) were the strongest predictors of
 9      between-study variability. The warm season estimates were larger than the cool season
10      estimates.  The influences of copollutants were inconsistent, but they found a potential influence
11      of summertime PM2 5.
12           As stated earlier, the combined O3 excess mortality risk estimates from the meta-analyses
13      by Bell et al., Ito et al., and Levy et al. were very  consistent. Although the analyses were
14      conducted independently, there was considerable  overlap among the estimates used in the three
15      meta-analyses; thus, the agreement in the combined risk estimates was not unexpected.  The
16      common findings among these three meta-analyses, aside from the consistency in their  combined
17      estimates, include:  (1) no difference in estimates  between GAM studies using default versus
18      stringent convergence criteria; (2) estimates were larger in warm seasons; and (3) no strong
19      indication of PM confounding. Both Bell et al. and Levy et al. studies found that the estimates at
20      lag 0-day were larger than longer lags.  Both the Bell et al. and Ito et al. studies suggested
21      evidence of publication bias.  These three studies, along with the earlier meta-analyses, provide
22      strong evidence that O3  is associated with mortality. The combined effect estimates from the
23      various meta-analyses ranged from 1.5 to 2.5% excess risk in all cause mortality.
24
25      7.4.5   Seasonal Variation in Ozone-Mortality Risk Estimates
26           Since the seasonal cycle of O3 follows the seasonal cycle of temperature (which is
27      inversely related to the mortality seasonal cycle),  inadequate adjustment of temporal trends in
28      the regression model may lead to negative O3-mortality risk estimates. In addition, as discussed
29      in Section 7.1.3.5, in some cities low-level O3 during the winter may be negatively correlated
30      with PM and other primary pollutants, resulting in negative correlations between O3 and
        August 2005                               7-84        DRAFT-DO NOT QUOTE OR CITE

-------
 1      mortality even in short-term relationships.  The confounding effect by season could be
 2      substantially reduced by conducting season-stratified analyses.
 3           A fewer number of O3-mortality studies performed seasonal analyses.  Figure 7-19 presents
 4      the studies that reported O3 risk estimates for all cause mortality by season.  For those studies
 5      that obtained O3 risk estimates for each of the four seasons, only  summer and winter results are
 6      shown. The estimates for year-round data analyses, when available, also are shown for
 7      comparisons.  In all the studies, the O3 risk estimates are larger during the warm season than the
 8      cool season, with the all year estimates generally in between the two seasonal estimates.
 9           In three U.S. and European multicity studies (Gryparis et al., 2004; Samet et al., 2000
10      [reanalysis Dominici et al., 2003]; Schwartz, 2005), season-stratified analyses indicated that the
11      O3-mortality effect estimates were significant and positive in the  warm season,  with larger
12      effects observed compared to the year-round analyses. The effect estimates from the cool season
13      were notably smaller and less significant. In the case of the U.S.  90 cities study (of which
14      80 cities had O3 data available) the winter (December, January, and February) mortality estimate
15      was negative, which was most likely attributable to the inverse relationship between O3 and PM
16      in the winter.
17           In the U.S. 95 communities study by Bell et al. (2004), no significant difference was
18      observed between the estimates from all available data and warm season only data (April-
19      October); cool season only analyses were not performed. The warm season effect estimate using
20      the 7-day constrained distributed lag model was 0.78% (95% PI:  0.26, 1.30) excess risk per 20
21      ppb increase in 24-h avg O3, compared to 1.04% (95% PI: 0.54,  1.55) calculated  using all
22      available data.  In the 55  communities with year-round O3 data, the all year effect estimate was
23      0.96% (95% PI: 0.32,  1.57).
24           All  three recent meta-analyses (Bell et al., 2005; Ito et al., 2005; Levy et al.  2005), found
25      that the estimates for warm seasons were larger than all year estimates.  In Bell et al., the warm
26      season estimate was 3.02% [95% PI: 1.45, 4.63], compared to the all year estimate of 1.75%
27      [95% PI:  1.10, 2.37]. In the subset of 10 cities examined in Ito et al., the warm season and all
28      year estimates were 3.5% [95% CI:  2.1, 4.9] and 2.2% [95% CI:  0.8, 3.6], respectively.
29      Likewise, Levy et al. observed a 3.38% [95% CI: 2.27, 4.42] excess risk in the warm season
30      compared to a 1.64% [95% CI:  1.25, 2.03] excess risk using all year data. All  results presented
31      are percent excess risk in mortality per standardized increment.

        August 2005                              7-85       DRAFT-DO NOT QUOTE OR CITE

-------
                                                         % 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 Bernadino County
                      and Riverside County, CA
                      Gamble (1998): Dallas, TX-


          Moolgavkar et al. (1995): Philadelphia, PA-


        Chocket al. (2000): Pittsburgh, PA (age 0-74)-


             Vedal et al. (2003): Vancouver, Canada -




           Gryparis et al. (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-


             Cifuentes et al. (2000): Santiago, Chile-


                Ostro et al. (1996): Santiago, Chile-




           Simpson et al. (1997): Brisbane, Australia —



                   Kim etal. (2004): Seoul, Korea-
                                                 -10
                                   10
                                    I
                                   20
                                    I
30
                                           U.S. and Canada
                   -e-
                       *
                     -e-
lag 1 -2
       lagO
| Latin America I
            lag 0-6


             Iag1


             lag 0


             lagO
                            •  SE not given; significant at p = 0.05 level
                          O   SE not given; not significant at p = 0.05 level
                            lagO
                                                                  -X-
                                                               -e-
                             lag 0-1
                                lagO
                             lag 1
                                   lagO
                          lagO
                         lag 1-2
                                             lagO
                                                                           lag 1
Figure 7-19.   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.
August 2005
            7-86
               DRAFT-DO NOT QUOTE OR CITE

-------
 1           Studies that conducted analysis by season indicate that O3-mortality risk estimates are often
 2      larger in the warm season compared to the colder season. The seasonal dependence of
 3      O3-mortality effects complicates interpretation of O3 risk estimates calculated from year-round
 4      data without adequate adjustment of temporal trends.
 5
 6      7.4.6   Ozone-Mortality Risk Estimates Adjusting for PM Exposure
 7           The confounding between "winter type" pollution (e.g., CO, SO2, and NO2) and O3 is not
 8      of great concern because the peaks of these pollutants do not strongly coincide. The main
 9      confounders of interest for O3, especially for the northeast U.S., are "summer haze" type
10      pollutants such as acid aerosols and sulfates. Since very few studies had these chemical
11      measurements, PM (especially PM2 5), may serve as surrogates.  However, due to the expected
12      high correlation among the constituents of the "summer haze mix," multipollutant models
13      including these pollutants may result in unstable coefficients, and therefore, an interpretation of
14      such results requires some caution.
15           Figure 7-20 shows the O3 risk estimates with and without adjustment for PM indices using
16      all year data in studies that conducted two-pollutant analyses. Approximately half of the O3 risk
17      estimates slightly increased while the other half slightly decreased in value with the inclusion of
18      PM in the models. In general, the O3-mortality risk estimates were robust to adjustment for PM
19      in the models, with the exception of Los Angeles, CA data with PM10 (Kinney et al., 1995) and
20      Mexico City data with TSP (Borja-Aburto et al., 1997).
21           The U.S. 95 communities study by Bell et al. (2004) examined the sensitivity of acute
22      O3-mortality effects to potential confounding by PM10. Restricting analysis to days when both
23      O3 and  PM10 data were available, the community-specific O3-mortality effect estimates as well as
24      the national average results indicated that O3 was robust to adjustment for PM10 (Bell et al.,
25      2004).  There  was insufficient data available to examine potential confounding by PM2 5.
26      One study (Lipfert et al., 2000a) reported  O3 risk estimates with and without sulfate adjustment.
27      Lipfert  et al. (2000a) calculated O3 risk estimates based on mean (45 ppb) less background (not
28      stated)  levels of 1-h max O3 in seven counties in Pennsylvania and New Jersey. The O3 risk
29      estimate was not substantially affected by the addition of sulfate in the model (3.2% versus
30      3.0% with sulfate) and remained statistically significant.
        August 2005                              7-87        DRAFT-DO NOT QUOTE OR CITE

-------
                                         -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 7-20.   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.
August 2005
                DRAFT-DO NOT QUOTE OR CITE

-------
 1           Several O3-mortality studies examined the effect of confounding by PM indices in different
 2      seasons (Figure 7-21). In analyses using all year data and warm season only data, O3 risk
 3      estimates were once again fairly robust to adjustment for PM indices, with values showing both
 4      slight increases and decreases with the inclusion of PM in the model. In the analyses using cool
 5      season data only, the O3 risk estimates all increased slightly with the adjustment of PM indices,
 6      although none reached statistical significance.
 7           The three recent meta-analyses (Bell et al., 2005; Ito et al., 2005; Levy et al. 2005) all
 8      examined the influence of PM on O3 risk estimates. No substantial influence was observed in
 9      any of these studies. In the analysis by Bell et al., the combined estimate without PM adjustment
10      was 1.75% (95% PI:  1.10, 2.37) from 41  estimates, and the combined estimate with PM
11      adjustment was 1.95% (95% PI: -0.06, 4.00) from 11 estimates per 20 ppb increase in 24-h avg
12      O3.  In the meta-analysis of 15 cities by Ito et al., the combined estimate was 1.6% (95% CI: 1.1,
13      2.2) and 1.5% (95% CI:  0.8,  2.2) per 20 ppb in 24-h avg O3 without and with PM adjustment,
14      respectively. The additional time-series analysis of six cities by Ito et al. found that the
15      influence of PM by season varied across alternative weather models but was never substantial.
16      Levy et al. examined the regression relationships between O3 and PM indices (PM10 and PM2 5)
17      with O3-mortality effect estimates for all year and by season. Positive slopes, which might
18      indicate potential confounding, were observed for PM2 5 on O3 risk estimates in the  summer and
19      all year periods, but the relationships were weak. The effect of one causal variable  (i.e., O3)
20      is expected to be overestimated when a second causal variable (e.g., PM) is excluded from the
21      analysis if the two variables are positively correlated and act in the same direction.  However,
22      the results from these meta-analyses as  well as several single- and multiple-city studies indicate
23      that copollutants generally do not appear to substantially confound the association between O3
24      and mortality.
25
26      7.4.7   Ozone Risk Estimates for Specific Causes of Mortality
27           In addition to all cause mortality,  several studies examined broad underlying causes of
28      mortality, such as cardiovascular and respiratory causes.  The U.S. 95 communities  study
29      (Bell et al., 2004) analyzed O3 effect estimates from cardiovascular and respiratory  mortality.
30      Significant effects were seen  at 0- and 2-day lags with results similar to total mortality.  The
31      national average estimate from the constrained distributed lag model was slightly greater for

        August 2005                               7-89       DRAFT-DO NOT QUOTE OR CITE

-------
                                                     % Change in Mortality
                                  -15
        Ostro (1995): San Bernadino County
               and Riverside County, CA —
                          with PM2.5
    Moolgavkaretal. (1995): Philadelphia, PA •
                           withTSP -
                            03 only -
                           with TSP •
  Chock et al. (2000): Pittsburgh, PA (age 0-74) •
                          with PM10 •
                            O3 only -
                          with PM10 •
                            03 only —
                          with PM10 -
     Gryparis et al. (2004): 21 European cities —
                            03 only —
     Anderson et al. (1996): London, England •
                            with BS -
                            03 only •
                            with BS -
                            O3 only •
                            with BS -
         Ostro et al. (1996): Santiago, Chile -
                            03 only —
             Kim et al. (2004): Seoul, Korea -
                          with PM10 -
                            O3 only —
                          with PM10 -
         -10
         J	
         -5
          I
5
I
10
                           -e-
                              -e-
                      -e-
                      -e-
                                   I Europe
                                    Latin America
                                    Asia
X All year
Warm  Q Cool
                                        lag 0
                                              lag 1
                                            lagO
                                                                                 lag 0-1
                                                 lagO
                                    -e-
                                                                           -e-
                                                                                      Iag1
                                            -X-
                                                    lag 1
Figure 7-21.  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.
August 2005
          7-90
           DRAFT-DO NOT QUOTE OR CITE

-------
 1      cardiopulmonary deaths than deaths from all causes, with an excess risk of 1.28% (95% PI:
 2      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
 3      preceding week.  In a related study, Huang et al. (2005) examined O3 effects on cardiopulmonary
 4      mortality during the summers (June to September) of 1987 to 1994 in 19 large U.S. cities from
 5      the NMMAPS database.  In the 7-day distributed lag model, the O3 effect estimate was
 6      2.52% (95% PI:  0.94, 4.10) excess risk in cardiopulmonary mortality per 20 ppb increase in
 7      24-h avg O3.
 8           Figure 7-22 presents the effect estimates of the association between O3 and cardiovascular
 9      mortality for all year and warm season analyses.  All studies, with the exception of Ponka et al.
10      (1998), showed positive associations between O3 and cardiovascular mortality.  However, as
11      with all cause mortality, there appears to be heterogeneity in the effect estimates across studies.
12      The cardiovascular mortality estimate from the meta-analysis by Bell et al. (2005) appears to be
13      close to the mode of the effect estimates from the various studies, as shown in Figure 7-22.  This
14      is expected as many of these  studies are included in the meta-analysis.  Bell et al. observed that
15      the posterior mean estimate for cardiovascular causes (2.23% [95% PI:  1.36, 3.08] excess risk
16      per 20 ppb increase in 24-h avg O3 from 25 estimates) was slightly larger than that for total
17      mortality (1.75% [95% PI:  1.10, 2.37] excess risk from 41 estimates).  However, since
18      cardiovascular deaths account for the largest fraction (over 40%)  of total deaths, it is not
19      surprising that the risk estimates for cardiovascular mortality are somewhat similar to those from
20      all cause mortality.  Overall, the cardiovascular mortality risk estimates in the current literature
21      show consistently positive  associations with some heterogeneity (most estimates fall within the
22      range of 1 to 8%  per 40 ppb increase in 1-h avg O3).
23           Several studies observed that the risk estimates for the respiratory category were larger
24      than the cardiovascular and total nonaccidental categories (e.g., Anderson et al., 1996; Gouveia
25      and Fletcher, 2000b; Gryparis et al., 2004; Zmirou et al., 1998).  In the European 21 multicities
26      study (Gryparis et al., 2004), the warm  season effect estimate for  respiratory mortality was
27      6.75% (95% CI:  4.38, 9.10) excess risk per 30 ppb increase in 8-h max O3, compared  to 2.70%
28      (95% CI: 1.29, 4.32) for cardiovascular mortality and 1.82% (95% CI:  0.99, 3.06) for total
29      mortality. In contrast, other studies have found that the risk estimates for the respiratory
30      category were smaller or even negative while the risk estimates for total or cardiovascular
31      categories were positive (e.g., Borja-Aburto et al., 1998; Bremner et al., 1999; Lipfert  et al.,

        August 2005                               7-91        DRAFT-DO NOT QUOTE OR CITE

-------
                                                                                 % Change in Cardiovascular Mortality
OQ
 to
 o
 o
 
 o
 H
 W
 O

 O
 HH
 H
 W
                     Ponka et al. (1998): Helsinki, Finland (age < 65) -

                          Gwynn et al. (2000): Buffalo, NY region -

                       * Le Tertre et al. (2002a): Strasbourg, France -

                        * Goldberg et al. (2001): Montreal, Canada -

                        Lippmann el al, (2000; reanalysis Ito, 2003): _
                                         Detroit Area, Ml
                      * Moolgavkar (2003): Los Angeles County, CA -

                        * Le Tertre et al. (2002a): Le Havre, France -

                          * Le Tertre et al. (2002a): Lyon. France -

                        Peters et al (2000b):  NE Bavaria, Germany —
                         Hoek et al. (2000; reanalysis Hoek, 2003); _
                                         the Netherlands "

                          * Le Tertre et al. (2002a): Paris, France -

                            De Leon et al. (2003): New York City -

                         Morgan et al. (1998b): Sydney, Australia -
                          Fairley (1999; reanalysis Fairley, 2003):
                                    Santa Clara County, CA"

                         Anderson et al, (1996): London, England -

                 Prescott et al. (1998):  Edinburgh, Scotland (age > 65) -

                        * Le Tertre et al. (2002a): Toulouse, France -

                           Sunyer et al. (1996): Barcelona, Spain -

                            * Moolgavkar (2003): Cook County. IL -

                         Simpson etal. (1997): Brisbane, Australia -

                         * Le Tertre et al. (2002a): Rouen. France -

                          Vedal et al. (2003): Vancouver, Canada -



                         Gryparis et al. (2004): 21 European Cities —

                     Bell et al. (2005): Meta-analysis of 25 estimates,
                             from 18 U.S. and non-U.S. studies"
0 -15 -10 -5 0 5 10 15 20 25 30 35 4
i i i i i i i i i i i
Single-city Studies



X All year
• Warm
	 v. 	 i-,rj n

-K- lagO
	 |on n

33 ° ° 9

-X- lag 1
— K- lagO
	 ,\^, 	 inn n









Mullicity and Meta-analysis Studies
-•— lag 0-1
-X™ various short-term lags (single/multiday lags of 0 to 2)
           Figure 7-22.   Ozone-associated cardiovascular mortality risk estimates (95% CI) per standardized increment
                              (see Section 7.1.3.2).  Analyses include all ages unless otherwise noted.  *Note that Goldberg et al.
                              (2001), Le Tertre et al. (2002a), and Moolgavkar (2003) performed analyses using Poisson GAM
                              with default convergence criteria.

-------
 1      2000a; Morgan et al., 1998b).  The apparent inconsistencies across studies may be due in part to
 2      the differences in model specifications, but they may also reflect the lower statistical power
 3      associated with the smaller daily counts of the respiratory category (usually accounting for less
 4      than 10% of total deaths) compared to the larger daily counts for the cardiovascular category
 5      (approximately 40 to 50% of total deaths). Thus, an examination of the differences in risk
 6      estimates across specific causes requires a large population and/or a long period of data
 7      collection. In the meta-analysis by Bell et al. (2005), which combined 23 estimates from
 8      17 studies for respiratory mortality, the effect estimate for respiratory causes was smaller (0.94%
 9      [95% PI:  -1.02, 2.96] excess risk per 20 ppb increase in 24-h avg O3) compared to the estimates
10      for total mortality (1.75% excess risk) and cardiovascular mortality (2.23% excess risk).
11           The analyses of a 9-year data set for the whole population of the Netherlands
12      (population = 14.8 million) provided risk estimates for more specific causes of mortality,
13      including COPD, pneumonia, and subcategories of cardiovascular causes (Hoek et al., 2000,
14      2001;  reanalysis Hoek, 2003).  The effect estimate for total nonaccidental mortality was
15      1.6%(95%CI:  0.9, 2.4) excess risk per 30 ppb increase in 8-h avg O3.  In comparison, the
16      excess risk estimates for pneumonia and COPD were 5.6% (95% CI:  1.8, 9.5) and 0.8% (95%
17      CI:  -2.4, 4.2), respectively. The effect estimates for some of the cardiovascular subcategories,
18      including heart failure (3.8% [95% CI: 0.5, 7.3]) and thrombosis-related disease (6.0% [95% CI:
19      1.1, 10.8]), showed greater risk estimates than that for total mortality. However, these
20      associations were not specific to O3.  For example, most of the pollutants examined, including
21      PM10,  BS, SO2, NO2, CO and NO3  , were associated with pneumonia. Therefore, it is difficult to
22      make a causal inference specific to O3 based on these results.
23           De Leon et al.  (2003) examined the role of contributing respiratory causes in the
24      association between air pollution and nonrespiratory mortality (circulatory and cancer) in
25      New York City during the period of 1985 to 1994.  The  main finding of this study was that the
26      estimated excess mortality risks for PM10 were  higher for nonrespiratory deaths that had
27      contributing respiratory causes compared to deaths without contributing  respiratory causes in the
28      older age (75+ years) group. This pattern was also seen for CO and SO2, but not for O3.
29      Therefore, this study did not suggest a role of contributing respiratory causes in the association
30      between O3 and  nonrespiratory causes of deaths.
        August 2005                               7-93        DRAFT-DO NOT QUOTE OR CITE

-------
 1           In summary, several single-city studies observed positive associations between ambient O3
 2      concentrations and cardiovascular mortality. In addition, a meta-analysis that examined specific
 3      causes of mortality found that the cardiovascular mortality risk estimates were higher than those
 4      for total mortality. The findings regarding the effect size for respiratory mortality have been less
 5      consistent, possibly due to lower statistical power in this subcategory of mortality.
 6
 7      7.4.8   Ozone-Mortality Risk Estimates for Specific Subpopulations
 8           Some studies examined O3-mortality risk estimates in potentially susceptible
 9      subpopulations, such as those with underlying cardiopulmonary disease.  Sunyer et al. (2002)
10      examined the association between air pollution and mortality in a cohort  of patients (467 men
11      and 611 women) with severe asthma in Barcelona, Spain during the period of 1986 to  1995.
12      A case-crossover study design was used to estimate excess odds of mortality adjusting for
13      weather and epidemics in three groups:  (1) those who had only one asthma emergency
14      department visit; (2) those who had more than one asthma emergency department visit; and
15      (3) those who had more than one asthma and COPD emergency department visit.  Those with
16      more than one asthma emergency department visit showed the strongest associations with the
17      examined air pollutants, with NO2 being the most significant predictor, followed by O3. Sunyer
18      et al.  reported a significant association between O3 and all cause deaths for this group during the
19      warm season, with an odds ratio of 2.82 (95% CI:  1.15, 6.87) per 40 ppb increase in 1-h max O3,
20      compared to an odds ratio of 1.03 (95% CI: 0.60, 1.78) for those with only one asthma
21      emergency department visit and 1.08 (95% CI:  0.60,  1.92) for the group with a concomitant
22      diagnosis of COPD.  In another Barcelona study, Saez et al. (1999) examined asthma mortality
23      death among persons aged 2 to 45 years.  Once again, O3 and NO2 were the only air pollutants
24      that were significantly associated with asthma mortality death. While the similarity of the
25      patterns of associations between O3 and NO2 makes it difficult to speculate on the specific causal
26      role of O3, the results of these studies suggest that individuals with  severe asthma may make up a
27      subpopulation that is sensitive to these pollutants.
28           Sunyer and Basagna (2001) also performed an analysis of emergency department visits by
29      a cohort with COPD.  The results from this study suggested that PM10, but not gases were
30      associated with mortality risks for the COPD cohort.  However, a Mexico City study by Tellez-
        August 2005                              7-94        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Rojo et al. (2000) observed a significant association between COPD mortality and O3, as well as
 2      PM10, among patients living outside a medical unit.  For a cumulative 5-day lag, an excess risk of
 3      15.6% (95% CI:  4.0, 28.4) per 40 ppb increase in 1-h max O3 was observed for COPD mortality.
 4           Goldberg et al. (2003) investigated the association between air pollution and daily
 5      mortality with congestive heart failure as the underlying cause of death in patients aged 65 years
 6      or more in Montreal, Quebec, Canada during the period of 1984 to 1993. Analysis was stratified
 7      into two groups, those whose underlying cause of death was congestive heart failure and those
 8      with a diagnosis of congestive heart failure one year before their death.  They found no
 9      association between daily mortality for congestive heart failure and any pollutants. However,
10      they did find associations between daily mortality and coefficient of haze, SO2, and NO2 among
11      those who were classified as having congestive heart failure before death. In the case of O3,
12      positive risk estimates were observed for year-round and warm season data; however, results
13      were not significant. While the 10-year study period for this data was long, the daily mean death
14      counts for the specific subcategory chosen was relatively small (0.7/day for mortality with
15      congestive heart failure as underlying cause of death and 4.0/day for total mortality in patients
16      previously diagnosed with congestive heart failure), limiting the power of the study.
17           In the meta-analysis by Bell et al. (2005), a combined estimate was obtained for the elderly
18      population (age 64 years and older or 65 years and older) using 10 estimates from 9 studies.  The
19      posterior mean estimate for the elderly category (2.92% [95% PI:  1.34, 4.51] per 20 ppb
20      increase in 24-h avg O3) was larger than that from all ages (1.75% [95% PI: 1.10, 2.37] from 41
21      estimates).  The results from this meta-analysis suggest that the elderly population may be
22      particularly susceptible to O3-related mortality.
23           Few studies have examined O3-mortality effects for specific subpopulations.  Among those
24      that investigated the effect of air pollution in populations with underlying cardiopulmonary
25      diseases, associations were not unique to O3 but were shared with other pollutants.  There is
26      suggestive evidence that severe asthmatics may be susceptible to the mortality effects associated
27      with NO2 and O3. In addition, the meta-analysis by Bell et al. (2005) suggests that the elderly
28      population may be more affected by O3.
29
        August 2005                               7-95        DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.4.9   Summary of Acute Ozone Effects on Mortality

 2         •  A substantial body of new data on acute mortality effects of O3 has emerged since the
 3            previous O3 AQCD.  While uncertainties remain in some areas, it can be concluded that
 4            robust associations have been identified between various measures of daily O3
 5            concentrations and increased risk of mortality. Most of the single-pollutant model
 6            estimates from single-city studies fell in the range between 0.5 to 5% excess deaths per
 7            standardized increment. The corresponding summary estimates in large multicity studies
 8            and meta-analyses ranged between 0.5 to 2.5%, with some studies noting heterogeneity
 9            across cities and studies.  These associations could not be readily explained by
10            confounding due to time, weather, nor copollutants, but model specifications likely
11            contributed to some of the observed heterogeneity in risk estimates across studies.
12
13         •  The majority of the available O3-mortality risk estimates were computed using all year
14            data.  The results from the studies that conducted analysis by season suggest that the O3
15            risk estimates were larger in the warm season. Some of the risk estimates in the cool
16            season were negative, possibly reflecting the negative correlation between low-level O3
17            and PM (and other primary pollutants) during that season. Thus, even with adjustment
18            for temporal trends, the O3 risk estimates obtained for year-round data may be
19            misleading. In locations with considerable seasonal variation, season-specific analyses
20            may better elucidate the effect of O3 on mortality.
21
22         •  The majority of the available O3-mortality risk estimates were computed for a single-day
23            lag. Choosing the optimal lag out of several lags examined may bias the single-day risk
24            estimate upward. However, recent findings from the largest U.S. 95 communities study
25            indicated that a strong association between O3 and mortality was observed with a 7-day
26            distributed lag model. Thus, it is possible that the effect of acute O3 exposure on
27            mortality persists over several days. Further research is needed to understand the
28            nature of cumulative effects.
29
30         •  Some studies examined specific subcategories of mortality, but most of these studies had
31            limited statistical power to detect associations due to the small daily mortality counts.
32            A recent meta-analysis indicated that there was a slightly greater risk of cardiovascular
33            mortality compared to total mortality.
34
35         •  Few studies examined the effect of O3 on mortality in subpopulations with underlying
36            cardiopulmonary diseases.  Similar to cause-specific mortality, these population-specific
37            studies had limited statistical power to detect associations. The evidence suggests that
38            individuals with severe asthma may be at increased risk of O3-related mortality; however,
39            similar results were seen with other pollutants.
40

41
        August 2005                               7-96        DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.5   EFFECTS OF CHRONIC OZONE EXPOSURE
 2      7.5.1   Summary of Key Findings on Studies of Health Effects and Chronic
 3             Ozone Exposure from the 1996 O3 AQCD
 4          The 1996 O3 AQCD concluded that there was insufficient evidence from the limited
 5      number of studies to determine whether long-term ambient O3 exposures resulted in chronic
 6      health  effects. However, the aggregate evidence suggested that chronic O3 exposure, along with
 7      other environmental factors, could be responsible for health effects in exposed populations.
 8
 9      7.5.2   Introduction to Morbidity Effects of Chronic Ozone Exposure
10          Several new longitudinal epidemiologic investigations have yielded information on health
11      effects of long-term O3 exposures. Epidemiologic interest in investigating long-term effects has
12      been motivated by several considerations. Animal toxicology studies carried out from the late
13      1980s onward demonstrated that long-term exposures can result in permanent changes in the
14      small airways of the lung, including remodeling of the airway architecture (specifically the distal
15      airways and centriacinar region) and deposition of collagen, as discussed earlier in Chapter 5.
16      These changes result from the damage and repair processes that occur with repeated exposure.
17      Indices of fibrosis also were found to persist after exposure in some of the studies. Collectively,
18      these findings provide a potential pathophysiologic basis for the changes in airway function
19      observed in children in longitudinal studies. Seasonal ambient patterns of exposure may be of
20      greater concern than continuous daily exposure.  In the classical study by Tyler et al. (1988),
21      young  monkeys with seasonal exposure to O3, but not those with daily exposure, experienced
22      increases in total lung  collagen content, chest wall compliance, and inspiratory capacity,
23      suggesting a delay in lung maturation in seasonally-exposed animals.
24          Controlled human exposure studies clearly demonstrated acute inflammation in the lung at
25      ambient exposure levels. Epidemiologic studies could examine whether repeated exposures over
26      multiple episode periods and/or multiple years would lead to persistent inflammation and result
27      in damage to the human lung, especially in the small, terminal bronchiolar regions where
28      vulnerability is greatest. However, the challenges to addressing these issues in epidemiologic
29      studies are formidable, and as a result there exists relatively limited literature in  this area.  Long-
30      term O3 concentrations tend to be correlated with long-term concentrations of other pollutants,
31      making specific attribution difficult.  Subtle pulmonary effects require health outcome measures

        August 2005                             7-97        DRAFT-DO NOT QUOTE OR CITE

-------
 1      that are sensitive, and must usually be directly collected from individual human subjects, rather
 2      than from administrative data bases.  Although these factors make chronic studies difficult and
 3      expensive to conduct, efforts must be made to design studies with adequate power to examine
 4      the hypothesis being tested. Epidemiologic studies have the potential to provide important new
 5      insights on the links between chronic exposure to O3 and the occurrence of human health effects.
 6           This section reviews studies published from 1996 onward in which health effects were
 7      tested in relation to O3 exposures extending from several weeks to many years (Table AX7-6 in
 8      the Chapter 7 Annex). The available literature falls into four general categories:  (1) studies
 9      examining seasonal changes in lung  function as related to O3 exposures in peak season;
10      (2) studies addressing smaller increases in lung function during childhood or decline of lung
11      function beyond childhood in relation to long-term O3 exposures; (3) studies addressing
12      respiratory inflammation in high versus low exposure groups or time periods; and (4) studies
13      addressing longitudinal and cross-sectional associations between long-term O3  exposures and
14      asthma development and prevalence.
15
16      7.5.3   Seasonal Ozone Effects on Lung Function
17           While it has been well-documented in both chamber and field studies that daily, multihour
18      exposures to O3 result in transient declines in lung function, much less is known about the effects
19      of repeated exposures to O3 over extended periods on lung function.  Several new studies
20      reported over the past decade have examined lung function changes over seasonal time periods
21      with differing levels  of O3 exposures (Frischer et al., 1999; Horak et al.,  2002a,b; Ihorst et al.,
22      2004; Kinney and Lippmann, 2000; Kopp et al., 2000). The seasonal effects of O3 are examined
23      first in this section. In the next section is a discussion of effects  over years, as opposed to over
24      seasons, in addition to multiyear analyses of seasonal studies.
25           In a large Austrian study, Frischer et al. (1999) collected repeated lung function
26      measurements in  1,150 school children (mean age 7.8 years) from nine towns that differed in
27      mean O3 levels. Lung function was measured in the spring and fall over a three-year period from
28      1994 to 1996, yielding six measurements per child.  Mean summertime O3 exposure ranged from
29      32.4 to 37.3 ppb during the three summers. Growth-related increases in lung function over the
30      summer season were reduced in relation to seasonal mean O3 levels.  Ozone was associated with
31      a change of -156.6 mL (95% CI: -209.5, -103.7) (central estimate:  -0.029 mL/day/ppb x  90

        August 2005                              7-98        DRAFT-DO NOT QUOTE OR CITE

-------
 1      days/year x 3 years x 20 ppb) in FEVj increase for each 20 ppb increase in mean 24-h avg O3
 2      concentrations over the three summers and -129.6 mL (95% CI: -193.1, -66.1) over the three
 3      winters.  When analyses were restricted to children who had spent the whole summer period in
 4      their community, the changes were greater, with an O3-related -183.6 mL (95% CI:  -278.9,
 5      - 88.3) change in FEVj increase over three summers.  Other pollutants (PM10, SO2, and NO2) had
 6      less consistent associations with changes in lung function. Horak et al. (2002a,b) extended the
 7      study of Frischer et al. (1999) with an additional year of data and stated that seasonal mean O3
 8      was associated with a negative effect on increases in lung function, confirming results from the
 9      previous three-year study. In an editorial, Tager (1999) stated that the Frischer et al. (1999) data
10      provided the first prospective evidence of an association between exposure to ambient air
11      pollution and alterations in lung function in children.  Tager further noted that the prospective
12      study design represented a substantial improvement over data derived from cross-sectional
13      studies and should be emulated. However, Tager also cautioned that it was difficult to attribute
14      the reported effects to O3 alone independently of copollutants.
15          Kopp et al. (2000), in a cohort  of 797 children in Austria and southwestern Germany,
16      reported smaller increases in lung function in children exposed to high (44 to 52 ppb O3) levels
17      of ambient O3.  Children residing in low O3 (24 to 33  ppb) areas experienced a 43 mL increase
18      in FEVj whereas those in high O3 areas only experienced a 16 mL increase during the summer of
19      1994.  Similar results were found in  data from the summer of 1995.  In another Austrian study,
20      Ihorst et al. (2004) examined 2,153 children with a median age of 7.6 years and reported summer
21      pulmonary function results revealing a significantly lower FVC and FEVj increase associated
22      with higher O3 exposures in the summer, but not in the winter.
23          In a pilot study (Kinney and Lippmann, 2000), 72 nonsmoking  adults (mean age 20 years)
24      from the second year class of students at the U. S. Military Academy at West Point, NY provided
25      two lung function measurements, one before and one  after a five-week long summer training
26      program at four locations. There was a greater decline in FEVj among students at the Fort Dix
27      location (78 mL) as compared to  students at the other locations (31 mL).  Ozone levels at Fort
28      Dix averaged 71 ppb (mean of daily  1-h max O3) over the summer training period versus mean
29      values of 55 to 62 ppb at the other three locations. In addition to the  higher mean O3 level, Fort
30      Dix had greater peak O3 values (23 hours >100 ppb) compared to the other locations (1 hour
31      >100 ppb).  Ambient levels of other  pollutants, PM10 and SO2,  were relatively low during the

        August 2005                              7-99        DRAFT-DO NOT QUOTE OR CITE

-------
 1      study and did not vary across the four sites. Though conclusions are limited by the small size of
 2      the study, results are consistent with a seasonal decline in lung function that may be due, in part,
 3      to O3 exposures. An exploratory observation from this study was that there appeared to be a
 4      larger decline for those subjects who completed their post-summer lung function measurements
 5      in the first two weeks after returning from training compared to those measured three to four
 6      weeks after training, which is consistent with some degree of rebound of function following the
 7      summer exposure period.
 8           Collectively, the above studies indicate that seasonal O3 exposure is associated with
 9      smaller increases in lung function in children. The study by Kinney and Lippman (2000)
10      provide limited evidence that seasonal O3 also may affect lung function in adults, though the
11      effect may be somewhat transient.
12
13      7.5.4   Chronic Ozone Exposure Effects on Lung Function and
14             Respiratory Symptoms
15          Lung capacity grows during childhood and adolescence as body size increases, reaches
16      a maximum during the 20s, and then begins to decline steadily and progressively with age.
17      There has long been concern that long-term exposure to air pollution might lead to slower
18      growth in lung capacity, diminished maximally attained capacity, and/or more rapid decline
19      in capacity with age. The concern arises by analogy with cigarette smoking, where it is well-
20      documented that lung function declines more rapidly with age in a dose-dependent manner
21      among adults who smoke cigarettes. Adults who stop smoking return to a normal rate of decline
22      in capacity, although there is no evidence that they regain the capacity previously lost due to
23      smoking (Burchfiel et al., 1995). Because O3 is a strong respiratory irritant and is associated
24      with acute lung function declines as well as inflammation and re-structuring of the respiratory
25      airways, it seems plausible that there might be a negative impact of long-term O3 exposures on
26      lung function.  Exposures that negatively affect increases in lung function during childhood, in
27      particular, might have greater long-term risks. Thus, studies of effects on the rates of increases
28      in lung function in children are especially important.
29           Several studies published over the past decade have examined the relationship between
30      lung function and long-term O3 exposure.  The most extensive and robust study of respiratory
31      effects in  relation to long-term air pollution exposures among children in the U.S. is the

        August 2005                             7-100       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Children's Health Study carried out in 12 communities of southern California starting in 1993
 2      (Avol et al., 2001; Gauderman et al., 2000, 2002, 2004a,b; Peters et al., 1999a,b). The first
 3      cohort included children from the fourth, seventh, and tenth grades. A total of 3,676 students
 4      completed questionnaires regarding their lifetime residential histories, historic and current health
 5      status, residential characteristics, and physical activity. Among those students, 3,293 also
 6      performed pulmonary function tests at the time of enrollment.  Peters et al. (1999a) examined the
 7      relationship between long-term (1986-1990) O3 exposures and self-reports of respiratory
 8      symptoms and asthma in a cross-sectional analysis. For outcomes of current  asthma, bronchitis,
 9      cough, and wheeze, the reported odds ratios were 0.95 (95% CI: 0.70,1.29),  1.14 (95% CI:
10      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
11      increase in 1-h max O3. In another cross-sectional analysis examining the relationship between
12      lung function at baseline and levels of air pollution in the community, there was evidence that
13      annual mean O3 levels were  associated with decreased FVC, FEVb PEF, and FEF25.75 (the latter
14      two being statistically significant) among females but not males (Peters et al., 1999b).
15           Avol et al. (2001) examined 110 children from the first cohort who had moved from the
16      participating communities in southern California to other states to determine whether changes in
17      air quality caused by relocation were  associated with changes in annual increases in lung
18      function. With the exception of FEVj, the O3 effect estimates for all other spirometric
19      parameters were negative, but the associations were not as strong as those observed for PM10.
20           A second cohort of fourth graders (n = 1,678) were recruited in 1996 and followed over
21      four years to examine the association between long-term exposure to air pollution and changes in
22      lung function (Gaunderman  et al., 2002).  In general, smaller increases in various lung function
23      parameters were observed in communities with higher 4-year average O3 levels (for examples,
24      see Figure 7-23). The strongest effect of O3 was on PEF — children from the least-polluted
25      community had a 1.21% (95% CI:  0.36, 2.06) greater increase in PEF compared to those from
26      the most-polluted communities. However, in the 4-year and 8-year longitudinal analysis of the
27      first cohort, Gauderman et al. (2000, 2004) stated that the results provided little evidence that
28      long-term exposure to ambient O3 was associated with significant deficits in the growth rate of
29      lung function in children.
30           In both cohorts of fourth graders, stratified analyses by time spent outdoors indicated a
31      stronger association between long-term O3 exposure  and smaller increases in lung function in

        August 2005                              7-101       DRAFT-DO NOT QUOTE OR CITE

-------
                       m 2-  12.5-
                       i>
                       cm
                      < U-  12.0"
                      •0 .=
                       tf> $  11.5-1
                      .5. <"
                      2 E
                      < «  11.01
                            10.5
   SD
                               25    30    35    40   45   50   55   60   65   70
                                               8-h avg O3 (ppb)
                            12.0
                            10.5-
                            10.0
                            9.5
                               25    30    35    40   45   50   55   60   65   70
                                               8-h avg O3 (ppb)
       Figure 7-23.  Adjusted average annual increases in FEVt and maximal midexpiratory flow
                    (MMEF) versus the mean 8-h avg O3 (10 a.m. to 6 p.m.) concentration over a
                    4-year period in the 12 southern California communities of the Children's
                    Health Study.
             AL = Alpine; AT = Atascadero; LA = Lake Arrowhead; LB = Long Beach; LE = Lake Elsinore;
             LM = Lompoc; LN = Lancaster; ML = Mira Loma; RV = Riverside; SD = San Dimas;
             SM = Santa Maria; UP = Upland
       Source: Gauderman et al. (2002).
1      children who spent more time outdoors, as shown in Table 7-3 (Gauderman et al., 2002).
2      A study by Jedrychowski et al. (2001) found a link between repeated respiratory symptoms and
3      smaller lung function increases. Gauderman et al., therefore, suggested that the observation of
4      reduced increases in lung function with increasing annual average air pollution might be a
5      consequence of repeated acute respiratory events after short-term increases in pollutant levels.
       August 2005
7-102
DRAFT-DO NOT QUOTE OR CITE

-------
          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
FVC

FEVj

MMEF

PEF

Cohort b
Cohort 1
Cohort 2
Cohort 1
Cohort 2
Cohort 1
Cohort 2
Cohort 1
Cohort 2
More Time Outdoors °
% Change (95 %CI)d
-0.02% (-0.57, 0.54)
-0.57% (-1.03, -0.09)
-0.25% (-1.18, 0.68)
-0.68% (-1.36, 0.00)
-0.55% (-2.08, 1.01)
-0.48% (-1.71, 0.78)
-0.77% (-2.03, 0.52)
-1.33% (-2.43, -0.24)
Less Time Outdoors °
% Change (95 %CI)d
-0.04% (-0.45, 0.37)
-0.06% (-0.76, 0.66)
-0.05% (-0.58, 0.49)
-0.29% (-1.02, 0.46)
0.23% (0.89, 1.36)
-0.80% (-2.07, 0.50)
0.25% (-0.65, 1.16)
-0.71% (-1.71, 0.30)
         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.).
 1      The findings of larger deficits in children who spend more time outdoors in the afternoon adds
 2      some support to the possibility.  Results from this study also indicate the importance of reducing
 3      exposure misclassification.  Most long-term epidemiologic studies, including the Children's
 4      Health Study, estimated O3 exposure using centrally-located ambient monitors. Gonzales et al.
 5      (2003) and Kiinzli et al. (1997) evaluated the use of retrospective questionnaires to reconstruct
 6      past time-activity and location pattern information.  Both studies found that questionnaires or
 7      activity diaries might improve the assessment of chronic exposure in epidemiologic studies.
 8           In a study conducted in Austria and Germany, Ihorst et al. (2004) found that there were no
 9      associations between increases in lung function and mean summer O3 levels for FVC and FEVj
10      over a 3.5-year period, in contrast to the significant seasonal effects discussed in the earlier

        August 2005                               7-103        DRAFT-DO NOT QUOTE OR CITE

-------
 1      section. Unlike the reduced increases in lung function parameters over the first two summers
 2      among children in high O3 areas, a greater increase was observed during the third summer and no
 3      difference was observed during the fourth summer. The authors then concluded that medium-
 4      term effects on school children lung function were possibly present but were not detected over a
 5      3- to 5-year period due to partial reversibility.  The study by Frischer et al. (1999) showed results
 6      similar to the Ihorst et al. (2004) study. Although O3 was related to smaller increases in lung
 7      function when three years of data were analyzed collectively, the magnitude and direction of the
 8      effect changed throughout the years. Ozone was associated with a change of -34.0 mL (95% CI:
 9      -58.7, -9.3) in FEVi increase in the first year, compared to +7.3 mL (95% CI:  -20.8, 35.6) in
10      the third year for each 20 ppb increase in mean 24-h avg O3 (Frischer et al., 1999).
11          Calderon-Garciduefias et al. (2003) examined chest X-rays and lung function in
12      174 children from Mexico City and 27 control children from low pollution areas (Tuxpam and
13      Tlaxcala,  Mexico). The Mexico City children exhibited lung hyperinflation (67%), interstitial
14      markings  (49%), and a mild restrictive pattern by spirometry (10%).  In children with increased
15      interstitial markings, FEF75 values were significantly declined (r = 0.42, p < 0.003).
16      No significant abnormalities were observed in the control children. In another study of similar
17      design, the prevalence of respiratory symptoms was higher in Mexico City children compared to
18      children from  the clean coastal town of Manzanillo, Mexico (Calderon-Garciduefias et al., 1999).
19          A study  by Gong et al. (1998b) examined lung function changes in 164 nonsmoking adults
20      (mean age 45 years) from a high O3 community in southern California, recruited from a cohort
21      of 208 who had been tested on two previous occasions. In the earlier analysis by Detels et al.
22      (1987), a  significant decline in lung function was observed from 1977/1978 to 1982/1983. In
23      contrast, Gong et al. observed a slight increase in FVC and FEVj from 1982/1983 to  1986/1987.
24      A consistent decline in FEVj/FVC ratio was observed at all three time points (p  < 0.0001).
25      Among the 45 subjects who further participated in the controlled exposure study (0.40 ppm O3
26      over 2 hours with intermittent exercise), acute changes in lung function were not associated with
27      long-term changes in lung function over a decade.
28          Evidence for a relationship between long-term O3 exposures and decrements in maximally
29      attained lung function was observed in a nationwide cohort of 520 first year students at Yale
30      College in New Haven, CT (Galizia and Kinney 1999; Kinney et al.,  1998). Each student
31      performed one lung function test in the spring of their first year at college. Ozone exposures

        August 2005                              7-104      DRAFT-DO NOT QUOTE OR CITE

-------
 1      were estimated by linking 10-year mean summer season 1-h max O3 levels at the nearest
 2      monitoring station to the residential locations reported each year from birth to the time of
 3      measurement.  Students who had lived four or more years in areas with long-term mean O3 levels
 4      above 80 ppb had significantly lower FEVj (-3.07% [95% CI: -0.22, -5.92]) and FEF25.75
 5      (-8.11% [95% CI:  -2.32, -13.90]) compared to their classmates with lower long-term O3
 6      exposures. Stratification by gender indicated that males had much larger effect estimates than
 7      females, which might reflect higher outdoor activity levels and corresponding higher O3
 8      exposure during childhood.
 9           A similar study of 130 first year college freshmen at the University of California at
10      Berkeley also reported significant effects of O3 on lung function (Kiinzli et al., 1997; Tager
11      et al.,  1998).  Enrollment was limited to students from either the San Francisco or Los Angeles,
12      CA metropolitan areas.  After controlling for city of origin, long-term O3 exposures were
13      associated with declines in FEF25.75 and FEF75  (forced expiratory flow after 75% of FVC has
14      been exhaled).  No effects were seen for PM10 and NO2.  Kiinzli and colleagues noted that
15      significant changes in these mid- and end-expiratory flow measures could be considered early
16      indicators for pathologic changes that might ultimately progress to COPD, as evidenced by
17      animal studies that showed that the primary site of O3 injury in the lung was the centriacinar
18      region (Chapter 5).
19           Sherwin et al. (2000) examined lungs from autopsies of young residents in Miami, FL and
20      Los Angeles, CA for centriacinar region inflammatory diseases. A trend towards greater degrees
21      of centriacinar region alterations was observed in the lungs of Los Angeles residents compared
22      to Miami residents, independent of a smoking  effect. The results suggest that the greater extent
23      and severity  of centriacinar region alterations might be related to the higher O3 levels in Los
24      Angeles. Beyond the challenge of differentiating the lifetime of exposure for subjects in the two
25      cities, various confounding factors also can impact this study. The pathogenesis of centriacinar
26      region alteration is undoubtedly multifactorial with respiratory infection and adverse
27      environmental influences being two major  considerations.  In addition, Sherwin et al. (2000)
28      noted  that the study was limited due to the  relatively small number of cases available.
29      Nonetheless, as observed by Tager (1993),  the use of human postmortem specimens is of interest
30      in future epidemiologic studies.
        August 2005                              7-105       DRAFT-DO NOT QUOTE OR CITE

-------
 1           The results of the southern California Children's Health Study, as well as those from the
 2     European studies, provide little evidence for impacts of long-term O3 exposures on lung function
 3     in children.  However, further study is needed to better address this difficult question. There is
 4     limited evidence that young adults who grew up in high O3 communities may have reduced lung
 5     function compared to those from low O3 communities.
 6
 7     7.5.5  Chronic Ozone Exposure and Respiratory Inflammation
 8           As noted in Chapter 6, human chamber studies have demonstrated that brief (2 to
 9     6.6 hours) exposures to O3 while exercising result in inflammation in the lung, including the
10     alveolar region where gas exchange takes place. This acute exposure effect is potentially
11     important for effects of chronic exposure because repeated inflammation can result in the release
12     of substances from inflammatory cells that can damage the sensitive cells lining the lung. Over
13     extended periods, repeated insults of this kind can lead to permanent damage to and restructuring
14     of the small airways and alveoli. In addition, since inflammation is a fundamental feature of
15     asthma, there is concern that O3-induced inflammation can exacerbate existing asthma or
16     perhaps promote the development of asthma among genetically pre-disposed individuals.
17     Several studies are discussed next, examining different outcomes related to inflammation.
18           In a study by Kinney et al. (1996b), bronchoalveolar lavage fluids were collected in the
19     summer and winter from a group of 19 adult joggers living and working on an island in
20     New York harbor.  The mean 1-h max O3 concentrations for a 3-month  period were 58 ppb
21     (maximum 110) in the summer and 32 ppb (maximum 64) in the winter. PM10 and NO2
22     concentrations were similar across the two seasons. There was little evidence for acute
23     inflammation in bronchoalveolar lavage fluids collected during  the summer as compared to that
24     collected from the same subjects in the winter.  However, there was evidence of enhanced cell
25     damage, as measured by lactate dehydrogenase, in the summer lavage fluids. These results
26     indicate that acute inflammatory responses may diminish with repeated exposures over the
27     course of a summer (which have been demonstrated in multiday chamber exposures,  Chapter 6,
28     Section 6.9) but cell damage may be ongoing.
29           Pollution effects in the nose can be viewed as a potential surrogate measure for effects that
30     may occur in the lungs, though doses to nasal tissues are usually higher for a given pollutant
31     concentration. In Chapter 5, morphological effects of O3 on the upper respiratory tract indicated

       August 2005                             7-106       DRAFT-DO NOT QUOTE OR CITE

-------
 1      quantitative changes in the nasal transitional respiratory epithelium. The persistent nature of the
 2      O3-induced mucous cell metaplasia in rats, as discussed in Chapter 5, suggests that O3 exposure
 3      may have the potential to induce similar long-lasting alterations in the airways of humans.
 4      A series of interesting studies in Mexico City have demonstrated inflammation and genetic
 5      damage to cells in the nasal passages of children chronically exposed to O3 and other air
 6      pollutants (Calderon-Garciduefias et al., 1995, 1997, 1999, 2001, 2003). Nasal lavage samples
 7      and nasal biopsies from children living in Mexico City were compared to those from children
 8      living in a clean coastal town with no detectable air pollutants. In the first study, urban children
 9      (n = 38) from Mexico City were found to have significantly higher polymorphonuclear leukocyte
10      counts and abnormal nasal cytologies compared to nonurban children (n = 28) (Calderon-
11      Garciduefias et al., 1995).  A more recent study of similar design examined nasal abnormalities
12      and serum cytokines in both urban and nonurban children (Calderon-Garciduefias et al., 2003).
13      Twenty-two percent of the 112 Mexico City children showed a grossly abnormal nasal mucosa.
14      No significant abnormalities were observed in the control children. In addition, the Mexico City
15      children had more serum interleukin-10 and interleukin-6, and less serum  interleukin-8 than
16      controls. Twenty-five children with whitish-gray nasal lesions showed a significant association
17      between tumor necrosis factor a and interleukin-8 (r = 0.89, p < 0.0001), which suggested the
18      potential importance of the nose in the production of proinflammatory cytokines.
19           Calderon-Garciduefias et al. (1997)  also observed that cells collected from the lining of the
20      nose had significantly higher amounts of DNA damage in the urban children in Mexico City
21      (n = 129) versus nonurban children (n = 19). Among exposed children, the extent of
22      DNA damage was greater in older children, who had spent more time outdoors and were more
23      engaged in physical activities compared to the younger children. Another study of 86 urban and
24      12 nonurban children reported similar findings, and also noted increased levels of specific DNA
25      mutations (Calderon-Garciduefias et al., 1999). Fortoul et al. (2003) examined DNA strand
26      breaks in nasal epithelial cells from asthmatic and nonasthmatic medical students in Mexico City
27      and noted greater genotoxic damage in  asthmatics.  These results indicate  that asthmatics may
28      have a greater susceptibility for DNA damage, or a decreased ability to repair it, compared to
29      nonasthmatic subjects. However, because of the complex mixture of pollutants present in
30      Mexico City, it is not possible to uniquely attribute these observed changes to O3 concentrations.
31

        August 2005                              7-107       DRAFT-DO NOT QUOTE OR CITE

-------
 1           Another outcome of inflammation was examined in a study by Frischer et al. (2001).
 2     In this cross-sectional study, urinary eosinophil protein was analyzed as a marker of eosinophil
 3     activation in 877 school children living in nine Austrian  communities with varying O3 exposure.
 4     The results indicated that O3 exposure was significantly associated with eosinophil
 5     inflammation.
 6           In the Mexico City studies, specific attribution of these adverse respiratory and genotoxic
 7     effects to O3 is difficult given the complex pollutant mixture present in the ambient air.
 8     In particular, the DNA effects seem more plausibly related to other components of urban air,
 9     such as semi-volatile organic compounds.  However, the inflammatory changes such as
10     increased eosinophil levels observed in the Austrian study would be consistent with known
11     effects of O3.
12
13     7.5.6   Risk of Asthma Development
14           Recent longitudinal cohort studies have reported associations between the onset of asthma
15     and long-term O3 exposures (McConnell et al., 2002; McDonnell et al., 1999).  Significant
16     associations between new cases of asthma among adult males and long-term O3 exposure were
17     observed in a cohort of nonsmoking adults in California  (Greer et al., 1993; McDonnell  et al.,
18     1999). The Adventist Health and Smog (AHSMOG) study cohort of 3,914 (age 27-87 years,
19     36% male) was drawn from nonsmoking, non-Hispanic white California Seventh Day
20     Adventists. Subjects were surveyed in 1977, 1987, and 1992. To be eligible, subjects had to
21     have lived  10 or more years within 5 miles of their current residence in 1977.  Residences from
22     1977  onward were followed and linked in time and space to interpolate concentrations of O3,
23     PM10, SO2, and NO2. New asthma  cases were defined as self-reported doctor-diagnosed asthma
24     at either the 1987 or 1992 follow-up questionnaire among those who had not reported having
25     asthma upon enrollment in 1977. During the 10-year follow-up (1977-1987), the incidence of
26     new asthma was 2.1% for males and 2.2% for females (Greer et al., 1993). A relative risk of
27     3.12 (95% CI: 1.16, 5.85) per 10 ppb increase in annual mean O3 (exposure metric not stated)
28     was observed in males, compared to a relative risk of 0.94 (95% CI:  0.65, 1.34) in females.
29     In the 15-year follow-up study (1977-1992), 3.2% of the eligible males and a slightly greater
30     4.3% of the eligible females developed adult asthma (McDonnell et al., 1999).  For males, the
31     relative risk of developing asthma was 2.27 (95% CI:  1.03, 4.87) per 30 ppb increase in 8-h avg

       August 2005                              7-108       DRAFT-DO NOT QUOTE OR CITE

-------
 1      O3 (9 a.m.-5 p.m.). Once again, there was no evidence of an association between O3 and new-
 2      onset asthma in females (relative risk of 0.85 [95% CI:  0.55,1.29]).  The lack of an association
 3      does not necessarily indicate no effect of O3 on the development of asthma among females.
 4      For example, differences in time-activity patterns in females and males may influence relative
 5      exposures to O3, leading to greater misclassification of exposure in females. The consistency of
 6      the results in the two studies with different follow-up times and indices of O3 exposure provides
 7      supportive evidence that long-term O3 exposure may be associated with asthma incidence in
 8      adult males. However, as the AHSMOG cohort was drawn from a narrow subject definition, the
 9      representativeness of this cohort to the general U.S. population may be limited.
10           A similar study of incident asthma cases in relation to O3 among children was carried out
11      in the Children's Health Study  (McConnell et al., 2002). 3,535 initially nonasthmatic children
12      (ages 9 to 16 years at enrollment) were followed for up to 5 years to identify new-onset asthma
13      cases. Communities were stratified by pollution levels, with six high-O3 communities (mean 1-h
14      max O3 of 75.4 ppb [SD 6.8] over four years) and six low-O3 communities (mean 50.1 ppb
15      [SD 11.0]).  A total of 265 children reported a new diagnosis of asthma during the follow-up
16      period. Asthma risk was not higher for residents of the  six high-O3 communities versus residents
17      of the six low-O3 communities.  However, within the high-O3 communities, asthma risk was
18      3.3 (95% CI: 1.9, 5.8) times greater for children who played three or more sports as compared
19      with children who played no sports.  This association was absent in the low-O3 communities
20      (relative risk of 0.8 [95% CI: 0.4,  1.6]). No associations with asthma were seen for PM10, PM25,
21      NO2, or inorganic acid vapors.  These results suggest effect modification of the impacts of O3 on
22      asthma risk by physical activity. Playing sports may indicate outdoor activity when O3 levels are
23      higher and an increased ventilation rate, which may lead to increased O3 exposure.  It should be
24      noted, however,  that these findings were based on a small number of new asthma cases (n = 29
25      among children who  played three or more sports) and it is not  clear to what extent the key
26      findings were based on a priori hypotheses.  Replication of these findings in other cohorts would
27      lend greater weight to a causal  interpretation.
28           Recent cross-sectional surveys have detected no associations between long-term O3
29      exposures and asthma prevalence,  asthma-related symptoms, or allergy to common aeroallergens
30      in children after controlling for covariates (Charpin et al., 1999; Kuo et al., 2002; Ramadour
31      et al., 2000). It should be noted that O3 levels were quite low in all cases, with a range of 16 to

        August 2005                              7-109       DRAFT-DO NOT QUOTE OR CITE

-------
 1      27 ppb for 8-h max O3. The longitudinal study design, which observes new onset of asthma
 2      prospectively, provides stronger evidence on the question of asthma development.
 3
 4      7.5.7   Respiratory Effects of Chronic Ozone Exposure on Susceptible
 5              Populations
 6           Studies on the effect of long-term O3 exposure on respiratory health has focused mostly on
 7      children, a potentially susceptible population. Ozone exposure was associated with smaller
 8      increases in lung function and respiratory inflammation in children. Other studies have
 9      investigated additional groups of potentially susceptible individuals. McConnell et al. (1999)
10      examined the association between O3 levels and the prevalence of chronic lower respiratory tract
11      symptoms in southern California children with asthma (n = 3,676). In this cross-sectional study,
12      bronchitis, phlegm, and cough were not associated with annual mean O3 concentrations in
13      children with asthma or wheeze.  All other pollutants examined, PM10, PM2 5, NO2, and gaseous
14      acid, was associated with an increase in phlegm, but not cough.
15           In another analysis from the Children's Health Study, McConnell et al. (2003) investigated
16      the relationship between air pollutants and bronchitic symptoms among 475 children with
17      asthma. For a 1 ppb increase in 8-h avg O3 concentrations averaged over 4 years, the between-
18      community odds ratio was 0.99 (95% CI: 0.98, 1.01) compared to the within-community odds
19      ratio of 1.06 (95% CI:  1.00,1.12).  The authors commented that if the larger within-community
20      effect estimates were correct, then other cross-sectional (between-community) studies might
21      have underestimated the true effect of air pollution on bronchitic symptoms in children.  These
22      differences might be attributable to confounding by poorly measured or unmeasured risk factors
23      that vary between communities. In two-pollutant models, the within-community effect estimates
24      for O3 were markedly reduced and in some cases no longer significant (odds ratios not provided).
25      However, given the high correlation between O3 and the other pollutants, a causal role for O3
26      should not be excluded.
27           One recent study examined a susceptible group not examined before. Goss et al. (2004)
28      investigated the effect of O3 on pulmonary exacerbations and lung function in individuals with
29      cystic fibrosis over the age of 6 years (n = 11,484). The study included patients enrolled in the
30      Cystic Fibrosis Foundation National Patient Registry.  The registry contained demographic and
31      clinical data collected annually at accredited centers for cystic fibrosis. In 1999 and 2000, the

        August 2005                             7-110       DRAFT-DO NOT QUOTE OR CITE

-------
 1      annual mean O3 concentration from 616 monitors in the U.S. EPA Aerometric Information
 2      Retrieval System (AIRS) was 51.0 ppb (SD 7.3). Exposure was assessed by linking air pollution
 3      values from AIRS with the patient's home zip code. No clear association was found between
 4      annual mean O3 and lung function parameters.  However, a 40 ppb increase in annual mean 1-h
 5      max O3 was associated with a 46% (95% CI: 13, 87) increase in the odds of two or more
 6      pulmonary exacerbations.  Significant excess odds of pulmonary exacerbations also were
 7      observed with increased annual mean PM10 and PM2 5 concentrations.
 8           In summary, some studies have identified and investigated potentially susceptible
 9      populations. Although effects are not specific to O3 exposure, the results suggest that O3 may
10      contribute to the adverse respiratory health responses observed in individuals with asthma and
11      cystic fibrosis.
12
13      7.5.8  Effects of Chronic Ozone Exposure on Mortality and Cancer Incidence
14           There is inconsistent and inconclusive evidence for a relationship between long-term O3
15      exposure and increased mortality and cancer risk (see Table AX7-7 in the Chapter 7 Annex).
16      In a large prospective cohort study of approximately 500,000 U.S. adults, Pope et al. (2002)
17      examined the effects of long-term exposure to air pollutants on mortality. All cause,
18      cardiopulmonary, lung cancer, and all other cause mortality risk estimates for long-term O3
19      exposure are shown in Figure 7-24. Consistent positive associations were not observed between
20      O3 and mortality. The mortality risk estimates  were larger when analyses were restricted to the
21      summer months (July to September) when O3 levels were generally higher. The O3-mortality
22      risk estimates were positive for all cause and cardiopulmonary mortality, with a marginally
23      significant estimate for cardiopulmonary mortality in the summer months.  A negative,
24      nonsignificant O3 risk estimate was observed for lung cancer mortality. Consistent positive and
25      significant effects of PM2 5 were observed for both lung cancer and cardiopulmonary mortality.
26           Lipfert et al. (2000b, 2003) reported positive effects on all cause mortality for peak O3
27      exposures (95th percentile levels) in the U.S. Veterans Cohort study of approximately 50,000
28      male middle-aged men recruited with a diagnosis of hypertension. The actual analysis involved
29      smaller subcohorts based on exposure and mortality follow-up periods. Four separate exposure
30      periods were associated with three mortality follow-up periods. In a preliminary screening of
31      regression results, Lipfert et al. (2000b) observed a negative association for mean O3 and a

        August 2005                             7-111        DRAFT-DO NOT QUOTE OR CITE

-------
               o
                t/l
                r£
                as
               _
               CD
               oc
1.50
1.40
1.30
1.20
1.10
1.00
0.90
0.80
All Cause Mortality Cardiopulmonary Mortality Lung Cancer Mortality All Other Cause Mortality
I II II
	 T. 	 J 	 ? 	 I 	 Y 	 I 	 I 	 I
2 1 II J (
i i i
> T X I
                        ^  ,&  ,&.

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 7-24.  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).
 1     positive relationship for peak O3; thus, peak O3 was used in subsequent analyses.  The mean of
 2     the peak values ranged from 85 to  140 ppb over the four exposure periods. For concurrent
 3     exposure periods, peak O3 was positively associated with all cause mortality, with a 9.4% (95%
 4     CI: 0.4, 18.4) excess risk per mean 95th percentile O3 less estimated background level (not
 5     stated). When exposure periods preceding death were considered, no association between O3
 6     and mortality was observed (-0.2% [95% CI: -12.5, 12.1]).  In a further analysis, Lipfert et al.
 7     (2003) reported the strongest positive association for concurrent exposure to peak O3 for the
 8     subset with low diastolic blood pressure during the period of 1982-1988. Once again, the O3
 9     effect was diminished when exposure (1982-1988) preceded mortality (1989-1996).
10          A long-term prospective cohort study (AHSMOG) of 6,338 nonsmoking, non-Hispanic
11     white individuals living in California examined the association between air pollutants and lung
12     cancer incidence (Beeson et al.,  1998). Over the follow-up period of 1977 to 1992, 20 females
       August 2005
7-112
DRAFT-DO NOT QUOTE OR CITE

-------
 1      (35% smokers, n = 7) and 16 males (37.5% smokers, n = 6) developed lung cancer.
 2      An association was observed between long-term O3 exposure and increased incidence of lung
 3      cancer in males only. The relative risk for incident lung cancer among males was 3.56 (95% CI:
 4      1.35, 9.42) for an interquartile range increase in hours per year (556 hours/year) when O3 levels
 5      exceeded 100 ppb (Beeson et al., 1998). A stronger association was observed in males who
 6      never smoked (4.48 [95% CI: 1.25,  16.04]) compared to those who smoked in the past (2.15
 7      [95% CI: 0.42, 10.89]) (Beeson et al.,  1998).
 8           A related study by Abbey et al. (1999) examined the effects of long-term O3 exposure on
 9      all cause (n = 1,575), cardiopulmonary (n = 1,029), nonmalignant respiratory (n = 410), and lung
10      cancer (n = 30) mortality in the same AHSMOG study population. A particular strength of this
11      study was the extensive effort devoted to assessing long-term air pollution exposures, including
12      interpolation to residential and work locations from monitoring sites over time and space.
13      No associations with long-term O3 exposure were observed for all cause, cardiopulmonary, and
14      nonmalignant respiratory mortality.  However, effects of O3 on lung cancer mortality confirmed
15      the results of the previous study by Beeson and colleagues. An association between lung cancer
16      mortality and chronic O3 exposure was observed in males only, with a relative risk of 4.19 (95%
17      CI:  1.81, 9.69) (Abbey et al., 1999). The gender-specific O3 effects may be partially attributable
18      to the differences in activity and time spent outdoors. The  questionnaires indicated that males
19      spent approximately twice as much time outdoors and performed more vigorous exercises
20      outdoors, especially during the summer, compared to the females. However, the very small
21      numbers of lung cancer deaths (n = 12 for males) raise concerns with regard to the precision of
22      the effect estimate, as evidenced by the wide confidence intervals.  The lack of an association of
23      chronic O3 exposure with other mortality outcomes, which had much larger samples sizes, also is
24      of concern.  A study by Pereira et al. (2005) provides supportive evidence of an  association
25      between O3 and increase risk of cancer.  The correlation between average air pollution data from
26      1981 to 1990 and cases of larynx and lung cancer in 1997 were assessed in communities of
27      Sao Paulo, Brazil. Of all the pollutants examined (PM10, NO2, NOX, SO2, CO, and O3), O3
28      was best correlated with cases of larynx (r = 0.9929, p = 0.007) and lung cancer (r = 0.7234,
29      p = 0.277).
30          Few studies have examined the effect of chronic O3 exposure on mortality outcomes and
31      incidence of cancer.  Consistent associations with long-term O3 exposure were not observed for

        August 2005                              7-113       DRAFT-DO NOT QUOTE OR CITE

-------
 1      all cause and cardiopulmonary mortality. There is limited evidence supportive of an association
 2      between O3 exposure and lung cancer incidence and mortality; however, the small number of
 3      lung cancer cases, differential effects by gender, and the lack of O3 effects on other mortality
 4      outcomes raise concerns regarding plausibility.
 5
 6      7.5.9   Effects of Ozone on Birth-Related Health Outcomes
 7           In recent years, air pollution epidemiologic studies have examined impacts on birth-related
 8      endpoints including intrauterine, perinatal, postneonatal, and infant deaths; premature births;
 9      intrauterine growth retardation; very low birth weight (weight <1500 grams) and low birth
10      weight (weight <2500 grams); and birth defects.  However, the majority of these studies did not
11      examine the effect of O3. In the limited studies that investigated O3, no associations were
12      observed between O3 and birth outcomes, with the exception of birth defects. The following is a
13      synopsis of the literature on this topic.
14           Pereira et al. (1998) investigated impacts of air pollution on intrauterine mortality in  Sao
15      Paulo, Brazil during 1991 and 1992. NO2, SO2,  CO, O3, and PM10 were examined. Intrauterine
16      mortality was most significantly associated with NO2, and less for SO2 and CO. No association
17      was found for O3 or PM10. Pereira et al. also sampled blood from the umbilical cord of healthy
18      non-smoking pregnant women soon after delivery in 1995 and analyzed for levels of
19      carboxyhemoglobin. They found an association between carboxyhemoglobin and ambient CO
20      after adjusting for passive smoking and weight, suggesting an impact of CO on the fetus.
21      Loomis et al. (1999) examined the association between  air pollutants and infant mortality in
22      Mexico City in the years 1993 to 1995. NO2, SO2, CO,  O3, and PM25 were examined. They
23      reported that the strongest association was found for PM2 5 with a 3- to 5-day cumulative lag.
24      They noted that infant mortality also was associated with NO2  and O3 at a 3- to 5-day lag, but not
25      as consistently as PM2 5. There have been air pollution studies that examined postneonatal
26      mortality (Bobak and Leon, 1992; Bobak and Leon, 1999; Kaiser et al., 2004; Woodruff et al.,
27      1997), but these  studies did not examine O3.
28           Ritz and Yu (1999) investigated the effects of ambient CO on low birth weight among
29      children born in  southern California between 1989 and 1993. They focused on CO because
30      "a biologic mechanism for fetal effects has been proposed for CO, but not for other air
        August 2005                             7-114        DRAFT-DO NOT QUOTE OR CITE

-------
 1      pollutants." They found that exposure to higher levels of ambient CO during the last trimester
 2      was associated with an increased risk for low birth weight.  Using available data, they also
 3      estimated the last-trimester exposures for NO2, O3, and PM10. NO2 and PM10 were positively
 4      associated with CO, but O3 was negatively associated with CO (r = -0.65). Ha et al. (2001;
 5      a GAM analysis using default convergence parameters) examined CO, NO2, SO2, O3, and TSP
 6      for their associations with low birth weight in Seoul, Korea during the years 1996 and 1997.
 7      They estimated first and third trimester exposures by averaging daily air pollution levels during
 8      the corresponding days for the registered births. Ha et al. found that first trimester exposures of
 9      CO, NO2, SO2, and TSP were associated with increased risk of low birth weight, whereas O3 was
10      associated with a decreased risk. The opposite pattern was  observed for third trimester
11      exposures, with an increased risk of low birth weight found only for O3. When exposures from
12      both trimesters were examined simultaneously, the associations of first trimester exposures of
13      CO, NO2, SO2, and TSP with increased risk of low birth weight remained; however, the
14      association between third trimester O3 exposure and low birth weight was diminished. Based  on
15      these results, Ha et al. concluded that exposures to CO, NO2, SO2, and TSP in the first trimester
16      were risk factors for low birth weight.  Note that neither of these studies examined the air
17      pollution effect by season.  Other studies that examined the associations between air pollution
18      and low birth weight (Bobak, 2000; Bobak and Leon, 1999; Lin et al., 2001; Maisonet et al.,
19      2001; Wang et al., 1997) did not examine O3 data, and found associations between low birth
20      weight and either one or more of CO, SO2, NO2 and PM indices.  Collectively, these results do
21      not indicate strong evidence of the role of O3 in low birth weight.
22           Two studies by Dejmek et al. (1999; 2000) examined  the relationship between ambient air
23      pollution and risk of intrauterine growth retardation in a highly polluted area of Northern
24      Bohemia (Teplice District).  Both studies, however, focused on PM indices and did not analyze
25      gaseous pollutants.
26           A few studies have examined the association between air pollution and premature births
27      (Bobak, 2000; Ritz et al., 2000; Xu et al., 1995), but only Ritz et al. (2000) included O3 in their
28      analysis.  Ritz et al. evaluated the effect of air pollution exposure during pregnancy on the
29      occurrence of preterm birth in a cohort of 97,518 neonates born in southern California.  CO, NO2
30      SO2, O3, and PM10 data measured at 17 air quality monitoring stations were used to estimate the
        August 2005                              7-115       DRAFT-DO NOT QUOTE OR CITE

-------
 1      average exposures for the first month and the last 6 weeks of pregnancy.  They found
 2      associations between PM10 levels averaged for the last 6 weeks of pregnancy as well as PM10
 3      levels averaged over the first month of pregnancy.  Similar but weaker associations were found
 4      for CO. No association was found for O3.  The reported correlation matrix indicated that O3 was
 5      negatively correlated with CO (r = -0.45)  and only weakly correlated with PM10 (r = 0.2).  The
 6      results from Beijing, China (Xu et al., 1995) and the Czech Republic (Bobak, 2000) suggested
 7      that SO2 and TSP were associated with preterm births. Considering that O3 tends to be
 8      negatively correlated with winter-type pollutants, O3 is unlikely to be an important risk factor for
 9      preterm births.
10           Ritz et al. (2002) evaluated the effect of air pollution on the occurrence of birth defects in
11      neonates and fetuses delivered in southern California from 1987 to 1993 as ascertained by the
12      California Birth Defects Monitoring Program. They averaged air pollution (CO, O3, PM10, and
13      NO2) levels measured at the assigned ambient station over the first, second, and third month of
14      gestation.  Conventional, polytomous, and hierarchical logistic regressions were used to estimate
15      odds ratios for subgroups of cardiac and orofacial defects.  Concentration-response relationships
16      were observed for second month CO exposure on ventricular septal defects, and second month
17      O3 exposure on aortic artery and valve defects, pulmonary artery and valve anomalies,  and
18      conotruncal defects.  The odds ratios observed for these outcomes were similar and quite large
19      (e.g., the odds ratios comparing the highest [monthly 24-h avg mean  34.9 ppb] to lowest
20      [monthly mean 6.4 ppb] O3 quartiles ranged from 2.0 to 2.7), and were not sensitive in
21      multipollutant models.  Ritz et al. reported that they did not observe consistently increased risks
22      and concentration-response patterns for NO2 and PM10 after controlling for the effects of CO and
23      O3.  Results from this study were in contrast to those with other birth-related outcomes in that
24      both CO and O3, presumably negatively  correlated pollutants, were associated with birth defects.
25      Further, O3 showed associations with more birth defect outcomes compared to CO. It should be
26      noted, however, that the concentration-response relationships were quite specific to exposures
27      during the second month. Associations with third month exposures were often negative (though
28      not significantly).  Since both CO and O3 show strong seasonal peaks, it is possible that seasonal
29      confounding could have played some role  in these associations. This is the only study  to date
30      that examined the relationship between air pollution and birth defects.
        August 2005                              7-116        DRAFT-DO NOT QUOTE OR CITE

-------
 1           In summary, O3 was not an important predictor of birth-related outcomes including

 2      intrauterine and infant mortality, premature births, and low birth weight. Birth-related outcomes

 3      generally appear to be associated with air pollutants that tend to peak in the winter and are

 4      possibly traffic-related, most notably CO.  The strong results for CO are consistent with its

 5      ability to cross the placental barrier and the high affinity that hemoglobin in fetal blood has for

 6      binding with it. However, most of these studies did not analyze the data by season, and therefore

 7      seasonal confounding may have influenced the reported associations. One study reported

 8      associations between exposures to O3 in the second month of pregnancy and birth defects.  Since

 9      the O3 effect estimates were relatively large (odds ratios >2.0 at the highest O3 quartile), the

10      potential role of O3 on birth defects should be further investigated.
11

12      7.5.10  Summary of Chronic Ozone Exposure Effects on Morbidity
13              and Mortality

14         • In the past decade, important new longitudinal studies have examined the effect of chronic
15           O3 exposure on respiratory health outcomes, including seasonal declines in lung function,
16           increases in inflammation, and development of asthma in children and adults.  Seasonal
17           O3 effects on lung function have been reported in several studies; however, it remains
18           uncertain to what extent these changes are transient.  There is suggestive evidence that
19           chronic exposure to O3 also may be associated with airway inflammation. In contrast to
20           the supportive evidence from chronic animal studies, epidemiologic studies of new asthma
21           development and longer-term lung function declines remain inconclusive at present.
22
23         • Few studies have investigated the effect of long-term O3 exposure on mortality and cancer
24           incidence.  Uncertainties regarding the exposure period of relevance, differential effects
25           by gender, and inconsistencies across outcomes raise concerns regarding plausibility.
26           There is currently little evidence for a relationship between chronic O3 exposure and
27           increased risk of mortality.
28
29         • A limited number of studies have examined the relationship between air pollution and
30           birth-related health outcomes, including mortality, premature births, low birth weights,
31           and birth defects. The most consistent associations with various birth outcomes were
32           observed for CO. One study reported a large effect of O3 on cardiac defects. The
33           potential role  of O3 on birth defects needs to be further examined.
34

35
        August 2005                              7-117        DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.6   INTERPRETIVE ASSESSMENT OF THE EVIDENCE IN
 2            EPIDEMIOLOGIC STUDIES OF OZONE HEALTH EFFECTS
 3      7.6.1   Introduction
 4          In the 1996 O3 AQCD, the epidemiologic section focused primarily on individual-level
 5      camp and exercise studies, and studies of hospital admissions and emergency room visits.  The
 6      field studies indicated concentration-response relationships of O3 exposure from the ambient air
 7      with declines in pulmonary function, increases in respiratory symptoms, and exacerbation of
 8      asthma, especially in children. Numerous new studies provide additional evidence for
 9      evaluating associations between O3 exposure and the above respiratory health outcomes. The
10      1996 O3 AQCD review of aggregate population time-series studies indicated an association
11      between ambient O3 concentrations and increased hospitalizations. Limited studies examined
12      the O3-mortality relationship.  The current O3 AQCD further presents results from time-series
13      studies that have addressed previously unresolved issues regarding potential linkages between
14      ambient O3 concentrations and health outcomes, particularly mortality. Daily time-series studies
15      minimize  confounding by population characteristics (e.g., cigarette smoking, diet, occupation)
16      by following the same population from day to day. However, confounders operating over
17      shorter time scales can affect O3 risk estimates in these studies.
18          In this section, the issues and attendant uncertainties that affect the interpretation of O3
19      health effects will be discussed. The use of various indices to represent O3 exposure in
20      epidemiologic studies is discussed first. Also, of interest is the issue of confounding by temporal
21      factors, meteorological factors, and copollutants.  The shape of the concentration-response
22      function and heterogeneity of O3 effects also will be discussed briefly.  All of these topics are of
23      much importance for characterizing and interpreting ambient O3-health effects associations.
24
25      7.6.2   Ozone Exposure Indices
26          Three O3 indices were used most often to indicate daily O3 exposure: maximum 1-h
27      average (1-h max); maximum 8-h average (8-h max); and 24-h average (24-h avg). The 8-h max
28      O3 is a frequently used index in newer epidemiologic studies, as it best reflects the current U.S.
29      EPA NAAQS. The O3 exposure indices are highly correlated as indicated in several studies.
30      In the 21 European multicities acute mortality study (Gryparis et al., 2004), 1-h max O3 was
31      found to be highly correlated with 8-h max O3, with a median correlation coefficient of 0.98

        August 2005                              7-118       DRAFT-DO NOT QUOTE OR CITE

-------
 1      (range 0.91-0.99).  Among single-city studies, the 1-h max O3 and 8-h max O3 also were found
 2      to have correlation coefficients ranging from 0.91 to 0.99 in various cities such as Atlanta, GA
 3      (Tolbert et al., 2000; White et al., 1994); southern New England (Gent et al., 2003); Ontario,
 4      Canada (Burnett et al., 1994); and Mexico City (Loomis  et al., 1996; Romieu et al., 1995).
 5      In addition, 1-h max O3 was highly correlated with 24-h avg O3, as observed in the Mexico City
 6      study by Loomis et al. (1996) (r = 0.77) and in the Ontario, Canada study by Burnett et al. (1994)
 7      (r = 0.87).
 8           All studies discussed in Sections 7.2 to 7.5 were examined for presentation of the three O3
 9      exposure indices.  Several presented the concentration data and correlations among 1-h max, 8-h
10      max, and 24-h avg O3 ambient measures.  Some presented the associated risk estimates of
11      comparable analyses for the three exposure indices.  No papers provided a statistical analysis
12      comparing results  from the different indices. Summary of the available data is provided below
13      starting with two multicity mortality studies.
14           In the large U.S. 95 communities study by Bell et al. (2004), increases in O3-associated
15      daily mortality were estimated using all three O3 indices. The increments used in this document
16      to standardize expressions of excess risks are 40 ppb for  1-h max O3, 30 ppb for 8-h max O3, and
17      20 ppb for 24-h avg O3, as discussed in Section 7.1.3.2. For these increments, the effect
18      estimates calculated by Bell et al. (2004) using all available data were 1.34% (95% PI: 0.84,
19      1.85), 1.28% (95% PI:  0.88, 1.73), and 1.04% (95% PI:  0.54, 1.55) excess risk in mortality for
20      1-h max O3, 8-h max O3, and 24-avg O3, respectively. A statistical test examining differences
21      among these risk estimates indicated that there were no significant differences by exposure
22      index. In the European study of 21 cities (of the 23 cities, two did not have 8-h max O3 data),
23      the O3-mortality effect estimate for the summer season was slightly smaller for 8-h  max O3,
24      1.82% (95% CI: 0.99, 3.06) excess risk, compared to 1-h max O3, 2.59% (95% CI: 1.32,4.10)
25      excess risk; however, the two risk estimates were not significantly different (Gryparis et al.,
26      2004).
27           Several single-city mortality studies examined multiple O3 exposure indices (Anderson
28      et al., 1996; Dab et al., 1996; Sunyer et al., 2002; Zmirou et al., 1996; Borja-Aburto et al., 1997).
29      These studies did not differentiate risk estimates by exposure index as the results were
30      considered similar. Hospital admission studies also  provided limited data for O3 index
31      comparisons. Schouten et al. (1996) found similar O3 effects on total respiratory hospitalizations

        August 2005                              7-119        DRAFT-DO NOT QUOTE OR CITE

-------
 1      from 8-h max O3 and 1-h max O3 in the summer.  Both indices resulted in a 4.0% excess risk per
 2      standardized increment.  For emergency department visits, the examples of Delfino et al.
 3      (1998b) and Weisel et al. (2002) indicated no difference in effect estimate when using various O3
 4      indices. Tolbert et al. (2000) noted an increase in emergency room visits of 4.0% per standard
 5      deviation increase (approximately 20 ppb) for both 1-h max O3 and 8-h max O3 as being
 6      expected since the correlation between the indices was 0.99.
 7          Peak flow asthma panel  studies generally used only one index; thus, there were limited
 8      data available for comparison. One respiratory symptom study (Gent et al., 2003) examined
 9      both 1-h max O3 and 8-h max  O3 but noted no differences in the results. Only one FEVj panel
10      study examined more than one O3 exposure index.  Chen et al. (1999) examined 1-h max O3 and
11      24-h avg O3 and reported a decrement in FEVj of -25.6 mL (95% CI: -49.1, -2.1) for  1-h max
12      O3 and -13.6 mL (95% CI: -33.2, 6.0) for 24-h avg O3 in children at a 1-day lag. For 2- and
13      7-day lags, smaller differences were observed between the two indices.  Despite the apparent
14      differences, the effect estimates calculated using  1-h max O3 and 24-h avg O3 concentrations
15      were not found to be significantly different for any of the lags examined.
16          Limited information is available to reach conclusions for comparison of the three indices
17      1-h max O3, 8-h max O3, and 24-h avg O3. Studies conducted in various cities have observed
18      very high correlations among the daily O3 indices.  For the same distributional increment, the
19      excess health risk estimates and significance of associations were generally comparable for the
20      three O3 indices across all outcomes. The high correlation among the indices presents a
21      challenge in distinguishing the most appropriate measure for epidemiologic studies.  Exploratory
22      analyses using various O3 exposure indices are valuable in understanding relationships.
23      However, to address the issue of multiple hypothesis testing, hypotheses that are confirmatory
24      and exploratory should be decided a priori and reported accordingly.
25
26      7.6.3   Confounding by Temporal Trends and Meteorologic Effects  in
27             Time-Series Studies
28          The challenge of analyzing acute O3 effects in time-series  studies is to avoid bias due to
29      confounding by daily to seasonal temporal factors.  On a seasonal scale, the analysis must
30      remove the influence of the strong seasonal cycles that usually exist in both health outcomes and
31      O3. On a daily scale, weather  factors and other air pollutants also may confound the association

        August 2005                             7-120      DRAFT-DO NOT QUOTE OR CITE

-------
 1      of interest.  This section discusses the interpretation of effect estimates after adjusting for
 2      temporal trends and meteorologic effects.
 3
 4      7.6.3.1  Assessment of Ozone Effects after Adjusting for Temporal Trends and
 5              Meteorologic Effects
 6           The relationship between O3 and health outcomes are significantly affected by temporal
 7      trends and meteorological factors, namely temperature.  Analyses of the association between
 8      health outcomes and O3 concentrations using raw data, therefore, can be misleading. In Diaz
 9      et al. (1999), a U-shaped relationship was observed between mortality and O3 concentrations, in
10      which the negative portion of the slope was likely due to the opposing seasonal cycles in
11      mortality (high in winter) and temperature (low in winter). Goldberg and Burnett (2003) report a
12      positive slope for the temperature-mortality relationship being fitted most tightly in the mild
13      temperature range where mortality effects of temperature are not expected.  It is possible that
14      temperature has mortality effects in the mild temperature range, however because daily
15      fluctuations of air pollution, especially  O3, are strongly influenced by weather conditions,
16      ascribing the association between temperature and mortality entirely to effects of temperature
17      may underestimate the effects of air pollution.
18           Sensitivity analyses specifically for O3 effects were performed in the U.S. 95 communities
19      data by Bell et al. (2004).  They found that varying the degrees of freedom from 7 to 21  per year
20      did not significantly affect the O3-mortality estimates, with effect estimates ranging from 0.82 to
21      1.08% excess risk per 20 ppb increase in 24-h avg O3 during the previous week. Using more
22      degrees of freedom in temporal trend fitting (i.e., controlling shorter temporal fluctuations)
23      means ascribing more details of daily health outcomes to unmeasured potential confounders and
24      possibly taking away real weather and air pollution effects.  However, results from this large
25      multicity study indicated that O3 effects were robust to aggressive smoothing of temporal trends.
26      In a related analysis of 19 U.S. cities by Huang et al. (2005), sensitivity of summertime O3 risk
27      estimates to varying degrees of freedom (4 to 16 per year) for temporal trend adjustment was
28      examined.   The extent of change in the risk estimates, while varied from city to city (graphically
29      presented), was not substantial. Huang et al. concluded that the risk estimates were robust to the
30      adjustment for long-term trends.
31           Ito et al. (2005) examined sensitivity of O3-mortality risk estimates to the extent of
32      temporal trend adjustment and to alternative weather model specifications using data from seven

        August 2005                              7-121        DRAFT-DO NOT QUOTE OR CITE

-------
 1      U.S. cities (Cook County, IL; Detroit, MI; Houston, TX; Minneapolis, MN; New York City;
 2      Philadelphia, PA; and St. Louis, MO). They found that varying the degrees of freedom from 4 to
 3      26 per year did not substantially or systematically affect the O3-mortality estimates, except for
 4      Cook County where the percent excess O3-mortality risk estimates were considerably reduced
 5      when the temporal adjustment term with 26 degrees of freedom was applied.  Ito et al. noted that
 6      the O3 risk estimates were generally more sensitive to alternative weather models than to the
 7      degrees of freedom for temporal adjustment.
 8           Schwartz (2005) examined the sensitivity of the O3-mortality relationship to methods used
 9      to control for temperature. Initially, temperature lagged 0 and 1 day was controlled using
10      nonlinear regression splines with 3 degrees of freedom each. In a comparison analysis, control
11      days were restricted to a subset that was matched on temperature. The effect estimates for all
12      year data using nonlinear regression splines (0.8% [95% CI:  0.1, 1.4] excess risk per 40 ppb
13      increase in 1-h max O3) and temperature matched controls (0.9% [95% CI:  0.04, 1.8] excess
14      risk) were not significantly different.  Results were similar when restricting analysis to warm
15      season only data.
16           Temporal cycles in daily hospital admissions or emergency department visits are often
17      considerably more episodic and variable than is usually the case for daily mortality.  As a result,
18      smoothing functions that have been developed and tuned for analyses of daily mortality data
19      may not work as well at removing cyclic patterns from morbidity counts. Two methods are
20      commonly used to adjust for temporal trends. The pre-adjustment method involves applying the
21      adjustment to both outcome and air pollution variables prior to the regression analysis. The
22      co-adjustment method involves applying the adjustment as part of the regression analysis, by
23      fitting a function of time while simultaneously fitting the regression effect of air pollution and
24      weather factors. As shown in a hospital admissions study by Burnett et  al. (2001;  used Poisson
25      GAM with default convergence criteria), the co-adjustment approach may lead to biased air
26      pollution effect estimates in cases where both outcome and pollution variables exhibit strong
27      seasonal cycles. Using year-round data, pre-adjustment followed by regression analysis yielded
28      a 14% (95% CI: 5, 24) increase in admissions per 40 ppb increase in 1-h max O3 with a
29      multiday lag of 0 to 4 days. The co-adjustment method resulted in a 7% (95% CI: 3, 11)
30      decrease in admissions. When the authors limited the analysis to the warm season (May-
31      August), both methods yielded similar results (32% [95% CI: 21, 44] versus 30% [95% CI:  17,

        August 2005                              7-122       DRAFT-DO NOT QUOTE OR CITE

-------
 1      45] increase for co-adjustment and pre-adjustment, respectively) implying that stratification by
 2      season can remove a significant amount of the confounding seasonality (which also may include
 3      seasonally-varying population behavior and ventilation conditions). This finding may be
 4      important to consider in reviewing the acute O3 mortality and morbidity literature since the vast
 5      majority of studies published over the past decade have used the co-adjustment method.
 6      However, the use of pre-adjustment versus co-adjustment in time-series studies is an unresolved
 7      issue. More empirical research in different locales is needed to evaluate the merits of these two
 8      methods as far as O3 is concerned, and to determine what endpoints may be affected.
 9           More sensitivity analysis of O3 effect estimates to the extent of adjustment for temporal
10      trends and meteorological factors is needed, but perhaps it is equally as important to evaluate the
11      epidemiologic  adequacy of a given adjustment.  For  example, do the fitted mortality series
12      sufficiently depict influenza epidemics? Or, when larger degrees of freedom (e.g., 12 degrees of
13      freedom per year) are used, what "unmeasured" confounders, other than weather and pollution,
14      are the investigators trying to adjust? Even in PM studies that conducted  sensitivity analyses,
15      investigators rarely stated assumptions  clearly and not enough discussions were provided as to
16      potential reasons for the sensitivity of results.
17           Given their relationship to health  outcomes and O3 exposure, adjusting for temporal trends
18      and meteorologic factors is critical to obtain meaningful O3 effect estimates.  While the
19      prevailing analytical approaches fit the  data flexibly, the estimated effects of meteorologic
20      variables and their impact on the adjusted O3 effects  are not adequately discussed.  More work is
21      needed in this area to reduce the uncertainty involved in the epidemiologic interpretation of O3
22      effect estimates.
23
24      7.6.3.2  Importance of Season-Specific Estimates of Ozone Health Effects
25           Analysis  of O3 health effects is further complicated as relationships  of O3 with other
26      pollutants and with temperature appear to change across seasons.  Moolgavkar et al. (1995)
27      examined the relationship between daily mortality and air pollution by season in Philadelphia,
28      PA for the period of 1973 to 1988. During the summer, there was a positive relationship
29      between O3 and TSP, as well as O3 and SO2.  In  contrast, the relationship  of O3 with TSP and
30      SO2 inversed during the winter. Ozone showed  positive associations only in the summer when
31      the mean O3 concentration was the highest.  The effect of O3 on mortality was negative (though

        August 2005                               7-123        DRAFT-DO NOT QUOTE OR CITE

-------
 1      not significantly) in the winter when the mean O3 concentration was low. In the summer
 2      multipollutant model, O3 was the only pollutant that remained significant.  Similar results were
 3      found in another Philadelphia study by Moolgavkar and Luebeck (1996). Both studies did not
 4      analyze year-round data, therefore the relationship between the excess risk estimates for all year
 5      and each season could not be compared. The results from these studies,  however, suggest that
 6      year-round analyses may mask the positive (or negative) associations that may exist in particular
 7      seasons.
 8           Ito et al. (2005) examined O3-mortality associations in seven U.S. cities, but also described
 9      the relationship between O3 and PM for summer months (June-August) and winter months
10      (December-February) in these cities (see Figure 7-25). The O3-PM relationships were positive in
11      the summer and negative in the winter in all of these cities, except in Houston, where the O3-PM
12      association was not clearly positive in the warmer months but positive in colder months.
13      Ito et al. found that O3-mortality associations were mostly weaker, null, or even negative in the
14      winter compared to the summer in most of these cities. Once again, the exception was Houston
15      where the cold season O3-mortality association was positive and larger than those for year-round
16      or warmer months. Findings from this study suggest the influence of seasonal O3-PM
17      relationships on O3-mortality associations.
18           In the analyses of the U.S. 90 cities data (of which 80 cities had O3 data  available) by
19      Samet et al. (2000; reanalysis Dominici et al., 2003), the focus of the study was PM10, but O3 and
20      other gaseous pollutants also were analyzed in  single- and multiple-pollutant models. In the
21      reanalysis  (Dominici et al., 2003), O3 was associated with an excess risk of mortality in analyses
22      of all available data (0.4% [95% PI: 0.1, 0.7]) and summer only data (1.0% [95% PI: 0.5, 1.6];
23      however, a negative association was observed for the winter only analysis (-1.1%  [95% PI:
24      -2.2, 0.1]). A twofold greater effect was estimated using summer data compared to all available
25      data.  It should be noted that the analyses by Samet et al. and Dominici et al. used a weather
26      model specification that is more detailed than other studies in that it had multiple terms for
27      temperature and dewpoint (these two variables are generally highly correlated).  Thus, it is
28      possible that the high concurvity of O3 with these weather covariates may have produced these
29      conflicting results. Another possibility is that the apparent negative relationship between O3 and
30      mortality in the winter may have been due to confounding by PM. In the larger U.S. 95
31      communities study by Bell et al. (2004), the all available data and summer only analyses also

        August 2005                              7-124       DRAFT-DO NOT QUOTE OR CITE

-------
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\ 
-------
 1      observed in the U.S. 90 cities study. However, there are several differences between the two
 2      studies that might account for these results. First, the U.S. 95 communities study nearly doubled
 3      the study period by extending the analysis by six additional years (1987 to 2000 versus 1987 to
 4      1994) and included 15 additional cities to the original 80.  Also note that the warm seasons are
 5      characterized differently in the two studies. The U.S. 90 cities study defined summer as a three-
 6      month period of June through August, while the 95 communities study defined warm season as a
 7      seven-month period of April through October.  In addition, the results presented in the U.S. 90
 8      cities study were from a single-day lag model (lag 1-day) while the estimates from the 95
 9      communities study were calculated using a constrained distributed 7-day lag model.  The
10      difference in seasonal O3 effects observed in the two related studies might be attributable to
11      some of these factors.
12           Many studies reported larger excess mortality risks in the warm (or summer) season than in
13      the cool (or winter) season (see Figure 7-19 in Section 7.4.5). These studies showed cool season
14      risk estimates that were either smaller compared to warm season estimates or slightly negative.
15      Of the studies that analyzed data by season, only one study in Pittsburgh, PA (Chock et al.,
16      2000) showed negative risk estimates in the summer.  The studies that observed larger, positive
17      associations between O3 and mortality in warm seasons are consistent with the expectation that
18      O3, if harmful, should have a stronger  association with health outcomes in the summer when
19      exposure to O3 is higher.  However, the negative O3-mortality associations seen in the winter
20      suggest that further examination of this issue is required. Specifically, if the O3 level in the
21      winter is shown to be negatively  associated with factors (e.g., PM) that are positively associated
22      with mortality, then these potentially spurious negative O3-mortality associations can be
23      explained.   Several examples of this phenomenon also exist in morbidity studies investigating
24      the effect of O3 on daily hospital  admissions and emergency department visits (Anderson et al.,
25      1998; Burnett et al., 2001; Prescott et al.,  1998; Thompson et al., 2001).
26           Unlike the time-series studies examining outcomes of mortality, hospital admissions, and
27      emergency  department visits, most acute field studies did not perform year-round analyses.
28      These acute field studies that examined the relationship between O3 and lung function,
29      respiratory  symptoms, and inflammation focused primarily on the O3 effect during the warm
30      season when O3 levels were expected to be high and subjects spent more time outdoors and were
31      physically active.

        August 2005                              7-126       DRAFT-DO NOT QUOTE OR CITE

-------
 1           There are seasonal (e.g., air conditioning use) or seasonally-modified (e.g., time spent
 2      outdoors, air exchange rates) factors that affect the relationship between ambient concentrations
 3      and personal exposures to O3, as discussed in Section 3.9.  The influence of combinations of
 4      these factors across seasons on air pollution health effects can become quite complex. For
 5      example, longer time spent outdoors in the summer may increase personal exposure to O3 for
 6      some segment of the population, but the increased use of air conditioners may reduce exposures
 7      to ambient O3 for those who spend much of their time indoors. In the meta-analysis by Levy
 8      et al. (2005), the combined risk estimate from the warm season was greater (3.38% [95% CI:
 9      2.27, 4.42] per 40 ppb increase in 1-h max O3) compared to the estimate from all year data
10      (1.64% [95% CI: 1.25, 2.03]).  However, further analysis suggested that the O3-mortality risk
11      estimates were smaller in cities with high air conditioning prevalence.  These seasonal factors
12      that influence the relationship between ambient concentrations and personal exposures make the
13      interpretation of the concentration-response relationships obtained from analyses of year-round
14      data less straightforward.
15           In some cities, O3 is  only monitored during the warm season.  For example, 34% of the
16      communities in the U.S. 95 communities study only collected O3 data during the warm season
17      (Bell et al., 2004). The cities with larger populations and/or higher O3 concentrations generally
18      collected year-round data. There is some concern that differential data availability may
19      contribute to the seasonal differences in O3 health effects observed.
20           The potential influence of season on O3 effect estimates was examined using summary
21      density curves.  The O3 effect observed in all year data was compared to effects from warm
22      season and cool  season only data (Figures 7-26 and 7-27).  Summary probability density curves
23      were calculated to review the effect estimates from the various studies (see Annex Section
24      AX7-2 for further explanation of summary density curves). The summary density curves shown
25      in Figures 7-26 and 7-27 were smoothed by multiplying a constant to the standard error of each
26      effect estimate in the calculation of the individual distribution functions. Since the normal
27      distribution is unimodal, this constant will oversmooth when the density is multimodal.
28      In Figure 7-26, the summary density curves of O3-associated all cause  (nonaccidental) mortality
29      are presented (see Figure 7-19 in Section 7.4.5 for the effect estimates). The summary density
30      curves are calculated using results from 14 studies that reported at least two of the three
31      estimates. This figure indicates that 75% of the area under the density curve has a value greater

        August 2005                              7-127       DRAFT-DO NOT QUOTE OR CITE

-------
           =
           "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 7-26.  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 7-19 in Section 7.4.5 for the effect estimates).
 1     than zero for all year data compared to 78% for warm season data and 62% for cool season data.
 2     Therefore, both all year and warm season data generally indicate a positive O3 effect on
 3     mortality. The mean effect estimates for all year data and warm season only data are
 4     1.2%(SD 2.8)  and  1.3% (SD 2.6) excess risk in mortality per 40 ppb increase in 1-h max O3,
 5     respectively. A slightly larger mode of effects is observed for warm season data (0.9% excess
 6     risk) compared to all year data (0.7%). The cool season only data indicate that there is no excess
 7     risk (mean 0.1% [SD 3.1]) associated with O3 concentrations.
 8           Similar observations are made when examining the O3 effect on total respiratory hospital
 9     admissions (Figure 7-27).  Six studies provided season-specific estimates as well as all year
10     results (see Figure 7-10 in Section 7.3.3 for the effect estimates).  Once again, a large percent of
11     the area under the summary density curve is greater than zero when using all year and warm
12     season data, 92% and 84%, respectively, compared to cool season data, 49%. The mean O3
       August 2005
7-128
DRAFT-DO NOT QUOTE OR CITE

-------
                                     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 7-27.  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).
 1     effect estimates for warm season data only, 6.3% (SD 9.1) excess risk per 40 ppb increase in 1-h
 2     max O3, and all year analyses, 6.5% (SD 6.4) excess risk, are similar.  A larger mode of effects is
 3     observed for warm season data (3.9% excess risk) compared to all year data (1.8% excess risk).
 4     A small O3 effect (0.8% [SD 6.1] excess risk) is observed when using cool season data only.
 5          Integrating seasonal influences across the various health outcomes supports the view that
 6     O3 effects are different in the cool and warm seasons, with greater effects observed during the
 7     warm season.  As this relates to potentially higher O3 exposures during the warm season, the
 8     larger effects are consistent with causal  inference. Therefore, these results indicate that the focus
 9     should be on warm season data to derive quantitative relationships for the effect of O3 on health
10     outcomes.  This conclusion is supported by epidemiologic researchers who mainly examine
11     warm season as an a priori hypothesis. However, studying summer data only when all year data
       August 2005
7-129
DRAFT-DO NOT QUOTE OR CITE

-------
 1      are available weakens the power of the study since less data are analyzed. In addition, increased
 2      adverse health outcomes are observed in the winter, some of which may be attributable to O3.
 3      The O3 effect in the wintertime may be masked by the effects of PM due to the negative
 4      correlation between these variables (see Section 7.6.4.2 for further discussion). Therefore,
 5      analysis of all year data may be improved by adjusting for PM indices in addition to adequate
 6      adjustment of meteorological factors and temporal trends.
 7           Seasonality influences the relationship between O3 and health outcomes as it may serve as
 8      an indicator for time-varying factors, including temperature, copollutant concentrations,
 9      infiltration, and human activity patterns. Given the potentially significant effect of season, O3
10      effect estimates computed for year-round data need to be interpreted with caution. Small or no
11      effects may simply reflect the cancellation of positive associations in the summer and negative
12      associations in the winter, or the presence of confounding due to the strong seasonal character of
13      O3 concentrations.
14
15      7.6.4   Assessment  of Confounding by Copollutants
16           Potential confounding by daily variations in copollutants is another analytical issue to be
17      considered.  With respect to copollutants, daily variations in O3 tend to not correlate highly with
18      most other criteria pollutants (e.g., CO, NO2, SO2, PM10), but may be more correlated with
19      secondary fine PM (e.g., PM25, sulfates) measured during the summer months. Assessing the
20      independent health effects of two pollutants that are correlated over time is not straightforward.
21      If high correlations between O3 and PM or other gaseous pollutants exist in a given area, then
22      disentangling their relative individual contributions to observed health effects associations
23      becomes very difficult.  The changing relationship between O3 and other copollutants also is of
24      issue. In some urban locations, the correlation between PM indices and O3 is positive in the
25      summer and negative in the winter. This section will further discuss the correlation between O3
26      and copollutants and confounding of the O3 effect by copollutants.
27
28      7.6.4.1   Relationship between Personal Exposure to Ozone and Copollutants
29           Ambient levels of PM, NO2, SO2, and CO, measured at central monitoring sites, have been
30      found to be highly correlated to ambient O3 concentrations. A very limited number of studies
31      have examined the association between personal O3 concentrations and personal exposures to

        August 2005                              7-130        DRAFT-DO NOT QUOTE OR CITE

-------
 1      other copollutants.  An issue of particular interest is the correlation between personal exposure to
 2      O3 and personal exposure to the ambient component of PM2 5. Only one study examined
 3      personal exposure to PM of ambient origin. In a Baltimore, MD study of susceptible populations
 4      (older adults, individuals with COPD, and children), Sarnat et al. (2001) found that ambient 24-h
 5      avg O3 concentrations and ambient 24-h avg PM25 levels were positively associated (P = 0.84,
 6      r = 0.67) in the summer and negatively associated (P = -0.67, r = -0.67) in the winter.
 7      A significant association also was observed between ambient O3 concentrations and personal
 8      PM2 5 of ambient origin, with a mixed regression effect estimate of P = 0.37 (95% CI:  0.25,
 9      0.49) in the summer and P = -0.36 (95% CI: -0.31, -0.41) in the winter. However, no
10      relationship was found between 24-h avg personal O3  exposure and personal exposure to PM25
11      of ambient origin.  While the  results from this study provide limited evidence for a lack of an
12      association between personal  O3 levels and personal exposure to PM25 of ambient origin,
13      additional research  is necessary to address this issue.
14
15      7.6.4.2  Assessment of Confounding Using Multipollutant Regression Models
16           Multipollutant regression models are generally used to determine whether the pollutant-
17      specific effect is robust. However, due to the multicollinearity among O3 and pollutants, and the
18      changing correlations by seasons, multipollutant models may not adjust for potential
19      confounding adequately, especially when using year-round data. These limitations need to be
20      considered when evaluating results from multipollutant models. Results from the U.S. 90 cities
21      study, which included 80 cities with O3 data, indicated that while the addition  of PM10 in the
22      model did not substantially change the O3-mortality risk estimates, slight declines in the O3
23      effect were observed, as shown in Figure 7-28 (Samet et al., 2000; reanalysis Dominici et al.,
24      2003). In the extended U.S. 95 communities study (Bell et al., 2004), the city-specific O3-
25      mortality effects were robust to adjustment for PM10, as indicated by the nearly 1:1 ratio between
26      estimates with and without PM10 adjustment shown in Figure 7-29. This finding suggested that
27      PM10 generally did  not confound the association between O3 and mortality.  Limited data were
28      available to examine the potential confounding effect of PM2 5 on the O3-mortality relationship.
29      A weighted second-stage linear regression indicated that there was no association between long-
30      term PM2 5 average  and the community-specific O3-mortality effect estimate. Several other
31      mortality and morbidity studies have investigated confounding of O3 risk estimates using

        August 2005                              7-131        DRAFT-DO NOT QUOTE OR CITE

-------
                       o
                       s
                       G)
                       C
                       re
                       ^
                       O
1.0-
U.b-
0-


-0.5-
-1.0-

j <





La
•
i
<
,



go
i
k <


i


i
» i

i
•


La
• ;
! i -

i


31
i

> *





T
1 *
i
i



•
r i




Li
1
» t

1
1

ig2

^ <


1



) (

1


                               ABCDEABCDEABCDE
                                                  Models
       Figure 7-28.  Posterior means and 95% Pis of 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
                     +PM  +SO; E = O  + PM  + CO.
                          10
                                2
                                              10
        Source: Derived from Dominici et al. (2003).
 1     multipollutant models with year-round data, and most have reported that O3 effects were robust
 2     to adjustment for copollutants (see Figures 7-11 and 7-20 in Sections 7.3.3 and 7.4.6,
 3     respectively).
 4           The pollutant most correlated with O3 in the summer is sulfate (which is in the fine particle
 5     size range), especially in the eastern U.S.  Therefore, the main potential confounders of interest
 6     for O3 are PM2 5 and sulfate in the summer. Once again, the results from two-pollutant
 7     regression models with O3 and sulfate (or PM2 5) should be interpreted with caution because both
 8     of these pollutants are formed under the same atmospheric condition and are both part of the
 9     "summer haze" pollution mix.  A simple two-pollutant regression model does not address their
10     possible synergistic effects, and the high correlation between the two pollutants may lead to
11     unstable and  possibly misleading results. In any case, most studies that analyzed O3 with PM
12     indices did not have PM25 data and very few examined sulfate data. The studies that did have
13     PM25 data, including Santa Clara County, CA (Fairley, 1999;  reanalysis Fairley, 2003),
14     Philadelphia, PA (Lipfert et al., 2000a), and Detroit, MI (Lippmann et al., 2000;  reanalysis Ito,
       August 2005
7-132
DRAFT-DO NOT QUOTE OR CITE

-------
                           10
                           Q
                           Q.
                              15 -
                              10 -
                               5 -
                           S§   o
                              -5 -
                                                           = National Average Effect
                                    -5
                                                               I
                                                               10
                       I
                       15
                                            Without PM10 Adjustment
                                         (using only days with PM10 data)
       Figure 7-29.  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).
 1     2003), examined copollutant models for year-round data only, but O3-mortality risk estimates
 2     were not substantially affected by the addition of PM25. The updated analysis of Philadelphia
 3     and Detroit data by season suggested that O3-mortality risk estimates were not sensitive to
 4     adjustment for PM2 5 in all year or seasonal analyses (Ito et al., 2005). A mortality study by
 5     Lipfert et al. (2000a) also found that all year O3 risk estimates were not affected by the addition
 6     of sulfate.
 7          Other studies have estimated O3 health risks with copollutants in the model by season.
 8     Respiratory hospitalization studies conducted during the warm season in Canada observed
 9     consistent O3 risk estimates with the inclusion of PM25 in the model (Burnett et al., 1997b,
10     2001). In one of these studies (Burnett et al., 1997b), the effect of O3 also was adjusted for
11     sulfate. With the addition of sulfate in the model, the risk estimate for O3 on respiratory
       August 2005
7-133
DRAFT-DO NOT QUOTE OR CITE

-------
 1      hospitalizations remained relatively stable, from an 14.4% (95% CI:  8.7, 20.5) excess risk to a
 2      11.7% (95% CI:  5.6, 18.0) excess risk per 25 ppb increase in 12-h avg O3 at a 1-day lag.
 3      In contrast, the effects for sulfate were reduced in half after adjusting for O3. Amongst the
 4      mortality studies  (see Figure 7-21 in Section 7.4.6), adjusting for copollutants, in particular PM
 5      indices, did not substantially change the warm season O3-mortality effect estimates, with both
 6      slight reductions  and increases observed in the adjusted estimates.  In the analysis using cool
 7      season data only, the O3 effect estimates were generally negative, but none were statistically
 8      significant. The O3 risk estimates all increased slightly with the adjustment of PM indices.
 9      The inverse relationship between O3 and PM during the cool season most likely influenced the
10      O3-mortality effect estimates in the single-pollutant models.
11           In field studies, power to assess independent O3 effects may be limited by small sample
12      sizes and short follow-up times.  Yet, the O3 effect also was robust to the addition of copollutants
13      in multipollutant  models, with a few exceptions.  For example, the  effect of O3 on PEF was not
14      robust to adjustments for PM25 and  sulfate, in studies by Romieu et al. (1996) and Neas et al.
15      (1999). In general, however, O3 effects on respiratory symptoms (Romieu et al.,  1996), lung
16      function parameters (Brauer et al., 1996, Gold et al., 1999), and asthma medication use (Gent
17      et al., 2003) were robust to inclusion of PM2 5. Further, the effects  for O3 were observed to be
18      stronger than those for PM.
19           Multipollutant regression analyses indicated that O3 risk estimates, in general, were not
20      sensitive to the inclusion of copollutants, including PM2 5 and sulfate.  These results suggest that
21      the effect of O3 on respiratory health outcomes appears to be robust and independent of the
22      effects of other copollutants.
23
24      7.6.5   Concentration-Response Function and Threshold
25           An important consideration in characterizing the public health impacts associated with O3
26      exposure is whether the concentration-response relationship is linear across the full
27      concentration range or instead shows evidence of a population threshold. Of particular interest is
28      the shape of the concentration-response curve at and below the level of the current 8-h standard
29      of 80 ppb. The slope of the O3 concentration-response relationship has been explored in several
30      studies.
        August 2005                              7-134        DRAFT-DO NOT QUOTE OR CITE

-------
 1           To examine the shape of the concentration-response relationship between O3 and mortality,
 2      Gryparis et al. (2004) used meta-smoothing to combine smooth curves across the 23 European
 3      cities in a hierarchical model. For the summer period, while the estimated concentration-
 4      response curve did not appear to deviate significantly from linearity, there were indications of
 5      decreasing effectiveness at lower exposures.
 6           In the U.S. 95 communities study (Bell et al., 2004), effect estimates calculated using only
 7      days with 24-h avg O3 levels less than 60 ppb were compared to those using all data. At a lag of
 8      1 day, O3 was associated with an excess risk of 0.36% (95% PI: 0.12, 0.60) per 20 ppb increase
 9      in 24-h avg O3 using data from all days and only a slightly smaller risk of 0.30% (95% PI: 0.08,
10      0.54) when data were limited to days less than 60 ppb. These results suggest that if there is a
11      threshold, it must be notably lower than a 24-h avg O3 of 60 ppb.
12           Fairley (2003) reanalyzed the Santa Clara County mortality data using GAM with stringent
13      convergence criteria and examined a new exposure index for O3. He noted O3 concentrations
14      exceeding 60 ppb each hour and calculated a daily sum of these exceedances. Fairley's index
15      incorporates  measures of concentration and exposure duration; this index represents a linear
16      time-integrated concentration, also known  as dosage. The O3 index with the 60 ppb "threshold
17      level" was found to be significantly associated with mortality in single-pollutant models as well
18      as in multi-pollutant models.  Two other "threshold levels" were examined, 40 ppb  and 80 ppb.
19      Both produced statistically significant results in single-pollutant models.  These results suggest
20      that the threshold for O3-mortality effects, if it exists, is likely less than 40 ppb. The implication
21      for thresholds in terms of the three standard indices (i.e.,  1-h max, 8-h max, and 24-h avg) is
22      unclear, but there may be an empirical relationship.
23           Vedal et al. (2003) observed that the  annual mean 1-h daily max O3 concentration of
24      27.3 ppb in Vancouver, Canada, was lower than that in any of the 80 NMMAPS cities (Samet
25      et al., 2000); thus, a study in this city might focus better on the shape of the concentration-
26      response curve at lower levels.  In this study, an O3 effect was observed on total mortality at a
27      0-day lag during the  summer. Ozone effects on respiratory mortality at a 2-day lag and
28      cardiovascular mortality at a 0-day lag also were observed in the summer.  The effect of O3 on
29      mortality was robust in two-pollutant models. Vedal et al. questioned if O3, other gaseous
30      pollutants, and PM were acting as surrogate markers of pollutant sources that contain more toxic
31      compounds,  since the low measured concentrations were unlikely, in their opinion,  to cause the

        August 2005                              7-135        DRAFT-DO NOT QUOTE OR CITE

-------
 1      observed effects. They further stated that measurement error and interference by meteorological
 2      factors might have contributed to the inability to detect a threshold.  Vedal et al. (2003)
 3      concluded that O3 concentrations were associated with adverse effects on mortality even at low
 4      levels. Although this study supports the argument that there are no threshold concentrations
 5      below which adverse effects cannot be detected, the results must be interpreted with caution as
 6      concerns remain.
 7           Kim et al. (2004) investigated the presence of a threshold in O3-mortality  effects in Seoul,
 8      Korea by analyzing data using a log linear GAM (linear model), a cubic natural spline model
 9      (nonlinear model), and a B-mode splined model (threshold model). Models were stratified by
10      season and adjusted for PM10, long-term time trend, and meteorological variables. An estimated
11      threshold value of 47 ppb was observed for 1-h daily max  O3. None of the other pollutants
12      examined, including PM10, SO2, NO2, and CO, had a nonlinear association with mortality.  Using
13      summer data only, the B-spline model resulted in an excess mortality risk of 7.1% (95% CI:  3.1,
14      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)
15      calculated using the log linear model. If a threshold truly exists, results from the Kim et al. study
16      suggest that the use of log-linear models may underestimate the O3 effect on mortality at levels
17      ab ove the thre shol d.
18           Other studies examining the effect of O3 on mortality also have found suggestive evidence
19      for a possible threshold level. In a London, England study (Anderson et al., 1996), an adjusted
20      O3-mortality bubble plot suggested that a threshold might exist around 50 ppb for 8-h avg O3.
21      A study  by Simpson et al. (1997) in Brisbane, Australia observed a significant excess risk in
22      mortality only in the highest quintile of O3 exposure, which had a mean concentration of 42 ppb
23      for 1-h max O3.
24           Among several  studies with morbidity outcomes, examination of the shape of the
25      concentration-response function indicated evidence of an effect threshold.  In a study of all-age
26      respiratory hospital admissions in Toronto, Canada, effects of O3 appeared to become apparent
27      only above approximately 30 ppb daily 1-h max O3 (Burnett et al., 1997b).  In London, England,
28      Ponce de Leon et al.  (1996) observed an indication of a threshold in the O3 effect on
29      hospitalizations at 40 to 50 ppb for 8-h max O3 and 50 to 60 ppb for 1-h max O3. In a study of
30      emergency department visits for asthma in St. John, Canada, effects observed in the over
31      15 years age group were apparent only when data above the 95th percentile (75 ppb daily 1-h

        August 2005                              7-136        DRAFT-DO NOT QUOTE OR CITE

-------
 1      max O3) were included (Stieb et al., 1996).  However, other morbidity studies observed a
 2      monotonic increase in the concentration-response function, suggesting that there was no
 3      threshold in O3 effects on hospitalizations and emergency department visits (Burnett et al.,
 4      1997a; Jaffe et al., 2003; Petroeschevsky et al., 2001; Tenias et al., 1998).
 5           In a field study by Mortimer et al. (2002), the associations of ambient O3 levels with PEF
 6      and asthma symptoms were investigated in eight urban cities in the U.S.  The mean 8-h avg O3
 7      was 48 ppb, with less than 5% of days exceeding 80 ppb. Analysis performed using all data
 8      indicated that a 15 ppb change in 8-h avg O3 was associated with decrements in PEF (-0.59%
 9      [95% CI:  -1.05, -0.13]) and increased incidence of respiratory symptoms (odds ratio of 1.16
10      [95% CI:  1.02, 1.30]) over multiday lag periods. When data were restricted to days when
11      ambient O3 concentrations were less than 80 ppb, the O3 effects persisted, with a significant PEF
12      decline (-0.70% [95% CI: -1.29, -0.12]) and incidence of morning symptoms (odds ratio of
13      1.17 [95%  CI:  1.01,  1.35]). A study by Chen et al. (1999)  also found that there was no clear
14      threshold in the O3 effect on FEVj and FVC in Taiwanese school children.
15           The studies of both Brauer et al. (1996) and Korrick et al. (1998) demonstrate that
16      exposure duration and exercise level, in addition to O3 concentration, must be considered when
17      evaluating  thresholds.  In the study by Brauer et al., the mean O3 concentration during the
18      11-hour work shift was 26.0 ppb (SD 11.8).  Workers experienced a change of -180.0 mL (95%
19      CI: -227.0, -133.0) in FEVj levels the next morning per 40 ppb increase in 1-h max O3. The
20      hikers in the study by Korrick et al.  (1998) were exposed to mean O3 levels of 40 ppb (SD 12)
21      over the duration of their hike (mean 8  hours).  Korrick et al. observed a mean change of -62.5
22      mL (95% CI:  -115.3, -9.7) in pre-hike to post-hike FEVj per 30 ppb increase in 8-h avg O3
23      when all hikers were included in the analysis; however, when analysis was restricted to hikers
24      with wheeze or asthma, a larger change of-182.5 mL (95% CI:  -312.2, -52.9) was observed.
25      In both studies, large reductions in lung function were observed in subjects exposed to relatively
26      low levels  of O3 over multiple hours while active outdoors.
27           Note  that adjusting for seasonal cycles does not address the issue of the changing
28      relationship between O3 concentrations and personal exposure across seasons. The ambient O3
29      levels are lower in the cold season, but people are likely to be exposed to even lower levels of
30      O3 during this season due to the shorter time spent outdoors and the longer time spent indoors
31      with closed windows. This is in contrast to what occurs with fine particles, which can

        August 2005                             7-137        DRAFT-DO NOT QUOTE OR CITE

-------
 1      effectively penetrate the indoors.  Thus, a more "accurate" concentration-response relationship
 2      may need to be examined in a summer-only data set.  Even for summer data, however, an
 3      interpretation of the relationship is not straightforward because of the possible influence of the
 4      use of air conditioning (an effective remover of O3). Greater use of air conditioning is expected
 5      on hot days when the O3 level is higher, but the use of air conditioning may also vary from city
 6      to city and across social class within a city. Using PM25 and sulfate as an example, Brauer et al.
 7      (2002) observed that surrogate measures of exposure (i.e., those from centrally-located ambient
 8      monitors) that were not highly correlated with personal exposures obscured the presence of
 9      thresholds in epidemiologic studies of larger populations. Likewise, exposure measurement
10      error may reduce the ability to detect a threshold in O3 population studies that used ambient O3
11      concentrations as an indicator of personal exposure.
12           Limited studies have examined the issue of thresholds in O3 health effects studies.  Some
13      studies have found a low level threshold while others have found  no threshold in O3  effects.
14      Levy et al. (2001) states that the molecular effects of O3 are mediated by antioxidants in the lung
15      lining fluid, which raises the possibility that there may be a threshold levels below which O3
16      would have few or no adverse effects. However, due to the variability in individual  sensitivities
17      and antioxidant  levels, this threshold may not be seen at the population level.
18           From  1990 to 2004, the 10th percentile values (which represent the lower concentration
19      range) of the warm season (May to September) 8-h max O3 concentrations averaged for all
20      available monitors throughout the U.S. were approximately 40 ppb (see discussion in Section
21      3.2).  While no conclusion can be made regarding the threshold issue, the limited evidence
22      suggests that if there is a threshold level in O3 health effects, it is likely near the lower limit
23      of ambient O3 concentrations in the U.S.
24
25      7.6.6   Heterogeneity of Ozone Health Effects
26           As described in Chapter 3 of this AQCD, O3 concentrations tend to be more spatially
27      variable than PM2 5 concentrations in urban areas.  In addition, relative personal exposures to O3
28      likely vary by region. The geographic variability in O3 concentrations and personal  exposures
29      may contribute to the heterogeneity in observed O3 health effects. The degree of influence of the
30      geographic variability on heterogeneity in  effects will vary by study as study design affects
31      different aspects of exposure (e.g., time period and duration of exposure).

        August 2005                              7-138       DRAFT-DO NOT QUOTE OR CITE

-------
 1           More than 80% of the O3-mortality estimates from the various studies conducted in North
 2      America, South America, Europe, and Australia were between 0.5 and 5% excess risk per 40 ppb
 3      increase in 1-h max O3 using year-round data. In general, the O3-mortality estimates were
 4      greater when using summer only data compared to year-round data. Though not all statistically
 5      significant, most of the O3-mortality estimates were greater than zero, indicating a positive
 6      relationship between O3 exposure and mortality.  The O3 risk estimates from the numerous
 7      hospitalization and emergency department visit studies were generally larger in magnitude and
 8      more variable from study to  study compared to the mortality studies. These differences in the
 9      O3 effect estimates may be attributable to the greater variability in the outcome measure in
10      hospitalization studies compared to mortality studies, such as  more subcategories of outcome
11      and varying degrees of severity.
12           Three recent meta-analyses that included both U.S.  and non-U.S. studies found consistent
13      all-year combined point estimates: 1.75% (95% PI:  1.10, 2.40),  1.6% (95% CI:  1.1,2.0), and
14      1.64% (95% CI:  1.25, 2.03) per 20 ppb increase in 24-h average O3, for Bell et al. (2005),
15      Ito et al. (2005), and Levy et al. (2005), respectively. Bell et al. further observed that the pooled
16      estimate for U.S. studies (11 estimates), 1.69% (95% PI:  0.94, 2.78), was similar to the pooled
17      estimate for the non-U.S. studies (30 estimates), 1.85% (95%  PI: 0.94, 2.78).  Levy et al.
18      compared North American studies to European studies and  also found nearly identical effect
19      estimates.
20           As differences in study design, population, and data analysis may affect risk estimates,
21      studies that were conducted in multiple cities using standardized  methods were further examined
22      to investigate the geographic heterogeneity of O3 effects.  Bell et al. (2004) conducted a time-
23      series analysis of O3 and mortality in 95 U.S. communities from 1987 to 2000. A 20 ppb
24      increase in 24-h avg O3 levels in the previous week was associated with an increase of 1.04%
25      (95% PI: 0.54, 1.55) excess risk of mortality in the pooled  analysis of 95  communities using all
26      available data. Intercity heterogeneity was observed among the 95 communities, which the
27      authors noted  as plausible given the city-specific differences in pollution characteristics, the use
28      of air conditioning, time-activity patterns, and socioeconomic factors.  Although some
29      heterogeneity  was observed among the communities (see Figure 7-16 of Section 7.4.3), the
30      range of the community-specific Bayesian estimates was  fairly narrow.  Note that the
31      community-specific Bayesian estimates are  shrunken estimates of the percent changes in daily

        August 2005                              7-139        DRAFT-DO NOT QUOTE OR CITE

-------
 1      mortality.  The larger the heterogeneity (across-community variance relative to within-
 2      community variance), the less the Bayesian estimates shrink toward the national average. Of the
 3      95 U.S. communities, 93 had positive O3-mortality risk estimates. Only 5 had risk estimates
 4      greater than 2.0% per 20 ppb increase in 24-h avg O3 during the previous week, with all
 5      communities indicating an excess mortality risk less than 3.5%.
 6           Greater heterogeneity was observed in the European study of 23 cities in 14 countries
 7      (Gryparis et al., 2004).  In the year-round analyses,  only 8 of the 23 cities had positive
 8      O3-mortality effect estimates. However, in the analyses using summer data only, the risk
 9      estimates were positive in 19 of the 23 cities, with a range of 0.8 to 8% excess risk per 40 ppb
10      increase in 1-h max O3. The heterogeneity may be attributable to the considerable variability
11      among countries in factors that may influence the relationship between  ambient O3
12      concentrations and personal exposure to O3, such as climate, use of air conditioning, personal
13      activity patterns, and socioeconomic factors.  In addition, the variability in the concentration and
14      composition of copollutants by cities or countries may contribute to the heterogeneity in the
15      O3-mortality effects. For example, concentrations of NO2 may vary widely by region, depending
16      on the differences in traffic density.
17           Among the hospitalization studies, Burnett et  al. (1997a) conducted the largest study of
18      16 Canadian cities.  The mean daily 1-h max O3 was 31 ppb in the 16 cities. The pooled O3
19      estimate was 5.6% (95% CI:  3.4, 7.9) excess risk in respiratory hospitalization per 40 ppb
20      increase in 1-h max O3 using warm season data (April to December).  The risk estimates were
21      fairly homogenous across the 16 Canadian cities, ranging from 3.1% for Vancouver to 7.7%  for
22      Quebec City.
23           Anderson et al. (1997) investigated the association between O3 and hospital admissions for
24      COPD in five European cities — London, Paris, Amsterdam, Rotterdam, and Barcelona.  The
25      pooled effect estimate was 5.0% (95% CI: 2.6, 7.6) excess risk per 30 ppb increase in 8-h max
26      O3 for year-round data.  Results from the APHEA study showed similar variability to that from
27      the Burnett et al. (1997a) study. The year-round effects estimates were lower in the two Dutch
28      cities (2.5% excess risk) compared to that in Paris (7.7% excess risk); however, analyses
29      indicated that there was no  significant heterogeneity in effects by city.   The authors further noted
30      that among the pollutants examined (O3, BS, TSP, SO2, and NO2), O3 had the most consistent
31      and significant findings.

        August 2005                              7-140        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Among the field studies, various respiratory health outcomes were examined, including
 2      PEF, spirometric parameters, respiratory symptoms, and medication use.  Only one field study
 3      investigated the O3 effect in several locations (Mortimer et al., 2002).  Mortimer et al. (2002)
 4      investigated the association of ambient O3 concentrations with PEF and asthma symptoms in
 5      asthmatic children living in eight urban cities in the U.S. — St. Louis, MO; Chicago, IL; Detroit,
 6      MI; Cleveland, OH; Washington, DC; Baltimore, MD; East Harlem, NY; and Bronx, NY. In the
 7      analysis pooling data from all eight cities, a 30 ppb increase in 8-h avg O3 was associated with a
 8      decrement of -1.18% (95%CI: -2.10, -0.26) in morning PEF for a 5-day cumulative lag period.
 9      The percent changes in PEF were negative in all cities except for Baltimore, 0.49%. Among the
10      other seven cities, the percent changes in PEF were quite homogenous, with values ranging from
11      -1.08% for Washington, DC to -1.71% for St. Louis.  A 30 ppb increase in 8-h avg O3 also was
12      associated with an increased incidence of morning symptoms in the pooled analysis (odds ratio
13      of 1.35 [95% CI: 1.04, 1.69] for a 4-day cumulative lag period). In all cities except for
14      St. Louis, there was an increase in the incidence of morning symptoms. In these cities, the odds
15      ratios for incidence of morning symptoms varied more compared to the PEF measurements,
16      ranging from 1.19 for Chicago to 2.96 for Detroit.  The greater variance may indicate the lack of
17      standardization in the use of symptoms as a health outcome measure.
18           Most of the multicity and meta-analyses studies consistently found positive associations
19      between O3 and mortality. Consistent O3 effects on hospitalizations and various respiratory
20      health outcomes also were found.  The observed heterogeneity of O3 effects may be partially
21      attributable to the use of centrally-located ambient monitors to assess exposure.  There may be
22      differences in relative personal exposures to O3 due to varying factors, such as use of air
23      conditioning and activity patterns, that affect the relationship between personal exposure and
24      ambient concentrations. For example, Levy et al. (2005) found suggestive evidence that air
25      conditioning prevalence was a predictor of heterogeneity in O3 risk estimates in their meta-
26      analysis.  The variability in the concentration and composition of other pollutants present also
27      may contribute to the heterogeneity of the effect of O3 on health outcomes as confounding by
28      copollutants may vary by region.
29
        August 2005                              7-141        DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.6.7   Health Effects of Ozone in Susceptible Populations
 2           In this section, the effects of O3 on morbidity and mortality in potentially susceptible
 3      populations will be examined. In epidemiologic studies of O3 health effects, the most widely
 4      studied subpopulation was asthmatics.  Also of interest were the observed health effects of O3 on
 5      different age groups, particularly children and the elderly.  This section begins with a discussion
 6      of the O3-related health effects in asthmatics.
 7
 8      7.6.7.1   Health Effects Associated with Ambient Ozone Exposure in Asthmatics
 9           Epidemiologic studies of health effects from acute O3 exposure in asthmatics have
10      examined a range of outcomes:  pulmonary function, respiratory symptoms, inflammation,
11      emergency room visits, hospital admissions, and mortality. Chronic O3 exposure studies have
12      investigated similar outcomes, with the exception of emergency room visits and hospitalizations.
13      Both are discussed in the earlier text.  This subsection draws together this information to
14      examine whether the evidence indicates that O3 exposure impacts asthmatics.
15           In Germany and Mexico City, O3 exposure was associated with a decline in FEVj in
16      asthmatic adults and children (Hoppe et al., 1995a, 2003; Romieu et al., 2002). Change in FEVj
17      also was examined in a group of asthmatic hikers  in Mount Washington, NH (Korrick et al.,
18      1998). Compared to the healthy subjects, the asthmatic subjects experienced a four-fold greater
19      decline in FEVj with the same exposure to O3 (mean change of -1.08% [95% CI: -2.49, 0.33]
20      versus -4.47% [95% CI: -7.65, -1.29] per 30 ppb increase in 8-h avg O3). The results from the
21      hiker study are consistent with those observed in controlled human exposure studies (discussed
22      in Chapter 6),  which also indicate greater decrements in FEVj among mild asthmatics versus
23      nonasthmatic subjects with heavy intermittent exercise.
24           PEF was examined in panels of asthmatic children in several field studies (see Figures 7-1
25      and 7-2).  Collectively, most of the studies indicated decrements of morning PEF, though only a
26      few estimates  were statistically significant. One multicity study of eight urban areas in the U.S.
27      observed O3-related reductions in morning PEF that were not significant in each individual  city
28      (Mortimer et al., 2002); however, the analysis combining data from all eight cities indicated a
29      significant decline in PEF with a cumulative lag of 1 to 5 days of O3 exposure. The odds ratio
30      for the incidence of > 10% decline in morning PEF was greater than one, which was discussed by
31      the author as an indication that O3 exposure might be associated with clinically important

        August 2005                              7-142       DRAFT-DO NOT QUOTE OR CITE

-------
 1      changes in PEF in asthmatic children.  The study examined 846 asthmatic children, the largest
 2      asthma panel study reported.
 3           Mortimer et al. (2000) observed that the subpopulation of asthmatic children with a history
 4      of low birth weight or premature birth had greater O3-associated declines in PEF (mean change
 5      of -3.66% [95% CI:  -5.30, -2.02] per 30 ppb increase in 8-h avg O3) than normal birth weight
 6      children (-0.60% [95% CI: -1.58, 0.38]).  Low birth weight and prematurity are associated with
 7      reduced lung function, higher levels of airway reactivity and increased susceptibility to lung
 8      damage (Barker et al.,  1993; Rona et al.,  1993), which may explain why these factors are found
 9      to increase susceptibility to respiratory insults of air pollution in children.
10           Lung function parameters have been evaluated for clinical significance. A reversible 5 to
11      15% decline in FEVj in an individual may have clinical importance to asthma morbidity
12      (American Thoracic Society, 1991; Lebowitz et al.,  1987; Lippmann,  1988). The National
13      Institutes of Health (1997) has stated that PEF below 80% of the personal best indicates a need
14      for additional medication use in asthmatics.  At a population level the mean changes in lung
15      function attributable to O3 exposure do not generally exceed 10% changes in FEVj or PEF per
16      standardized increment of O3.  At an individual  level, a subpopulation of susceptible asthmatics
17      are likely experiencing clinically significant declines in lung function. Hoppe et al. (2003)
18      examined effects  of O3 on the lung function of potential risk groups, performing both group and
19      individual analyses. For the group mean values, consistent O3 effects were not detectable.
20      On an individual basis, a potential pattern of O3 sensitivity was observed. About 20% of the
21      asthmatics and children were regarded as O3 responders (i.e., individuals with >10% change in
22      FEVj) compared to only 5% of the elderly and athletes.  These results indicated that while the
23      population as a whole was not reacting to O3, susceptible individuals were experiencing
24      clinically significant declines in lung function in response to O3 exposure.
25           Respiratory symptom increases in asthma panels were examined in several field studies,
26      some of which also examined PEF as discussed above.  The outcome  definition of symptoms
27      varied among these studies. Collectively, the results are suggestive of a potential O3 effect on
28      respiratory symptoms, but the evidence is not strong in the available studies. Two U.S. studies
29      that examined larger panels might be better to draw  inferences from as the large sample size
30      provided greater power to examine the effect of O3 on respiratory symptoms.  The eight U.S.
31      urban cities study reported that morning symptoms in the 846 asthmatic children were most

        August 2005                              7-143        DRAFT-DO NOT QUOTE OR CITE

-------
 1      strongly associated with a 4-day cumulative lag of O3 concentrations (Mortimer et al., 2002).
 2      A New England study examined 271 asthmatic children and observed an O3 effect on a variety
 3      of respiratory symptoms at a lag of 1 day among the 130 subjects who used maintenance asthma
 4      medications (Gent et al., 2003).
 5           Few epidemiologic studies have examined airway inflammation in asthmatics.  A Mexico
 6      City study indicated that supplementation with antioxidants may modulate the impact of O3
 7      exposure on the small airways of children with moderate-to-severe asthma (Romieu et al., 2002).
 8      A related study indicated that asthmatic children with GSTM1 null genotype were found to be
 9      more susceptible to the impact of O3 exposure on small airways (Romieu et al, 2004). A chronic
10      exposure study in Mexico City examined DNA  strand breaks in nasal epithelial cells in
11      asthmatic and nonasthmatics medical students and noted greater genotoxic damage in asthmatics
12      (Fortoul et al., 2003).
13           Emergency department visits for asthmatics have been examined in several  studies and
14      range from negative to positive results (see Figure 7-8 in Section 7.3.2). Warm season studies
15      tended to yield positive outcomes, as expected based on earlier discussions. Two studies in
16      Atlanta, GA (Tolbert et al., 2000) and Valencia, Spain (Tenias et al., 1998) indicated positive
17      effects in warm season analyses.  Further, a Canadian study, one of the larger studies conducted
18      in the summertime, reported a large increase in asthma emergency department visits when the
19      daily 1-h max O3 concentration exceeded 75 ppb (Stieb et al., 1996). A three-city study in Ohio
20      also indicated an increased risk of asthma visits during the summer (Jaffe et al., 2003).  Other
21      studies of mostly year-long data tended to produce inconsistent results, with some finding
22      negative estimates (Atkinson et al., 1999a;  Castellsague et al., 1995; Thompson et al., 2001;
23      Tobias  etal., 1999).
24           Hospital admission studies that specifically examined asthmatics were fewer in number
25      than those that  examined total respiratory diseases. Associations were noted in all age groups in
26      studies conducted in Seattle, WA (Sheppard, 2003), New Jersey (Weisel et al., 2002), Toronto,
27      Canada (Burnett et al., 1999), London, England (Anderson et al., 1998), Brisbane, Australia
28      (Petroeschevsky et al., 2001), and Hong Kong (Wong et al., 1999a).  However, several other
29      studies, mostly examining the effect of O3 on asthmatic children, did not observe  a significant
30      relationship (Gouveia and Fletcher, 2000a;  Lin et al., 2003; Morgan et al.,  1998; Nauenberg and
31      Basu, 1999; Schouten et al., 1996).

        August 2005                              7-144        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Acute mortality related to asthma was examined in Barcelona, Spain (Saez et al., 1999;
 2      Sunyer et al., 2002). Severe asthmatics with more than one asthma emergency visit showed the
 3      strongest mortality associations with O3 (Sunyer et al., 2002).
 4           Recent reports from longitudinal cohort studies in California have reported associations
 5      between the onset of asthma and long-term O3 exposures (Greer et al., 1993; McConnell et al.,
 6      2002; McDonnell et al., 1999). In adult studies, associations were seen in males but not females
 7      (Greer et al., 1993; McDonnell et al.,  1999). Among children residing in high O3 communities,
 8      McConnell et al. (2002) observed that asthma risk was elevated for those who played three or
 9      more sports as compared with those who did not play sports.  Playing sports may indicate
10      outdoor activity and an increased ventilation rate which may lead to increased dose of O3.  These
11      outcomes would benefit from replication in other cohorts in regards to indicating weight of a
12      causal interpretation.
13           A few studies provide limited discussion of concentration-response functions and
14      thresholds.  In the eight U.S. urban cities study, the odds ratios for incidence of > 10% decline in
15      morning PEF and incidence of morning symptoms when excluding days with 8-h avg O3 greater
16      than 80 ppb were nearly identical to those including data from all days (Mortimer et al., 2002).
17      In the New England asthma panel study (Gent et al., 2003), some of the associations for
18      symptoms occurred at 1-h max O3 levels below 60 ppb. In the St. John, Canada study (Stieb
19      et al., 2003), an effect of O3 on emergency department visits was reported with evidence of a
20      threshold somewhere in the range below a 1-h max O3 of 75 ppb in the  15 years and over age
21      group.
22           Overall, asthma subjects have been examined across most health endpoints of interest.  The
23      results reported in these studies vary with some indicating a positive excess risk associated with
24      O3. While no endpoint in itself seems to indicate an unquestionable demonstration of an
25      association, studies with adequate  sample size and power consistently provide positive results,
26      especially during the summer months when higher O3 levels occur. This view is strengthened as
27      positive results are obtained cohesively across the varied outcomes. Therefore, based on the
28      evidence, it seems prudent to consider asthmatics  as a susceptible group that requires  protection
29      from O3 exposures.
30
        August 2005                              7-145       DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.6.7.2  Age-Related Differences in Ozone Effects
 2           Several mortality studies have investigated age-related differences in O3 effects. Among
 3      the studies that observed positive associations between O3 and mortality, a comparison of all age
 4      or younger age (<65 years of age) O3-mortality risk estimates to that of the elderly population
 5      (>65 years) indicates that, in general, the elderly population is more susceptible to O3 effects
 6      (Borja-Aburto et al.  1997; Bremner et al., 1999; Gouveia and Fletcher 2000b; O'Neill et al.,
 7      2004; Simpson et al., 1997; Sartor et al., 1995; Sunyer et al., 2002). For example, a study by
 8      Gouveia and Fletcher (2000b) examined the O3-mortality effect by age in Sao Paulo, Brazil.
 9      There were 151,756 deaths for all non-violent causes over the period of 1991 to 1993, of which
10      49% occurred in the elderly.  Among all ages, O3 was associated with a 0.6% (95% CI:  -0.8,
11      2.0) excess risk in all cause mortality per 40 ppb increase in 1-h max O3.  In comparison, in the
12      elderly population, the O3-mortality risk estimate was nearly threefold greater, 1.7%  (95% CI:
13      0.0, 3.3).  Similarly, a Mexico City study found that O3-mortality risk estimates were 1.3% (95%
14      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
15      all ages and the elderly, respectively (O'Neill et al., 2004).
16           The meta-analysis by Bell et al. (2005) found a larger effect estimate for the elderly (2.92%
17      [95% PI:  1.34, 4.51] per 20 ppb increase in 24-h avg O3) than for all ages (1.75% [95% PI:
18      1.10, 2.37]). In the large U.S. 95 communities study (Bell et al., 2004), effect estimates were
19      slightly higher for those aged 65 to 74 years, 1.40% (95% PI:  0.56, 2.25) excess risk per 20 ppb
20      increase in 24-h avg O3, compared to individuals less than 65 years and 75 years or greater,
21      1.00% (95% PI: 0.20, 1.85) and 1.04% (95% PI:  0.36, 1.75), respectively, using a constrained
22      distributed 7-day lag model. Bell et al. (2004) notes that despite somewhat similar effect
23      estimates, the absolute effect  of O3 is substantially greater in the elderly population due to the
24      higher underlying  mortality rates, which leads to a larger number of extra deaths for the elderly
25      compared to the general population.
26           Few mortality  studies examined another potentially susceptible age group, young children
27      under the age of 5  years. The results were mixed, with one Mexico City study showing a lower
28      risk of O3-related all cause mortality in young children compared to all ages and the elderly
29      (Borja-Aburto et al., 1997) and one Sao Paulo, Brazil study showing a greater risk in respiratory
30      mortality in young children compared to the elderly (Gouveia and Fletcher, 2000b).  It should be
        August 2005                              7-146        DRAFT-DO NOT QUOTE OR CITE

-------
 1      noted that approximately 10% of mortality occurred in young children, thus the statistical power
 2      to study the O3 effect in this age group was limited.
 3           With respect to age-specificity of associations between O3 and acute respiratory
 4      hospitalizations or emergency department visits, no clear pattern emerges from recent studies.
 5      Associations have been reported for all ages (Anderson et al., 1997; Burnett et al., 1995, 1997b,
 6      1999; Weisel et al., 2002), adults or elderly (Burnett et al., 1997a; Delfino et al., 1997b, 1998b;
 7      Moolgavkar et al., 1997; Schwartz et al., 1996; Yang et al., 2003), and children (Burnett et al.,
 8      2001; Gouveia and Fletcher, 2000a; Lin et al., 1999; Ponka and Virtanen, 1996; Tolbert et al.,
 9      2000; Yang et al., 2003). Interestingly, studies that have examined effects in multiple age strata
10      often have seen effects only in non-pediatric strata (Delfino et al., 1997b, 1998b;  Stieb et al.,
11      1996; Jones et al., 1995). Several studies that focused on children did not report significant O3
12      effects, though in some cases these studies are limited by small size, inadequate control of
13      seasonal patterns, or very low O3 levels (Lierl and Hornung, 2003; Lin et al., 2003; Thompson
14      et al., 2001). If O3 is causally related to exacerbations of respiratory diseases leading to hospital
15      usage, one would expect to see effects most prominently  among children, for whom asthma is
16      more prevalent and O3 exposures may be greater. However, once again, children only comprised
17      of about 20% of the total hospitalizations, which limits the power to examine age-specific O3
18      effects.
19           A few field studies compared the effect of O3 in different age groups.  Korrick et al. (1998)
20      examined changes in FEVj and FVC related to O3 exposure in a group of hikers ranging in age
21      from 18 to 64 years, and found that there was no association between O3 responsiveness and age.
22      Brauer et al. (1996), in a study of berry pickers aged 10 to 69  years, also observed that subject
23      age was not significantly associated with O3-related changes in lung function.  However, a study
24      by Hoppe et al. (1995a, 2003) observed that children, but not  seniors (69 to 95 years of age),
25      experienced a decline in lung function associated with O3 exposure. The results by Hoppe et al.
26      are consistent with the diminishing responses to O3 exposure with increasing age  observed in
27      clinical studies. The clinical studies by Drechsler-Parks (1995) and Bedi et al. (1989) found that
28      subjects aged 56 to 89 years had markedly reduced responses to O3 exposure compared to young
29      adults.
30           Many field studies focused on the effect of O3 on the respiratory health of school children.
31      In general, children experienced decrements in pulmonary function parameters, including PEF,

        August 2005                              7-147        DRAFT-DO NOT QUOTE OR CITE

-------
 1      FEVls and FVC (Castillejos et al., 1995; Chen et al., 1999; Gielen et al., 1997; Gold et al., 1999;
 2      Jalaludin et al., 2000; Mortimer et al., 2002; Romieu et al., 1996; Thurston et al., 1997).
 3      Increases in respiratory symptoms (Delfino et al., 2003; Gold et al., 1999; Neas et al., 1995;
 4      Romieu et al., 1996,  1997; Thurston et al., 1997) and asthma medication use (Delfino et al.,
 5      1996; Just et al., 2002; Ostro et al., 2001; Thuston et al.,  1997) also were observed in children.
 6      These respiratory health effects were observed in both healthy and asthmatic children. In one
 7      German study (Hoppe et al., 2003), juvenile asthmatics and healthy children were found to be
 8      particularly susceptible to O3 effects on lung function.  Approximately 20% of the children and
 9      asthmatics experienced a greater than 10% change in FEVb compared to only 5% of the elderly
10      population and athletes.
11           The American Academy of Pediatrics (2004) notes that children and infants are among the
12      most susceptible to many air pollutants, including O3.  Eighty percent of alveolar are formed
13      postnatally and changes in the lung continue through adolescence (Dietert et al., 2000).  Children
14      spend more time outdoors, which results in increased exposure to air pollutants (Wiley et al.,
15      1991a,b). Further, children have a high minute ventilation and high levels of physical activity
16      which increase their dose (Plunkett et al., 1992).
17           Collectively, there is supporting evidence of age-related differences in susceptibility to O3
18      health effects. The elderly population (>65 years of age) appear to be at increased risk of
19      O3-related mortality and hospitalizations, and children (<18 years of age) experience other
20      potentially adverse respiratory health outcomes with increased O3 exposure.  One epidemiologic
21      study also found that the lung function response to O3 exposure may be diminished in elderly
22      populations; this finding is further supported by evidence from clinical  studies.
23
24      7.6.8   Summary of Key Findings and Conclusions Derived From Ozone
25              Epidemiologic Studies
26           In the previous 1996 O3  AQCD, there was considerable evidence  of O3-related respiratory
27      health effects from individual-level camp and exercise studies, as well as some consistent
28      evidence from time-series studies of emergency room visits and hospitalizations. Since the 1996
29      document, more field studies have been conducted, with  some emphasis on  additional outcome
30      markers such as respiratory symptoms and asthma medication use. Another significant addition
31      to the current O3 AQCD is the substantial number of short-term O3 mortality studies. The recent

        August 2005                              7-148        DRAFT-DO NOT QUOTE OR CITE

-------
 1      publication of an analysis examining the relationship between O3 and mortality in 95 U.S.

 2      communities (Bell et al., 2004) and three meta-analysis on O3-mortality associations (Bell et al.,

 3      2005; Ito et al., 2005; Levy et al., 2005) also contribute significantly to the evidence base.

 4      Considering the wide variability in possible study designs and statistical model specification

 5      choices, the reported O3 risk estimates for the various health outcomes are in reasonably good

 6      agreement. In the case of O3-mortality time-series studies, combinations of choices in model

 7      specifications (the number of weather terms and degrees of freedom for smoothing of mortality-

 8      temporal trends) alone may explain the extent of the difference in O3 risk estimates across

 9      studies.  As use of time-series studies to investigate air pollution effects has become more

10      common, there has been a great effort to evaluate the issues surrounding these studies.

11           In this section, conclusions regarding O3 health effects from the epidemiologic evidence

12      and the issues that may affect the interpretation of the effect estimates are briefly summarized.

13      A more  integrative synthesis of all relevant information will be presented in Chapter 8 of this
14      AQCD.

15
16           (1)   Field/panel studies of acute (X effects. Results from recent field/panel  studies
17                 continue to confirm that short-term O3 exposure is associated with acute decrements
18                 in lung function and increased respiratory symptoms, particularly in children and
19                 asthmatics. There is also suggestive evidence that O3 is related to increased asthma
20                 medication use.  Taken together with the evidence from controlled human exposure
21                 studies, O3 is likely causally related to the various respiratory health outcomes.  The
22                 current evidence is limited but supportive of a potential effect of O3 on heart rate
23                 variability, ventricular arrhythmias, and the incidence of myocardial infarctions.
24
25           (2)   Acute O3 effects on emergency department visits and hospitalizations.  Large
26                 multicity studies, as well as many studies from individual cities have reported an
27                 association of O3 concentrations with respiratory and cardiovascular hospital
28                 admissions. Studies using year-round data noted some inconsistencies  in the O3
29                 effect on daily hospitalizations. However, studies with data restricted to the summer
30                 or warm season, in general, indicated positive and robust associations between
31                 ambient O3 concentrations and cardiopulmonary hospital admissions. Results for
32                 emergency department visits are less  consistent.
33
34           (3)   Acute O? effects on mortality. The majority of the studies suggest an elevated risk
35                 of all cause mortality associated with acute exposure to O3, especially in the summer
36                 or warm season when O3 levels are typically high. Slightly greater O3 effects were
37                 observed for cardiovascular mortality. Results from a recent, large U.S. multicity
38                 time-series study provide the strongest evidence to-date for O3 effects on acute
39                 mortality.  Recent meta-analyses also showed consistent risk estimates  that are
        August 2005                               7-149       DRAFT-DO NOT QUOTE OR CITE

-------
 1                 unlikely to be confounded by PM; however, future work is needed to better
 2                 understand the influence of model specifications on the risk coefficient.
 3
 4           (4)    Chronic O? exposure effects on morbidity and mortality.  Fewer studies have
 5                 investigated the effect of chronic O3 exposure on morbidity and mortality.  The
 6                 strongest evidence is for negative seasonal effects of O3 on lung function in adults
 7                 and children.  Less conclusive are longer-term studies investigating the association
 8                 of chronic O3 exposure on yearly lung function, asthma incidence, and respiratory
 9                 symptoms. Chronic O3-mortality studies observed inconsistencies across exposure
10                 periods, cause-specific mortality outcomes, and gender.
11
12           (5)    Exposure assessment. Exposure misclassification may result from the use of
13                 stationary ambient monitors to determine exposure in population studies. Although
14                 central ambient monitors do not explain the variance of individual personal
15                 exposures, significant correlations are found between aggregate personal O3
16                 measurements and O3 concentrations from ambient monitors. A simulation study
17                 indicated that the use of ambient monitor data will tend to underestimate the O3
18                 effect.  A better understanding of the factors that affect the relationship between
19                 ambient concentrations  and personal exposures will improve interpretation of the O3
20                 effect estimates.
21
22           (6)    Ozone exposure indices. The three most commonly used daily O3 exposure indices,
23                 1-h max O3, 8-max O3, and 24-h avg O3, were found to be highly correlated in
24                 studies conducted in various regions.  In addition, the effect estimates and
25                 significance of associations across all health outcomes were comparable when using
26                 the  standardized distributional increment of 40 ppb, 30 ppb, and 20 ppb for 1-h max
27                 O3,  8-h max O3, and 24-h avg O3, respectively.
28
29           (7)    Lag structures for (X exposure and effect.  The lag time between O3 exposure and
30                 effect may differ depending on various factors such as the specific health outcome
31                 of interest, the mechanism of effect, and preexisting health conditions.  The maj ority
32                 of the studies found an immediate O3 effect, with the strongest associations
33                 observed between health outcomes and O3 exposure on the same  day  and/or
34                 previous day.  Some studies found large cumulative effects of O3 over longer
35                 lag  periods, indicating that multiday lags also may be relevant for some health
36                 outcomes, including mortality.
37
38           (8)    Sensitivity to model specifications for temporal trends. Ozone effect estimates
39                 that were reported in studies whose main focus was PM often were calculated using
40                 the  same model specifications as PM. While the sensitivity of the O3 risk estimates
41                 to alternative model  specifications has not been throughly investigated, limited
42                 evidence indicates that O3 effects may be robust to various model specifications for
43                 temporal trend adjustment.
44
45           (9)    Influence of seasonal factors. An evaluation of the confounding effects of
46                 meteorologic factors and copollutants on O3 risk estimates is complicated by their
47                 changing relationships with O3 across seasons. In addition, seasonal or seasonally-


        August 2005                              7-150       DRAFT-DO NOT QUOTE OR CITE

-------
 1                 modified factors (e.g., air conditioning use, time spent outdoors) complicate
 2                 interpretation of all year effect estimates as they affect the relationship between
 3                 ambient concentrations  and personal exposures. Given the potentially significant
 4                 influence of season, season-specific analyses are more informative in assessing O3
 5                 health risks.
 6
 7           (10)  Confounding by copollutants.  Multipollutant regression models often are used to
 8                 adjust for confounding by copollutants. Although there is some concern regarding
 9                 the use of multipollutant models given the varying concurvity across pollutants,
10                 results generally suggest that the inclusion of copollutants into the models do not
11                 substantially affect O3 risk estimates.  These findings indicate that effects of O3 on
12                 various health outcomes are robust and independent of the effects of other
13                 copollutants.
14
15           (11)  Concentration-response function. In the limited mortality and morbidity studies  that
16                 have specifically examined the O3 concentration-response relationship, the evidence
17                 is inconclusive regarding the presence of an effect threshold.  Factors such as
18                 exposure measurement error may reduce the ability to detect a threshold in
19                 population studies.
20
21           (12)  Heterogeneity of O? health effects. Consistent O3 effect estimates generally were
22                 observed for mortality, hospitalizations, and other respiratory health outcomes in
23                 multicity studies.  Some of the observed geographic heterogeneity in effects may be
24                 attributable to the differences in relative personal exposure to O3, which is affected
25                 by factors such as  air conditioning prevalence and activity patterns, and the varying
26                 concentrations and compositions of copollutants present by region.
27
28           (13)  Ozone health effects in asthmatics. The effects of O3 on asthmatics have been
29                 examined widely in both time-series  studies and panel studies. Associations of
30                 O3 with various respiratory health outcomes, including lung function declines,
31                 increased respiratory symptoms, and emergency department visits, were observed.
32                 These findings, along with the pathophysiologic understanding of asthma as a
33                 chronic inflammatory disease, indicate that asthmatics may be a susceptible
34                 population that requires protection from O3 exposures.
35
36           (14)  Age-related differences in O3 health effects. Supporting evidence exists for
37                 heterogeneity in the effects of O3 by age.  The elderly population (>65 years of age)
38                 appear to be at greater risk of O3-related mortality and hospitalizations compared to
39                 all age or younger populations. In addition, potentially adverse respiratory health
40                 outcomes were associated with O3 exposure in children (<18 years of age). One
41                 epidemiologic study provided limited evidence that lung function responses to O3
42                 exposure is diminished in the elderly population.
43
44
        August 2005                               7-151        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES
  2
  3       Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. J.; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
  4             Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir. Crit.
  5             Care Med. 159: 373-382.
  6       American Academy of Pediatrics, Committee on Environmental Health. (2004) Ambient air pollution: health
  7             hazards to children. Pediatrics 114: 1699-1707.
  8       American Thoracic Society. (1991) Lung function testing: selection of reference values and interpretative strategies
  9             Am. Rev. Respir. Dis. 144: 1202-1218.
10       Anderson, H. R.; Ponce de Leon, A.; Bland, J. M; Bower, J. S.; Strachan, D. P. (1996) Air pollution and daily
11             mortality in London: 1987-92. Br. Med. J. 312: 665-669.
12       Anderson, H. R.; Spix, C.; Medina, S.; Schouten, J. P.; Castellsague, J.; Rossi, G.; Zmirou, D.; Touloumi, G.;
13             Wojtyniak, B.; Ponka, A.; Bacharova, L.; Schwartz, J.; Katsouyanni, K. (1997) Air pollution and daily
14             admissions for chronic obstructive pulmonary disease in 6 European cities: results from the APHEA project.
15             Eur. Respir. J.  10: 1064-1071.
16       Anderson, H. R.; Ponce de Leon, A.; Bland, J. M.; Bower, J. S.; Emberlin, J.; Strachen, D. P. (1998) Air pollution,
17             pollens, and daily admissions for asthma in London 1987-92. Thorax 53: 842-848.
18       Anderson, H. R.; Bremner, S. A.; Atkinson, R. W.; Harrison, R. M.; Walters, S. (2001) Paniculate matter and daily
19             mortality and hospital admissions in the west midlands conurbation of the United Kingdom: associations with
20             fine and coarse particles, black smoke and sulphate. Occup. Environ. Med.  58: 504-510.
21       Atkinson, R. W.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Bremner, S. A.; Ponce de Leon, A. (1999a)
22             Short-term associations between outdoor air pollution and visits to accident and emergency departments in
23             London for respiratory complaints. Eur. Respir. J. 13: 257-265.
24       Atkinson, R. W.; Bremner, S. A.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Ponce de Leon, A. (1999b)
25             Short-term associations between emergency hospital admissions for respiratory and cardiovascular disease
26             and outdoor air pollution in London. Arch. Environ. Health 54: 398-411.
27       Atkinson, R. W.; Anderson, H. R.; Sunyer, J.; Ayres, J.; Baccini, M.;  Vonk, J. M.; Boumghar, A.; Forastiere, F.;
28             Forsberg, B.; Touloumi, G.; Schwartz, J.; Katsouyanni, K. (2001) Acute effects of paniculate air pollution on
29             respiratory admissions: results from APHEA 2 project. Am. J. Respir. Crit. Care Med. 164: 1860-1866.
30       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,
31             K. B.; D'Arcy,  J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone exposure and lung function in children
32             attending a southern California summer camp. Presented at: 83rd annual meeting and exhibition of the Air &
33             Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management
34             Association; paper no. 90-150.3.
35       Avol, E. L.; Navidi, W. C.; Rappaport, E. B.; Peters, J. M. (1998) Acute effects of ambient ozone on asthmatic,
36             wheezy, and healthy children. Cambridge, MA: Health Effects Institute; research report no. 82.
37       Avol, E. L.; Gauderman, W. J.; Tan, S. M.; London, S. J.; Peters, J. M. (2001) Respiratory effects of relocating to
38             areas of differing air pollution levels. Am. J. Respir. Crit. Care Med. 164: 2067-2072.
39       Ballester, F.; Tenias, J. M.; Perez-Hoyos, S. (2001) Air pollution and emergency hospital admissions for
40             cardiovascular diseases in Valencia, Spain. J. Epidemiol. Community Health 55: 57-65.
41       Barker, D. J. P.; Gluckman, P. D.; Godfrey, K. M.; Harding, J. E.; Owens, J. A.; Robinson, J. S. (1993) Fetal
42             nutrition and cardiovascular disease in adult life. Lancet 341: 938-941.
43       Bates, D. V. (2005) Ambient ozone and mortality. Epidemiology 16: 427-429.
44       Bates, D. V.; Sizto, R. (1983) Relationship between air pollutant levels and hospital admissions in Southern Ontario.
45             Can. J. Public Health 74: 117-122.
46       Bates, D. V.; Sizto, R. (1987) Air pollution and hospital admissions in southern Ontario: the acid summer haze
47             effect. Environ. Res. 43: 317-331.
48       Bates, D. V.; Sizto, R. (1989) The Ontario Air Pollution study: identification of the causative agent. Environ. Health
49             Perspect. 79: 69-72.
50       Bates, D. V.; Baker-Anderson, M.; Sizto, R. (1990) Asthma attack periodicity: a study of hospital emergency visits
51             in Vancouver. Environ. Res. 51: 51-70.
52       Bedi, J. F.; Horvath, S. M.; Drechsler-Parks, D. M. (1989) Adaptation by older individuals repeatedly exposed to
53             0.45 parts per million ozone for two hours. JAPCA 39: 194-199.
54       Beeson, W. L.; Abbey, D. E.; Knutsen, S. F. (1998) Long-term concentrations of ambient air pollutants and incident
55             lung cancer in California adults: results from the AHSMOG study. Environ. Health Perspect. 106: 813-823.
         August 2005                                    7-152        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Bell, M. L.; McDermott, A.; Zeger, S. L.; Samet, J. M; Dominici, F. (2004) Ozone and short-term mortality in
  2             95 US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292: 2372-2378.
  3       Bell, M. L.; Dominici, F.;  Samet, J. M. (2005) A meta-analysis of time-series studies of ozone and mortality with
  4             comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
  5       Bobak, M. (2000) Outdoor air pollution, low birth weight, and prematurity. Environ. Health Perspect. 108: 173-176.
  6       Bobak, M.; Leon, D. A. (1992) Air pollution and infant mortality in the Czech Republic, 1986-1988. Lancet (8826):
  7             1010-1014.
  8       Bobak, M.; Leon, D. A. (1999) Pregnancy outcomes and outdoor air pollution: an ecological study in districts of the
  9             Czech Republic 1986-8. Occup. Environ. Med. 56: 539-543.
10       Borja-Aburto, V. H.; Loomis, D. P.; Bangdiwala, S. I.; Shy, C. M.; Rascon-Pacheco, R. A. (1997) Ozone, suspended
11             particulates, and daily mortality in Mexico City. Am. J. Epidemiol. 145: 258-268.
12       Borja-Aburto, V. H.; Castillejos, M.; Gold, D. R.; Bierzwinski, S.; Loomis, D. (1998) Mortality and ambient fine
13             particles in southwest Mexico City, 1993-1995. Environ. Health Perspect. 106: 849-855.
14       Bourcier, T.; Viboud, C.; Cohen, J.-C.; Thomas, F.; Bury, T.; Cadiot, L.; Mestre, O.; Flahault, A.; Borderie, V.;
15             Laroche, L. (2003)  Effects of air pollution and climatic conditions on the frequency of ophthalmological
16             emergency examinations. Br. J. Ophthalmol. 87: 809-811.
17       Braga, A. L. F.; Conceicao, G. M. S.; Pereira, L.  A. A.; Kishi, H. S.; Pereira, J. C. R.; Andrade, M. F.; Goncalves,
18             F. L. T.; Saldiva, P. H. N.; Latorre, M. R. D. O. (1999) Air pollution and pediatric respiratory hospital
19             admissions in Sao Paulo, Brazil. J. Environ. Med. 1: 95-102.
20       Brauer, M.; Brook, J. R. (1997) Ozone personal exposures and health effects for selected groups residing in the
21             Fraser Valley. In: Steyn, D. G.; Bottenheim, J. W., eds. The Lower Fraser Valley  Oxidants/Pacific '93 Field
22             Study. Atmos. Environ. 31: 2113-2121.
23       Brauer, M.; Blair, J.; Vedal,  S. (1996) Effect of ambient ozone exposure on lung function in farm workers. Am.  J.
24             Respir. Crit. Care Med. 154: 981-987.
25       Bremner, S. A.; Anderson, H. R.; Atkinson, R. W.; McMichael, A. J.; Strachan, D. P.; Bland, J. M.; Bower, J. S.
26             (1999) Short term associations between outdoor air pollution and mortality in London 1992-4. Occup.
27             Environ. Med. 56: 237-244.
28       Brook, R. D.; Franklin, B.; Cascio, W.; Hong, Y.; Howard, G.; Lipsett, M.; Luepker, R.; Mittleman, M.; Samet, J.;
29             Smith, S. C., Jr.; Tager, I. (2004) Air pollution and cardiovascular disease. A statement for healthcare
3 0             professionals from the Expert Panel on Population and Prevention Science of the American Heart
31             Association. Circulation 109: 2655-2671.
32       Buchdahl, R.; Parker, A.; Stebbings, T.; Babiker, A. (1996) Association between air pollution and acute childhood
33             wheezy episodes: prospective observational study. Br. Med. J. 312: 661-664.
34       Buchdahl, R.; Willems, C. D.; Vander, M.; Babiker, A. (2000) Associations between ambient ozone, hydrocarbons,
3 5             and childhood wheezy episodes: a prospective observational study in south east London. Occup. Environ.
36             Med. 57: 86-93.
37       Burchfiel, C. M.; Marcus, E. B.; Curb, J. D.; Maclean, C. J.; Vollmer, W. M.; Johnson, L. R.; Fong, K.; Rodriguez,
38             B. L.; Masaki, K. H.;  Buist, A. S. (1995) Effects of smoking and smoking cessation on longitudinal decline in
39             pulmonary function. Am. J. Respir. Crit. Care Med. 151: 1778-1785.
40       Burnett, R. T.; Dales, R. E.;  Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M.;
41             Dann, T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of respiratory
42             admissions to Ontario hospitals. Environ. Res. 65: 172-194.
43       Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
44             paniculate sulfate and admissions to Ontario hospitals for cardiac and respiratory  diseases. Am. J. Epidemiol.
45             142: 15-22.
46       Burnett, R. T.; Brook, J. R.;  Yung, W. T.; Dales, R. E.; Krewski, D. (1997a) Association between ozone and
47             hospitalization for respiratory diseases in 16 Canadian cities.  Environ. Res. 72: 24-31.
48       Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997b) The role of paniculate size and chemistry in the
49             association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
50             Environ. Health Perspect.  105: 614-620.
51       Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of paniculate and gaseous air
52             pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54: 130-139.
53       Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.;  Leech, J. A.; Cakmak, S.;
54             Krewski, D. (2001) Association between ozone and hospitalization for acute respiratory diseases in children
55             less than 2 years of age. Am. J. Epidemiol. 153:444-452.
         August 2005                                    7-153         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Burr, D.; Doss, H. (2005) A Bayesian semiparametric model for random-effects meta-analysis. J. Am. Stat. Assoc.
  2             100:242-251.
  3       Calderon-Garciduefias, L.; Rodriguez-Alcaraz, A.; Garcia, R.; Ramirez, L.; Barragan, G. (1995) Nasal inflammatory
  4             responses in children exposed to a polluted urban atmosphere. J. Toxicol. Environ. Health 45: 427-437.
  5       Calderon-Garciduefias, L.; Osnaya, N.; Rodriguez-Alcaraz, A.; Villarreal-Calderon, A. (1997) DNA damage in nasal
  6             respiratory epithelium from children exposed to urban pollution. Environ. Mol. Mutagen. 30:  11-20.
  7       Calderon-Garciduefias, L.; Wen-Wang, L.; Zhang, Y.-J.; Rodriguez-Alcaraz, A.; Osnaya, N.; Villarreal-Calderon,
  8             A.; Santella, R. M. (1999) 8-hydroxy-2'-deoxyguanosine, a major mutagenic oxidative DNA lesion, and DNA
  9             strand breaks in nasal respiratory epithelium of children exposed to urban pollution. Environ. Health Perspect.
10             107:469-474.
11       Calderon-Garciduefias, L.; Valencia-Salazar, G.; Rodriguez-Alcaraz, A.; Gambling, T. M.; Garcia, R.; Osnaya, N.;
12             Villarreal-Calderon, A.; Devlin, R. B.; Carson, J. L. (2001) Ultrastructural nasal pathology in children
13             chronically and sequentially exposed to air pollutants. Am. J. Respir. Cell Mol. Biol. 24: 132-138.
14       Calderon-Garciduefias, L.; Mora-Tiscarefio, A.; Fordham, L. A.; Valencia-Salazar, G.; Chung, C. J.;
15             Rodriguez-Alcaraz, A.; Paredes, R.; Variakojis, D.; Villarreal-Calderon, A.; Flores-Camacho, L.;
16             Antunez-Solis, A.; Henriquez-Roldan, C.; Hazucha, M. J. (2003) Respiratory damage in children exposed to
17             urban pollution. Pediatr. Pulmonol. 36: 148-161.
18       Cassino, C.; Ito, K.; Bader, L; Ciotoli, C.; Thurston, G.; Reibman, J. (1999) Cigarette smoking and ozone-associated
19             emergency department use for asthma by adults in New York City. Am. J. Respir. Crit. Care Med.
20             159: 1773-1779.
21       Castellsague, J.; Sunyer, J.; Saez, M.; Anto, J. M. (1995) Short-term association between air pollution and
22             emergency room visits for asthma in Barcelona. Thorax 50: 1051-1056.
23       Castillejos, M.; Gold, D. R.; Damokosh, A. L; Serrano, P.; Allen, G.; McDonnell, W. F.; Dockery, D.; Velasco,
24             S. R.; Hernandez, M.; Hayes, C. (1995) Acute effects of ozone on the pulmonary function of exercising
25             schoolchildren from Mexico City. Am. J. Respir. Crit. Care Med. 152:  1501-1507.
26       Chan, C.-C.; Wu, T.-H. (2005) Effects of ambient ozone exposure on mail carriers' peak expiratory flow rates.
27             Environ. Health Perspect. 113: 735-738.
28       Chang, C.-C.; Tsai, S.-S.; Ho, S.-C.; Yang, C.-Y. (2005) Air pollution and hospital admissions for cardiovascular
29             disease in Taipei, Taiwan. Environ. Res. 98: 114-119.
30       Charpin, D.; Pascal, L.; Birnbaum, J.; Armengaud, A.; Sambuc, R.; Lanteaume, A.; Vervloet, D. (1999) Gaseous air
31             pollution and atopy.  Clin. Exp. Allergy 29: 1474-1480.
32       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)
3 3             Adverse effect of air pollution on respiratory health of primary school children in Taiwan. Environ. Health
34             Perspect. 106:331-335.
35       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
36             the pulmonary function of children in primary school. Environ. Health Perspect. 107:  921-925.
37       Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. (2000) Elementary school absenteeism and air pollution.
38             Inhalation Toxicol. 12: 997-1016.
39       Chen, L.; Yang, W.; Jennison, B. L.; Goodrich, A.; Omaye, S. T. (2002) Air pollution and birth weight in northern
40             Nevada, 1991-1999. Inhalation Toxicol. 14: 141-157.
41       Chew, F. T.; Goh, D. Y. T.; Ooi, B. C.; Saharom, R.; Hui, J. K. S.; Lee, B. W. (1999) Association of ambient
42             air-pollution levels with acute asthma exacerbation among children in Singapore. Allergy (Copenhagen)
43             54:320-329.
44       Chock, D. P.; Winkler, S. L.; Chen, C. (2000) A study of the association between daily mortality and ambient air
45             pollutant concentrations in Pittsburgh, Pennsylvania.  J. Air Waste Manage. Assoc. 50: 1481-1500.
46       Cifuentes, L. A.; Vega,  J.; Kopfer, K.; Lave, L. B.  (2000) Effect of the  fine fraction of paniculate matter versus the
47             coarse mass and other pollutants on daily mortality in Santiago, Chile. J. Air Waste Manage. Assoc.
48             50: 1287-1298.
49       Clyde, M. A. (1999) Bayesian model averaging and model search strategies. In: Bernardo, J. M.; Berger, J. O.;
50             Dawid, A. P.; Smith, A. F. M., eds. Bayesian Statistics 6: proceedings of the Sixth Valencia International
51             Meeting, June; pp. 157-185. Oxford, UK. Oxford, UK: Clarendon Press.
52       Clyde, M. (2000) Model uncertainty and health effect studies for paniculate matter. Environmetrics 11: 745-763.
53       Clyde, M. A.; Guttorp, P.; Sullivan, E. (2000) Effects of ambient fine and coarse particles on mortality in Phoenix,
54             Arizona.  Seattle, WA: University of Washington, National Research Center for Statistics and the
55             Environment; NRCSE technical report series, NRCSE-TRS no. 040. Available:
56             http://www.nrcse.washington.edu/pdf/trs40_pm.pdf [18 October, 2004].
         August 2005                                    7-154         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Cody, R. P.; Weisel, C. P.; Birnbaum, G.; Lioy, P. J. (1992) The effect of ozone associated with summertime
  2             photochemical smog on the frequency of asthma visits to hospital emergency departments. Environ. Res.
  3             58: 184-194.
  4       Cross, D.; Nelson, H. S. (1991) The role of the peak flow meter in the diagnosis and management of asthma.
  5             J. Allergy Clin. Immunol.  87: 120-128.
  6       Cuijpers, C. E. J.; Swaen, G. M. H.; Wesseling, G.; Wouters, E. F. M. (1994) Acute respiratory effects of summer
  7             smog in primary school children. Toxicol. Lett. 72: 227-235.
  8       Dab, W.; Medina, S.; Quenel, P.; Le Moullec, Y.; Le Tertre, A.;  Thelot, B.; Monteil, C.; Lameloise, P.; Pirard, P.;
  9             Momas, I.; Ferry, R.; Festy, B. (1996) Short term respiratory health effects of ambient air pollution: results of
10             the APHEA project in Paris. In: St Leger, S., ed. The APHEA project. Short term effects of air pollution on
11             health: a European approach using epidemiological time series data. J. Epidemiol. Commun. Health
12             50(suppl. 1): S42-S46.
13       Dejmek, J.; Selevan, S. G.; Benes, I.; Solansky, I.; Sram, R. J. (1999) Fetal growth and maternal exposure to
14             paniculate matter during pregnancy. Environ. Health Perspect. 107: 475-480.
15       Dejmek, J.; Solansky, I.; Benes, I.; Lenicek, J.; Sram, R. J. (2000) The impact of polycyclic aromatic hydrocarbons
16             and fine particles on pregnancy outcome. Environ. Health Perspect. 108: 1159-1164.
17       De Leon, S. F.; Thurston, G. D.; Ito, K. (2003) Contribution of respiratory disease to nonrespiratory mortality
18             associations with air pollution. Am. J. Respir. Crit. CareMed. 167:  1117-1123.
19       Delfino, R. J.; Coate, B. D.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Koutrakis, P. (1996) Daily asthma severity in
20             relation to personal ozone exposure and outdoor fungal spores. Am. J. Respir. Crit. Care Med. 154: 633-641.
21       Delfino, R. J.; Zeiger, R. S.;  Seltzer, J. M.; Street, D. H.; Matteucci, R. M.; Anderson, P. R.; Koutrakis, P. (1997a)
22             The effect of outdoor fungal spore concentrations on daily asthma severity. Environ. Health Perspect.
23             105:622-635.
24       Delfino, R. J.; Murphy-Moulton, A. M.; Burnett, R. T.; Brook, J. R.; Becklake, M. R. (1997b) Effects of air
25             pollution on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J. Respir. Crit. Care
26             Med. 155: 568-576.
27       Delfino, R. J.; Zeiger, R. S.;  Seltzer, J. M.; Street, D. H. (1998a) Symptoms inpediatric asthmatics and air pollution:
28             differences in effects by symptom severity, anti-inflammatory medication use and paniculate averaging time.
29             Environ. Health Perspect.  106: 751-761.
30       Delfino, R. J.; Murphy-Moulton, A. M.; Becklake, M. R. (1998b) Emergency room visits for respiratory illnesses
31             among the elderly in Montreal: association with low level ozone exposure. Environ. Res. 76: 67-77.
32       Delfino, R. J.; Gone, H.; Linn, W. S.; Pellizzari, E. D.; Hu, Y. (2003) Asthma symptoms in Hispanic children and
3 3             daily ambient exposures to toxic and criteria air pollutants. Environ. Health Perspect. Ill: 647-656.
34       Delfino, R. J., Quintana, P. J. E.; Flora, J.; Gastanaga, V. M.; Samimi, B. S.; Kleinman, M. T.; Liu, L.-J.  S.;
35             Bufalino, C.; Wu, C.-F.; McLaren, C. E. (2004) Association of FEVj in asthmatic children with personal and
36             microenvironmental exposure to airborne paniculate matter. Environ. Health Perspect. 112: 932-941.
37       Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Squinazi, F.; Momas, I. (2002a) Short-term effects of low-level air
3 8             pollution on respiratory health of adults suffering from moderate to severe asthma. Environ. Res. A 89: 29-37.
39       Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Le Moullec, Y.; Momas,  I. (2002b) Effects of air pollution on adults
40             with chronic obstructive pulmonary disease. Arch. Environ. Health 57: 554-560.
41       Detels, R.; Tashkin, D. P.; Sayre, J. W.; Rokaw, S. N.; Coulson,  A. H.; Massey, F. J., Jr.; Wegman, D. H. (1987)
42             The UCLA population studies of chronic obstructive respiratory disease: 9. lung function changes associated
43             with chronic exposure to photochemical oxidants; a cohort study among never-smokers. Chest 92: 594-603.
44       Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
45             Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the
46             lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
47       Diaz, J.; Garcia, R.; Ribera, P.; Alberdi, J. C.; Hernandez, E.; Pajares, M. S.; Otero, A. (1999) Modeling of air
48             pollution and its relationship with mortality and morbidity in Madrid, Spain. Int. Arch. Occup. Environ.
49             Health 72: 366-376.
50       Dietert, R. R.; Etzel, R. A.; Chen, D.; Halonen, M.; Holladay, S. D.; Jarabek, A. M.; Landreth, K.; Peden, D. B.;
51             Pinkerton, K.; Smialowicz, R.  J.; Zoetis, T. (2000) Workshop to identify critical window of exposure for
52             children's health: immune  and respiratory systems work group summary. Environ. Health Perspect. Suppl.
53             108(3): 483-490.
54       Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with  particulates
55             and acid aerosols. Environ. Res. 59: 362-373.
         August 2005                                    7-155         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Dockery, D. W.; Luttmann-Gibson, H.; Rich, D. Q.; Link, M. S.; Mittleman, M. A.; Gold, D. R.; Koutrakis, P.;
  2             Schwartz, J. D.; Verrier, R. L. (2005) Association of air pollution with increased incidence of ventricular
  3             tachyarrhythmias recorded by implanted cardioverter defibrillators. Environ. Health Perspect. 113: 670-674.
  4       Dominici, F.; McDermott, A.; Zeger, S. L.; Samet, J. M. (2002) On the use of generalized additive models in
  5             time-series studies of air pollution and health. Am. J. Epidemiol.  156: 193-203.
  6       Dominici, F.; McDermott, A.; Daniels, M.; Zeger, S. L.; Samet, J. M. (2003) Mortality among residents of 90 cities.
  7             In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA: Health
  8             Effects Institute; pp. 9-24. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
  9       Drechsler-Parks, D. M. (1995) The dose-response relationship in older men exposed to ozone. Exp. Gerontol.
10             30:65-75.
11       Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
12             Perspect. 107: 637-641.
13       Fairley, D. (2003) Mortality and air pollution for Santa Clara County, California, 1989-1996. In: Revised analyses of
14             time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute; pp.
15             97-106. Available: http://www.healtheffects.org/news.htm [16 May, 2003].
16       Flachaire, E.; Nunez, O. (2002) Estimation of income distribution and detection of subpopulations: an explanatory
17             model. Paris, France: Document de travail de la MSB, EUREQua 2002.86, Universite Parisl
18             Pantheon-Sorbonne. Available:
19             http://eurequa.univ-parisl.fr/membres/flachaire/research/Flachaire_Nunez_02.pdf [14 June, 2005].
20       Fortoul, T. L; Valverde, M.; Lopez M. D. C.; Bizarro, P.; Lopez, L; Sanchez,  I.; Colin-Barenque, L.; Avila-Costa,
21             M. R.; Rojas, E.; Ostrosky-Shejet, P. (2003) Single-cell gel electrophoresis assay of nasal epithelium and
22             leukocytes  from asthmatic and nonasthmatic subjects in Mexico City. Arch. Environ. Health 58: 348-352.
23       Friedman, M. S.; Powell, K. E.; Hutwagner, L.; Graham, L. M.; Teague, W. G. (2001) Impact of changes in
24             transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality and
25             childhood asthma. JAMA J. Am. Med. Assoc. 285: 897-905.
26       Frischer, T. M.; Ktihr, J.; Pullwitt, A.; Meinert, R.; Forster,  J.; Studnicka, M.; Koren, H. (1993) Ambient ozone
27             causes upper airways inflammation in children. Am.  Rev. Respir. Dis.  148: 961-964.
28       Frischer, T.; Pullwitt, A.; Kuehr, K.; Meinert, R.; Haschke,  N.; Studnicka, M.; Lubec, G. (1997)  Aromatic
29             hydroxylation in nasal lavage fluid following ambient ozone exposure. Free Radical Biol. Med. 22: 201-207.
30       Frischer, T.; Studnicka, M.; Gartner, C.; Tauber, E.; Horak, F.; Veiter, A.; Spengler, J.; Ktihr, J.; Urbanek, R. (1999)
31             Lung function growth and ambient ozone: a three-year population study in school children. Am. J. Respir.
32             Crit. Care Med. 160: 390-396.
33       Frischer, T.; Studnicka, M.; Halmerbauer, G.; Horak, F.; Gartner, C.; Tauber, E.; Koller, D. Y. (2001) Ambient
34             ozone exposure is associated with eosinophil activation in healthy children. Clin. Exp. Allergy 31: 1213-1219.
35       Fuhlbrigge, A.; Kitch, B.; Paltiel, A. D.; Kuntz, K. M.; Neumann, P. J.; Dockery, D. W.; Weiss,  S. T. (2001) FEVj is
36             associated with risk of asthma attacks in a pediatric population. J. Allergy Clin. Immunol. 107: 61-67.
37       Galbraith, R. F. (1994) Some applications of radial plots. J.  Am. Stat. Assoc. 89: 1232-1242.
3 8       Galizia, A.; Kinney, P. L. (1999) Long-term residence in areas of high ozone: associations with respiratory health in
39             a nationwide sample of nonsmoking young adults. Environ. Health Perspect. 107: 675-679.
40       Gamble, J. L. (1998) Effects of ambient air pollution on daily mortality: a time series analysis of Dallas, Texas,
41             1990-1994. Presented at: 91st annual meeting and exhibition of the Air & Waste  Management Association;
42             June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association;  paper no. 98-MP26.03.
43       Garcia-Aymerich, J.; Tobias, A.; Anto, J. M.; Sunyer, J. (2000) Air pollution and mortality in a cohort of patients
44             with chronic obstructive pulmonary disease: a time series analysis.  J. Epidemiol.  Community Health
45             54:73-74.
46       Garty, B. Z.; Kosman, E.; Ganor, E.; Berger, V.; Garty, L.; Wietzen, T.; Waisman, Y.; Mimouni, M.; Waisel, Y.
47             (1998) Emergency room visits of asthmatic children, relation to air pollution, weather, and airborne allergens.
48             Ann.  Allergy Asthma Immunol. 81: 563-570.
49       Gauderman, W. J.; McConnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
50             Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and lung
51             function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
52       Gauderman, W. J.; Gilliland, G. F.; Vora, H.; Avol, E.; Stram, D.; McConnell, R.; Thomas, D.; Lurmann, F.;
53             Margolis, H. G.; Rappaport, E. B.; Berhane, K.; Peters, J. M. (2002) Association between air pollution and
54             lung function growth in southern California children: results from a second cohort. Am. J. Respir. Crit. Care
55             Med. 166: 76-84.
         August 2005                                   7-156         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Gauderman, W. I; Avol, E.; Gilliland, F.; Vora, H.; Thomas, D.; Berhane, K.; McConnell, R.; Kuenzli, N.;
  2             Lurmann, F.; Rappaport, E.; Margolis, H.; Bates, D.; Peters, J. (2004a) The effect of air pollution on lung
  3             development from 10 to 18 years of age. N. Engl. J. Med. 351: 1057-1067.
  4       Gauderman, W. J.; Avol, E.; Gilliland, F. (2004b) Air pollution and lung function [reply letter]. N. Engl. J. Med.
  5             351:2653.
  6       Gent, J. F.; Triche, E. W.; Holford, T. R.; Belanger, K.; Bracken, M. B.; Beckett, W. S.; Leaderer, B. P. (2003)
  7             Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA
  8             J. Am. Med. Assoc. 290: 1859-1867.
  9       George, E. I. (1999) Comment on "Hoeting, J. A.; Madigan, D.; Raftery, A. E.; Volinsky, C. T. 1999. Bayesian
10             model averaging: a tutorial. Stat. Sci. 14: 409-412."
11       Gielen, M. H.;  VanDer Zee, S. C.; Van Wijnen, J. H.; Van Steen, C. J.; Brunekreef, B. (1997) Acute effects of
12             summer air pollution on respiratory health of asthmatic children. Am. J. Respir. Crit. Care Med.
13             155:2105-2108.
14       Gilliland, F. D.; Berhane, K.; Rappaport, E. B.; Thomas, D. C.; Avol, E.; Gauderman, W. J.; London, S. J.;
15             Margolis, H. G.; McConnell, R.; Islam, K. T.; Peters, J. M. (2001) The effects of ambient air pollution on
16             school absenteeism due to respiratory illnesses. Epidemiology 12: 43-54.
17       Gold, D. R.; Damokosh, A. I.; Pope, C. A., Ill; Dockery, D. W.; McDonnell, W. F.; Serrano, P.; Retama, A.;
18             Castillejos, M. (1999) Paniculate and ozone pollutant effects on the respiratory function of children in
19             southwest Mexico City. Epidemiology 10: 8-16.
20       Gold, D. R.; Litonjua, A.; Schwartz, J.; Lovett, E.; Larson, A.; Nearing, B.; Allen, G.; Verrier, M.; Cherry, R.;
21             Verrier, R. (2000) Ambient pollution and heart rate variability. Circulation 101: 1267-1273.
22       Gold, D. R.; Schwartz, J.; Litonjua, A.; Verrier, R.; Zanobetti, A. (2003) Ambient pollution and reduced heart rate
23             variability. In:  Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
24             Health Effects Institute; pp. 107-112. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
25             [18 October, 2004].
26       Goldberg, M. S.; Burnett, R. T. (2003) Revised analysis of the Montreal time-series study. In: Revised analyses of
27             time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute;
28             pp. 113-132. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [13 August 2003).
29       Goldberg, M. S.; Burnett, R. T.; Brook, J.; Bailar, J. C., Ill; Valois, M.-F.;  Vincent, R. (2001) Associations between
30             daily cause-specific mortality and concentrations of ground-level ozone in Montreal, Quebec. Am. J.
31             Epidemiol.  154: 817-826.
32       Goldberg, M. S.; Burnett, R. T.; Valois, M.-F.; Flegel, K.; Bailar, J. C., Ill; Brook, J.; Vincent, R.; Radon, K. (2003)
3 3             Associations between ambient air pollution and daily mortality among persons with congestive heart failure.
34             Environ. Res. 91: 8-20.
35       Gong, H., Jr.; Wong, R.; Sarma, R. J.; Linn, W. S.; Sullivan, E. D.; Shamoo, D. A.; Anderson, K. R.; Prasad, S. B.
36             (1998a) Cardiovascular effects of ozone exposure in human volunteers. Am. J. Respir. Crit. Care Med.
37             158: 538-546.
38       Gong, H., Jr.; Simmons, M. S.; Linn, W. S.; McDonnell, W. F.; Westerdahl, D. (1998b) Relationship between acute
3 9             ozone responsiveness and chronic loss of lung function in residents  of a high-ozone community. Arch.
40             Environ. Health 53: 313-319.
41       Gonzales, M.; Ngo, L.; Hammond, S. K.; Tager, I. (2003) Validation of a questionnaire and microenvironmental
42             model for estimating past exposures to ozone. Int. J. Environ. Health Res. 13: 249-260.
43       Goodman, S. N. (2005) The methodologic ozone effect. Epidemiology  16:  430-435.
44       Goss, C. H.; Newsom, S. A.; Schildcrout, J.  S.; Sheppard, L.; Kaufman, J.  D. (2004) Effect of ambient air pollution
45             on pulmonary exacerbations and lung function in cystic fibrosis. Am. J. Respir. Crit. Care Med. 169: 816-821.
46       Gouveia, N.; Fletcher, T. (2000a) Respiratory diseases in children and outdoor air pollution in Sao Paulo, Brazil: a
47             time series analysis. Occup. Environ. Med. 57: 477-483.
48       Gouveia, N.; Fletcher, T. (2000b) Time series analysis of air pollution and mortality: effects by cause, age and
49             socioeconomic status. J. Epidemiol. Community Health 54: 750-755.
50       Gouveia, N.; Bremner, S. A.; Novaes, H. M. D. (2004) Association between ambient air pollution and birth weight
51             in Sao Paulo, Brazil. J. Epidemiol. Community Health 58: 11-17.
52       Greer, J. R.; Abbey, D. E.; Burchette, R. J. (1993) Asthma related to occupational and ambient air pollutants in
53             nonsmokers. J. Occup. Med. 35: 909-915.
         August 2005                                    7-157         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Gryparis, A.; Forsberg, B.; Katsouyarmi, K.; Analitis, A.; Touloumi, G.; Schwartz, J.; Samoli, E.; Medina, S.;
  2             Anderson, H. R.; Niciu, E. M.; Wichmann, H.-E.; Kriz, B.; Kosnik, M.; Skorkovsky, J.; Vonk, J. M.;
  3             Dortbudak, Z. (2004) Acute effects of ozone on mortality from the "air pollution and health: a European
  4             approach" project. Am. J. Respir. Crit. Care Med. 170: 1080-1087.
  5       Gwynn, R. C.; Thurston, G. D. (2001) The burden of air pollution: impacts among racial minorities. Environ. Health
  6             Perspect. Suppl. 109(4): 501-506.
  7       Gwynn, R. C.; Burnett, R. T.; Thurston, G. D. (2000) A time-series analysis of acidic paniculate matter and daily
  8             mortality and morbidity in the Buffalo, New  York, region. Environ. Health Perspect. 108: 125-133.
  9       Ha, E.-H.; Hong, Y.-C.; Lee, B.-E.; Woo, B.-H.; Schwartz, J.; Christiani, D. C. (2001) Is air pollution a risk factor
10             for low birth weight in Seoul? Epidemiology 12:  643-648.
11       Hagen, J. A.; Nafstad,  P.; Skrondal, A.; Bjerkly, S.; Magnus, P. (2000) Associations between outdoor air pollutants
12             and hospitalization for respiratory diseases. Epidemiology 11: 136-140.
13       Hajat, S.; Haines, A.; Goubet, S. A.; Atkinson, R. W.; Anderson, H. R. (1999) Association of air pollution with daily
14             GP consultations for asthma and other lower respiratory conditions in London. Thorax 54:  597-605.
15       Hajat, S.; Anderson, H. R.; Atkinson, R. W.; Haines, A. (2002) Effects of air pollution on general practitioner
16             consultations for upper respiratory diseases in London. Occup. Environ. Med. 59: 294-299.
17       Hankinson, J. L.; Odencrantz, J.  R.; Fedan, K. B. (1999) Spirometric reference values from a sample of the general
18             U.S. population. Am. J. Respir. Crit.  Care Med. 159: 179-187.
19       Health Effects Institute. (2003) Revised analyses of time-series studies of air pollution and health. Boston, MA:
20             Health Effects Institute; special report. Available: Available:
21             http://www.healtheffects.org/Pubs/TimeSeries.pdf [27 June 2003].
22       Hedley, A. J.; Wong, C.-M.;  Thach, T. Q.; Ma, S.; Lam, T.-H.; Anderson, H. R. (2002) Cardiorespiratory and
23             all-cause mortality after restrictions on sulphur content of fuel in Hong Kong: an intervention study. Lancet
24             360: 1646-1652.
25       Hernandez-Garduno, E.; Perez-Neria, J.; Paccagnella, A. M.; Pina-Garcia, M.; Munguia-Castro, M.;
26             Catalan-Vazquez, M.; Rojas-Ramos, M. (1997) Air pollution and respiratory health in Mexico City. J. Occup.
27             Environ. Med. 39: 299-307.
28       Higgins, I. T. T.; D'Arcy, J. B.; Gibbons, D. I.; Avol, E. L.; Gross, K. B. (1990) Effect of exposures to ambient
29             ozone on ventilatory lung function in children. Am. Rev. Respir. Dis. 141:  1136-1146.
30       Hill, A.  B. (1965) The environment and disease: association or causation? Proc. R. Soc. Med. 58: 295-300.
31       Hiltermann,  T. J. N.; Stolk, J.; Van der Zee, S. C.; Brunekreef, B.; De Bruijne, C.  R.; Fischer, P. H.; Ameling, C. B.;
32             Sterk, P. J.; Hiemstra,  P. S.; Van Bree, L. (1998)  Asthma severity and susceptibility to air pollution. Eur.
33             Respir. J. 11:686-693.
34       Hoek, G. (2003) Daily mortality and air pollution in The Netherlands. In: Revised analyses of time-series studies of
35             air pollution and health. Special report. Boston, MA: Health Effects Institute; pp. 133-142. Available:
36             http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
37       Hoek, G.;  Brunekreef,  B. (1995) Effect of photochemical air pollution on acute respiratory symptoms in children.
3 8             Am. J. Respir. Crit. Care Med. 151: 27-32.
39       Hoek, G.;  Brunekreef,  B.; Verhoeff, A.; Van Wijnen, J.; Fischer, P. (2000) Daily mortality and air pollution in the
40             Netherlands. J. Air Waste Manage. Assoc. 50:  1380-1389.
41       Hoek, G.;  Brunekreef,  B.; Fischer, P.; Van Wijnen,  J. (2001) The  association between air pollution and heart failure,
42             arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time series study.
43             Epidemiology 12: 355-357.
44       Hoeting, J. A.; Madigan, D.;  Raftery, A. E.; Volinsky, C. T. (1999) Bayesian model averaging: a tutorial. Stat. Sci.
45             14:382-417.
46       Holguin, F.;  Tellez-Rojo, M.  M.; Hernandez, M.; Cortez, M.; Chow, J. C.; Watson, J. G.; Mannino, D.; Romieu, I.
47             (2003) Air pollution and heart rate variability among the elderly in Mexico City. Epidemiology 14: 521-527.
48       Holmen, A.; Blomqvist, J.; Frindberg, H.; Johnelius, Y.; Eriksson, N. E.; Henricson, K. A.; Herrstrom, P.;
49             Hogstedt, B. (1997)  Frequency of patients with acute asthma in relation to ozone, nitrogen dioxide, other
50             pollutants of ambient air and meteorological  observations. Int. Arch. Occup. Environ. Health 69: 317-322.
51       Hong, Y.-C.; Lee, J.-T.; Kim, H.; Ha, E.-H.; Schwartz, J.; Christiani, D. C.  (2002) Effects of air pollutants on acute
52             stroke mortality. Environ. Health Perspect. 110: 187-191.
53       Hoppe, P.; Praml, G.; Rabe, G.; Lindner, J.; Fruhmann,  G.; Kessel, R. (1995a) Environmental ozone field study on
54             pulmonary and  subjective responses of assumed risk groups. Environ. Res. 71: 109-121.
5 5       Hoppe, P.; Lindner, J.; Praml, G.; Bronner, N. (1995b) Effects of environmental ozone on the lung function of senior
56             citizens. Int. J. Biometeorol. 38:  122-125.
         August 2005                                   7-158         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Hoppe, P.; Peters, A.; Rabe, G.; Praml, G.; Lindner, I; Jakobi, G.; Fruhmann, G.; Nowak, D. (2003) Environmental
  2             ozone effects in different population subgroups. Int. J. Hyg. Environ. Health 206: 505-516.
  3       Horak, F., Jr.; Studnicka, M; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.; Prisoner, T. (2002a)
  4             Paniculate matter and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren.
  5             Eur. Respir. J. 19: 838-845.
  6       Horak, F., Jr.; Studnicka, M.; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.; Frischer, T. (2002b)
  7             Paniculate matter and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren
  8             [author response]. Eur. Respir. J. 20: 1355.
  9       Huang, Y.; Dominici, F.; Bell, M. L. (2005) Bayesian hierarchical distributed lag models for summer ozone
10             exposure and cardio-respiratory mortality. Environmetrics 16: 547-562.
11       Hubbell, B. J.; Hallberg, A.; McCubbin, D. R.; Post, E. (2005) Health-related benefits of attaining the 8-hr ozone
12             standard. Environ. HealthPerspect. 113: 73-82.
13       Hwang, J.-S.; Chan, C.-C. (2002) Effects of air pollution on daily clinic visits for lower respiratory tract illness.
14             Am. J. Epidemiol. 155: 1-10.
15       Ihorst, G.; Frischer, T.; Horak, F.; Schumacher, M.; Kopp, M.; Forster, J.; Mattes, J.; Kuehr, J. (2004) Long- and
16             medium-term ozone effects on lung growth including a broad spectrum of exposure. Eur. Respir. J.
17             23:292-299.
18       Ilabaca, M.; Olaeta, I.; Campos, E.; Villaire, J.; Tellez-Rojo, M. M.; Romieu, I. (1999) Association between levels
19             of fine paniculate and emergency visits for pneumonia and other respiratory illnesses among children in
20             Santiago, Chile. J. Air Waste Manage. Assoc. 49:  154-163.
21       Ito, K. (2003) Associations of paniculate matter components with daily mortality and morbidity in Detroit,
22             Michigan. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
23             Health Effects Institute; pp. 143-156. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May,
24             2004].
25       Ito, K. (2004) Revised ozone risk estimates for daily mortality in Detroit, Michigan [personal communication with
26             attachments to Jee Young Kim]. New York,  NY: New York University School of Medicine, Nelson Institute
27             of Environmental Medicine; October 31.
28       Ito, K.; Thurston, G. D. (1996) Daily PM10/mortality associations: an investigation of at-risk subpopulations.
29             J. Exposure Anal. Environ. Epidemiol. 6: 79-95.
30       Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and
31             meta-analysis. Epidemiology 16: 446-457.
32       Jaffe, D. H.; Singer, M. E.; Rimm, A. A. (2003) Air pollution and emergency department visits for asthma among
33             Ohio Medicaid recipients, 1991-1996. Environ. Res. 91: 21-28.
34       Jalaludin, B. B.; Chey, T.; O'Toole, B. I.; Smith, W. T.; Capon, A. G.; Leeder, S. R. (2000)  Acute effects of low
3 5             levels of ambient ozone on peak expiratory flow rate in a cohort of Australian children. Int. J. Epidemiol.
36             29: 549-557.
37       Jalaludin, B. B.; O'Toole, B. I.; Leeder, S. R. (2004) Acute effects of urban ambient air pollution on respiratory
3 8             symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ
39             Res. 95: 32-42.
40       Jedrychowski, W.; Maugeri, U.; Bianchi, I.; Flak, E. (2001) Transient or persistent asthma-like symptoms and lung
41             growth over 2-year follow-up in pre-adolescent children. J. Epidemiol. Biostat. 6: 229-233.
42       Jenkins, P. L.; Phillips, T. J.; Mulberg, E. J.; Hui, S. P. (1992) Activity patterns of Californians: use of and proximity
43             to indoor pollutant sources. Atmos. Environ. Part A 26: 2141-2148.
44       Jones, M. C. (1983) The projection pursuit algorithm for exploratory data analysis  [Ph.D. thesis]. Bath, England:
45             University of Bath.
46       Jones, G. N.; Sletten, C.; Mandry, C.; Brantley, P. J. (1995) Ozone level effect on respiratory illness: an
47             investigation of emergency department visits. South. Med. J. 88: 1049-1056.
48       Just, J.;  Segala, C.; Sahraoui, F.; Priol, G.; Grimfeld, A.;  Neukirch, F. (2002) Short-term health effects of paniculate
49             and photochemical air pollution in asthmatic children. Eur. Respir. J. 20: 899-906.
50       Kaiser, R.; Romieu, L; Medina, S.; Schwartz, J.; Krzyzanowski, M.; Ktinzli, N. (2004) Air pollution attributable
51             postneonatal infant mortality in U.S. metropolitan areas: a risk assessment study. Environmental Health:
52             a global access science source. Available: http;//www.ehjournal.net/content/3/l/4 [5 July, 2005].
53       Kim, S.-Y.; Lee, J.-T.; Hong, Y.-C.; Ahn, K.-J.; Kim, H.  (2004) Determining the threshold effect of ozone on daily
54             mortality: an analysis of ozone and mortality in Seoul, Korea, 1995-1999. Environ. Res. 94: 113-119.
55       Kinney, P. L.; Lippmann, M. (2000) Respiratory effects of seasonal exposures to ozone and particles. Arch. Environ.
56             Health 55: 210-216.
         August 2005                                    7-159        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Kinney, P. L.; Ozkaynak, H. (1991) Associations of daily mortality and air pollution in Los Angeles County.
  2             Environ. Res. 54: 99-120.
  3       Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
  4             In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
  5             mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
  6       Kinney, P. L.; Thurston, G. D.; Raizenne, M. (1996a) The effects of ambient ozone on lung function in children:
  7             a reanalysis of six summer camp studies. Environ. Health Perspect. 104: 170-174.
  8       Kinney, P. L.; Nilsen, D. M.; Lippmann, M.; Brescia, M.; Gordon, T.; McGovern, T.; El Fawal, H.; Devlin, R. B.;
  9             Rom, W. N. (1996b) Biomarkers of lung inflammation in recreational joggers exposed to ozone. Am. J.
10             Respir. Crit. Care Med.  154: 1430-1435.
11       Kinney, P. L.; Aggarwal, M.; Nikiforov, S. V.; Nadas, A. (1998) Methods development for epidemiologic
12             investigations of the health effects of prolonged ozone exposure. Part III: an approach to retrospective
13             estimation of lifetime ozone exposure using a questionnaire and ambient monitoring data (U.S. sites).
14             Cambridge, MA: Health Effects Institute; research report no. 81; pp.  79-108.
15       Klemm, R. J.; Mason, R. M., Jr. (2000) Aerosol Research and Inhalation Epidemiological Study (ARIES):
16             air quality and daily mortality statistical modeling—interim results. J. Air.  Waste Manage. Assoc.
17             50: 1433-1439.
18       Klemm, R. J.; Lipfert, F. W.; Wyzga, R. E.; Gust, C. (2004) Daily mortality and air pollution in Atlanta: two years
19             of data from ARIES. Inhalation Toxicol. 16(suppl. 1): 131-141.
20       Kochi, I.; Hubbell, B.; Kramer, R. (2005) An empirical Bayes approach to combining and comparing estimates of
21             the value of a statistical life for environmental policy analysis. Available:
22             http://www.nicholas.duke.edu/people/faculty/kramer/VSLrevised%204-6-05.pdf [11 August, 2005].
23       Koken, P. J.; Piver, W. T.; Ye, F.; Elixhauser, A.; Olsen, L. M.; Portier, C. J. (2003) Temperature, air pollution,
24             and hospitalization for cardiovascular diseases among elderly people in Denver. Environ. Health Perspect.
25             111:  1312-1317.
26       Koop, G.; Tole, L. (2004) Measuring the health effects of air pollution: to what extent can we really say that people
27             are dying from bad air?  J. Environ. Econ. Manage. 47: 30-54.
28       Kopp, M. V.; Ulmer, C.; Ihorst, G.; Seydewitz, H. H.;  Frischer, T.; Forster,  J.; Kuehr, J. (1999) Upper airway
29             inflammation in children exposed to ambient ozone and potential signs of adaptation. Eur. Respir. J.
30             14:854-861.
31       Kopp, M. V.; Bonnet, W.; Frischer, T.; Ulmer, C.; Studnicka, M.; Ihorst,  G.; Gardner, C.; Forster, J.; Urbanek, R.;
32             Kuehr, J. (2000) Effects of ambient ozone on lung function in  children over a two-summer period. Eur.
33             Respir. J.  16: 893-900.
34       Koren, H. S.; Hatch, G. E.; Graham, D. E. (1990) Nasal lavage  as a tool in assessing acute inflammation in response
35             to inhaled pollutants. Toxicology 60: 15-25.
36       Korrick, S. A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K. D.; Rosner, B. A.;
37             Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
38             Environ. Health Perspect. 106: 93-99.
39       Krzyzanowski, M.; Quackenboss, J. J.; Lebowitz, M. D.  (1992) Relation of peak expiratory flow rates and symptoms
40             to ambient ozone. Arch. Environ. Health 47: 107-115.
41       Ktinzli, N.; Lurmann, F.; Segal, M.; Ngo, L.; Balmes, J.; Tager, I. B.  (1997) Association between lifetime ambient
42             ozone exposure and pulmonary function in college freshmen—results of a pilot study. Environ. Res.  72: 8-23.
43       Kuo, H. W.; Lai, J. S.; Lee, M.  C.; Tai, R. C.; Lee, M. C. (2002) Respiratory effects of air pollutants among
44             asthmatics in central Taiwan. Arch. Environ. Health 57:  194-200.
45       Kwon, H.-J.; Cho, S.-H.; Nyberg, F.; Pershagen, G. (2001) Effects of ambient air pollution on daily mortality in a
46             cohort of patients with congestive heart failure. Epidemiology 12:413-419.
47       Lagerkvist, B. J.; Bernard, A.; Blomberg, A.; Bergstrom, E.; Forsberg, B.; Holmstrom, K.; Karp, K.; Lundstrom,
48             N.-G.; Segerstedt, B.; Svensson, M.; Nordberg, G. (2004) Pulmonary epithelial integrity in children:
49             relationship to ambient ozone exposure and swimming pool attendance. Environ. Health Perspect.
50             112:  1768-1771.
51       Langstaff, J. (2003) Percentiles of 1996-2000 ozone concentrations [memorandum to Joe Pinto]. Research Triangle
52             Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
53             September 17.
         August 2005                                    7-160         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Lebowitz, M. D.; Camilli, A. E.; Bronnimann, D.; Quackenboss, J. (1987) The significance and meaningfulness of
  2             intraindividual changes in objective test results as responses to air contaminants. Presented at: 80th annual
  3             meeting of the Air Pollution Control Association; June; New York, NY. Pittsburgh, PA: Air Pollution Control
  4             Association; paper no. 87-32.1.
  5       Lebowitz, M. D.; Quackenboss, J. J.; Krzyzanowski, M. (1991) Acute respiratory effects of prolonged ambient
  6             ozone. In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment:
  7             papers from an international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste
  8             Management Association; pp. 111-119. (A&WMA transactions series no. TR-19).
  9       Lee, J.-T.; Schwartz, J. (1999) Reanalysis of the effects of air pollution on daily mortality in Seoul, Korea:
10             a case-crossover design. Environ. Health Perspect. 107: 633-636.
11       Lee, J.-T.; Shin, D.; Chung, Y. (1999) Air pollution and daily mortality in Seoul and Ulsan, Korea. Environ. Health
12             Perspect. 107:  149-154.
13       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
14             and asthma among children in Seoul, Korea. Epidemiology 13: 481-484.
15       Lee, K.; Parkhurst, W. J.; Xue, J.; Ozkaynak, H.; Neuberg, D.; Spengler, J. D. (2004) Outdoor/indoor/presonal ozone
16             exposures of children in Nashville, Tennessee. J. Air Waste Manage. Assoc.  54: 352-359.
17       Le Tertre, A.; Quenel, P.; Eilstein, D.; Medina, S.; Prouvost, H.; Pascal, L.; Boumghar, A.; Saviuc, P.; Zeghnoun,
18             A.; Filleul, L.; Declercq, C.; Cassadou, S.; Le Coaster, C. (2002a) Short-term effects of air pollution on
19             mortality in nine French cities: a quantitative summary. Arch. Environ. Health 57: 311-319.
20       Le Tertre, A.; Medina, S.; Samoli, E.; Forsberg, B.; Michelozzi, P.; Boumghar, A.; Vonk, J. M.; Bellini, A.;
21             Atkinson, R.; Ayres, J. G.;  Sunyer, J.;  Schwartz, J.; Katsouyanni, K. (2002b) Short term effects of paniculate
22             air pollution on cardiovascular diseases in eight European cities. J. Epidemiol. Community Health
23             56:773-779.
24       Levy, J. I.; Carrothers, T. J.; Tuomisto, J.  T.;  Hammitt, J. K.; Evans, J. S. (2001) Assessing the public health benefits
25             of reduced ozone concentrations. Environ. Health Perspect. 109: 1215-1226.
26       Levy, J. I.; Chemerynski, S. M.; Sarnat, J. A. (2005) Ozone exposure and mortality, an empiric Bayes
27             metaregression analysis. Epidemiology 16: 458-468.
28       Liao, D.; Duan, Y.; Whitsel, E. A.; Zheng, Z.-J.; Heiss, G.; Chinchilli, V. M.; Lin, H.-M. (2004) Association of
29             higher levels of ambient criteria pollutants with impaired cardiac autonomic control: a population-based
30             study. Am. J. Epidemiol. 159: 768-777.
31       Lierl, M. B.; Hornung, R. W. (2003) Relationship of outdoor air quality to pediatric asthma exacerbations. Ann.
32             Allergy Asthma Immunol. 90: 28-33.
33       Lin, C. A.; Martins, M. A.; Farhat, S. C. L.; Pope, C. A., Ill; Conceicao, G. M. S.; Anastacio,  V. M.; Hatanaka, M.;
34             Andrade, W. C.; Hamaue, W. R.; Bohm, G. M.; Saldiva, P. H. N. (1999) Air pollution and respiratory illness
35             of children in Sao Paulo, Brazil. Paediatr. Perinat. Epidemiol. 13: 475-488.
36       Lin, M.-C; Yu, H.-S.; Tsai, S.-S.; Cheng, B.-H.; Hsu, T.-Y.; Wu, T.-N.; Yang, C.-Y. (2001) Adverse pregnancy
37             outcome in a petrochemical polluted area in Taiwan. J. Toxicol. Environ. Health Part A 63: 565-574.
3 8       Lin, M.; Chen, Y.; Burnett, R. T.;  Villeneuve, P. J.; Krewski, D. (2003) Effect of short-term exposure to gaseous
3 9             pollution on asthma hospitalisation in children: a bi-directional case-crossover analysis. J. Epidemiol.
40             Community Health 57: 50-55.
41       Lin, M.; Chen, Y.; Villeneuve, P. J.; Burnett, R. T.; Lemyre, L.; Hertzman, C.; McGrail, K. M.; Krewski, D. (2004)
42             Gaseous air pollutants and asthma hospitalization of children with low household income in Vancouver,
43             British Columbia, Canada. Am. J. Epidemiol. 159: 294-303.
44       Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996)
45             Short-term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal.
46             Environ. Epidemiol. 6: 449-472.
47       Linn, W. S.; Szlachcic, Y.; Gong, H., Jr.; Kinney, P. L.; Berhane, K. T. (2000) Air pollution and daily hospital
48             admissions in metropolitan Los Angeles. Environ. Health Perspect. 108: 427-434.
49       Lipfert, F. W.; Hammerstrom, T. (1992) Temporal patterns in air pollution and hospital admissions. Environ. Res.
50             59: 374-399.
51       Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000a) Daily mortality in the Philadelphia metropolitan area and
52             size-classified paniculate matter. J. Air Waste Manage. Assoc. 50: 1501-1513.
53       Lipfert, F. W.; Perry, H. M., Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2000b) The Washington
54             University-EPRI veterans' cohort mortality study: preliminary results. In: Grant, L. D.,  ed. PM2000:
55             paniculate matter and health. Inhalation Toxicol. 12(suppl. 4): 41-73.
         August 2005                                    7-161         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Lipfert, F. W.; Perry, H. M, Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2003) Air pollution, blood
  2             pressure, and their long-term associations with mortality. Inhalation Toxicol. 15: 493-512.
  3       Lippmann, M. (1988) Health significance of pulmonary function responses to airborne irritants. JAPCA
  4             38:881-887.
  5       Lippmann, M.; Spektor, D. M. (1998) Peak flow rate changes in O3 exposed children: spirometry vs miniWright
  6             flow meters. J. Exposure Anal. Environ. Epidemiol. 8: 101-107.
  7       Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
  8             mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no.
  9             95.
10       Lipsett, M.; Hurley, S.; Ostro, B. (1997) Air pollution and emergency room visits for asthma in Santa Clara County,
11             California. Environ. Health Perspect.  105: 216-222.
12       Liu, L.-J. S.; Koutrakis, P.; Leech, J.; Broder, I. (1995) Assessment of ozone exposures in the greater metropolitan
13             Toronto area. J.  Air Waste Manage. Assoc. 45: 223-234.
14       Loomis, D. P.; Borja-Aburto, V. H.; Bangdiwala, S. I.; Shy, C. M. (1996) Ozone exposure and daily mortality in
15             Mexico City: a time-series analysis. Cambridge, MA: Health Effects Institute; research report no. 75.
16       Loomis, D.; Castillejos, M.; Gold, D. R.; McDonnell, W.; Borja-Aburto, V. H. (1999) Air pollution and infant
17             mortality in Mexico City. Epidemiology  10: 118-123.
18       Luginaah, I. N.; Fung,  K. Y.; Gorey, K. M.; Webster, G.; Wills, C. (2005) Association of Ambient Air Pollution
19             with Respiratory Hospitalization in a Government Designated  "Area of Concern": The Case of Windsor,
20             Ontario. Environ. Health Perspect.  113: 290-296.
21       Lumley, T.; Sheppard,  L. (2000) Assessing seasonal confounding and model selection bias in air pollution
22             epidemiology using positive and negative control analyses. Environmetrics 11: 705-717.
23       Maisonet, M.; Bush, T. J.; Correa, A.; Jaakkola, J. J. K. (2001) Relation between ambient air pollution and low birth
24             weight in the northeastern United States.  Environ. Health Perspect. Suppl.  109(3): 351-356.
25       Mann, J. K.; Tager, I. B.; Lurmann, F.; Segal, M.; Quesenberry, C. P., Jr.; Lugg, M. M.; Shan, J.; Van denEeden,
26             S. K. (2002) Air pollution and hospital admissions for ischemic heart disease in persons with congestive heart
27             failure or arrhythmia. Environ. Health Perspect. 110: 1247-1252.
28       Martins, L. C.; Latorre, M. R. D. O.; Saldiva, P. H. N.; Braga, A. L. F. (2002) Air pollution and emergency room
29             visits due to chronic lower respiratory diseases in the elderly: an ecological time-series study in Sao Paulo,
30             Brazil. J. Occup. Environ. Med. 44: 622-627.
31       McConnell, R.;  Berhane, K.; Gilliland, F.; London, S. J.; Vora, H.; Avol, E.; Gauderman, W. J.; Margolis, H. G.;
32             Lurmann, F.; Thomas, D. C.; Peters, J. M. (1999) Air pollution and bronchitic symptoms in southern
33             California children with asthma. Environ. Health Perspect. 107: 757-760.
34       McConnell, R.;  Berhane, K.; Gilliland, F.; London, S. J.; Islam, T.; Gauderman, W. J.; Avol, E.; Margolis, H. G.;
35             Peters, J. M. (2002) Asthma in exercising children exposed to ozone: a cohort study. Lancet 359: 386-391.
36       McConnell, R.;  Berhane, K.; Gilliland, F.; Molitor, J.; Thomas, D.; Lurmann, F.; Avol, E.; Gauderman, W. J.;
37             Peters, J. M. (2003) Prospective study of air pollution and bronchitic symptoms in children with asthma.
38             Am. J. Respir. Crit. Care Med. 168:790-797.
39       McDonnell, W.  F.; Abbey, D. E.; Nishino, N.; Lebowitz, M. D. (1999)  Long-term ambient ozone concentration and
40             the incidence of asthma in nonsmoking adults: the ahsmog study. Environ.  Res. 80:  110-121.
41       Metzger, K. B.;  Tolbert, P. E.; Klein, M.; Peel, J. L.; Flanders, W. D.; Todd, K. H.; Mulholland, J. A.; Ryan, P. B.;
42             Frumkin, H. (2004) Ambient air pollution and cardiovascular emergency department visits. Epidemiology
43             15:46-56.
44       Mitchell, H.; Senturia,  Y.; Gergen, P.; Baker, D.; Joseph, C.; McNiff-Mortimer, K.; Wedner, H. J.; Crain, E.;
45             Eggleston, P.; Evans, R., Ill; Kattan, M.; Kercsmar, C.; Leickly,  F.; Malveaux, F.; Smartt, E.; Weiss, K.
46             (1997) Design and methods of the National Cooperative Inner-City Asthma Study. Pediatr. Pulmonol.
47             24:237-252.
48       Moolgavkar, S.  H. (2003) Air pollution and daily mortality in two U.S.  counties: season-specific analyses and
49             exposure-response relationships. Inhalation Toxicol. 15: 877-907.
50       Moolgavkar, S.  H.; Luebeck, E. G. (1996) A critical review of the evidence on paniculate air pollution and
51             mortality. Epidemiology 7: 420-428.
52       Moolgavkar, S.  H.; Luebeck, E. G.; Hall, T.  A.; Anderson, E. L. (1995) Air pollution and daily mortality in
53             Philadelphia. Epidemiology 6: 476-484.
54       Moolgavkar, S.  H.; Luebeck, E. G.; Anderson, E. L. (1997) Air pollution and hospital admissions for respiratory
55             causes in Minneapolis-St. Paul and Birmingham. Epidemiology 8: 364-370.
         August 2005                                    7-162         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Morgan, G.; Corbett, S.; Wlodarczyk, J. (1998a) Air pollution and hospital admissions in Sydney, Australia, 1990 to
  2             1994. Am. J. Public Health 88: 1761-1766.
  3       Morgan, G.; Corbett, S.; Wlodarczyk, J.; Lewis, P. (1998b) Air pollution and daily mortality in Sydney, Australia,
  4             1989 through 1993. Am. J. Public Health 88: 759-764.
  5       Mortimer, K. M.; Tager, I. B.; Dockery, D. W.; Neas, L. M.; Redline, S. (2000) The effect of ozone on inner-city
  6             children with asthma: identification of susceptible subgroups. Am. J. Respir. Crit. Care Med. 162: 1838-1845.
  7       Mortimer, K. M.; Neas, L. M.; Dockery, D. W.; Redline, S.; Tager, I. B. (2002) The effect of air pollution on
  8             inner-city children with asthma. Eur. Respir. J. 19: 699-705.
  9       Naeher, L. P.; Holford, T.  R.; Beckett, W. S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
10             Healthy women's PEF variations with ambient summer concentrations of PM10, PN2 5, SO42", H+, and O3. Am.
11             J. Respir. Crit. Care Med.  160: 117-125.
12       National Institutes of Health. (1997) Guidelines for the diagnosis and management of asthma: expert panel report 2.
13             Bethesda, MD: U.S. Department of Health and Human Services, National Heart, Lung, and Blood Institute;
14             publication no. 97-4051. Available: http://www.nhlbi.nih.gov/guidelines/asthma/asthgdln.pdf (11 April 2003).
15       Nauenberg, E.; Basu, K. (1999) Effect of insurance coverage on the relationship between asthma hospitalizations
16             and exposure to air pollution. Public Health Rep.  114: 135-148.
17       Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
18             effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
19       Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
20             pollution with twice daily  peak expiratory flow rate measurements in children. Am. J. Epidemiol.
21             141: 111-122.
22       Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Speizer, F. E. (1999) Fine particles and peak flow in children: acidity
23             versus mass. Epidemiology 10: 550-553.
24       Neidell, M. J. (2004) Air pollution, health, and socio-economic status: the effect of outdoor air quality on childhood
25             asthma. J. Health Econ. 23: 1209-1236.
26       Newhouse, C. P.; Levetin, B. S.;  Levetin, E. (2004) Correlation of environmental factors with asthma and rhinitis
27             symptoms in Tulsa, OK. Ann. Allergy Asthma Immunol.  92: 356-366.
28       Nutman, A.; Solomon, Y.; Mendel, S.; Nutman, J.; Hines, E.; Topilsky, M.; Kivity, S. (1998) The use of a neural
29             network for studying the relationship between air pollution and asthma-related emergency room visits. Respir.
30             Med.  92: 1199-1202.
31       Oftedal, B.; Nafstad, P.; Magnus, P.; Bjerkly, S.; Skrondal, A. (2003) Traffic related air pollution and acute hospital
32             admission for respiratory diseases inDrammen, Norway 1995-2000.  Eur. J. Epidemiol. 18: 671-675.
33       O'Neill, M. S.; Loomis, D.; Borja-Aburto, V. H. (2004) Ozone, area social conditions, and mortality in Mexico City.
34             Environ. Res. 94: 234-242.
3 5       Ostro, B.  (1995) Fine paniculate  air pollution and mortality in two Southern California counties. Environ. Res.
36             70:98-104.
37       Ostro, B.; Sanchez, J. M.;  Aranda, C.; Eskeland, G. S. (1996) Air pollution and mortality: results from a study of
38             Santiago, Chile. In: Lippmann, M., ed. Papers from the ISEA-ISEE annual meeting; September 1994;
39             Research Triangle Park, NC.  J. Exposure Anal. Environ. Epidemiol. 6: 97-114.
40       Ostro, B.; Lipsett, M.; Mann, J.; Braxton-Owens, H.; White, M.  (2001) Air pollution and exacerbation of asthma in
41             African-American children in Los Angeles. Epidemiology 12: 200-208.
42       Palli, D.;  Saieva, C.; Grechi, D.; Masala, G.; Zanna, L; Barbara, A.; Decarli, A.; Munnia, A.; Peluso, M. (2004)
43             DNA bulky adducts in a Mediterranean population correlate with environmental ozone concentration, an
44             indicator of photochemical smog. Int. J. Cancer 109: 17-23.
45       Park, H.;  Lee, B.; Ha, E.-H.; Lee, J.-T.; Kim, H.; Hong, Y.-C. (2002) Association of air pollution with school
46             absenteeism due to  illness. Arch. Pediatr. Adolesc. Med. 156:  1235-1239.
47       Park, S. K.; O'Neill, M. S.; Vokonas, P. S.; Sparrow, D.; Schwartz, J. (2005) Effects of air pollution on heart rate
48             variability: the VA normative aging study. Environ. Health Perspect. 113: 304-309.
49       Pearce, N.; Beasley, R.; Burgess, C.; Crane, J. (1998) Asthma epidemiology: principles and methods. New York,
50             NY: Oxford University Press.
51       Peel, J. L.; Tolbert, P. E.; Klein, M.; Metzger, K. B.; Flanders, W. D.; Knox, T.; Mulholland, J. A.; Ryan, P. B.;
52             Frumkin, H. (2005) Ambient air pollution and respiratory emergency department visits. Epidemiology
53             16: 164-174.
54       Pereira, L. A. A.; Loomis, D.; Conceicao, G. M. S.; Braga, A. L. F.; Areas, R. M.; Kishi, H. S.; Singer, J. M.; Bohm,
55             G. M.; Saldiva, P. H. N. (1998)  Association between air pollution and intrauterine  mortality in Sao Paulo,
56             Brazil. Environ. Health Perspect. 106:  325-329.
         August 2005                                    7-163         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Pereira, F. A. C.; De Assuncao, J. V.; Saldiva, P. H. N.; Pereira, L. A. A.; Mirra, A. P.; Braga, A. L. F. (2005)
  2             Influence of air pollution on the incidence of respiratory tract neoplasm. J. Air Waste Manage. Assoc.
  3             55:83-87.
  4       Peters, J. M; Avol, E.; Navidi, W.; London, S. J.; Gauderman, W. J.; Lurmann, F.; Linn, W. S.; Margolis, H.;
  5             Rappaport, E.; Gong, H., Jr.; Thomas, D. C. (1999a) A study of twelve southern California communities with
  6             differing levels and types of air pollution. I. Prevalence of respiratory morbidity. Am. J. Respir. Crit.  Care
  7             Med. 159: 760-767.
  8       Peters, J. M.; Avol, E.; Gauderman, W. J.; Linn, W. S.; Navidi, W.; London, S.  J.; Margolis, H.;  Rappaport, E.;
  9             Vora, H.; Gong, H., Jr.; Thomas, D. C. (1999b) A study of twelve southern California communities with
10             differing levels and types of air pollution. II. Effects on pulmonary function. Am. J. Respir. Crit. Care Med.
11             159:768-775.
12       Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M.; Baliff, J.; Oh, J. A.;  Allen, G.;
13             Monahan, K.; Dockery, D. W. (2000a) Air pollution and incidence of cardiac arrhythmia.  Epidemiology
14             11: 11-17.
15       Peters, A.; Skorkovsky, J.; Kotesovec, F.; Brynda, J.; Spix, C.; Wichmann, H. E.; Heinrich, J. (2000b) Associations
16             between mortality and air pollution in central Europe. Environ. Health Perspect. 108: 283-287.
17       Peters, A.; Dockery, D.  W.; Muller, J. E.; Mittleman, M. A. (2001) Increased paniculate air pollution and the
18             triggering of myocardial infarction. Circulation 103: 2810-2815.
19       Petroeschevsky, A.; Simpson, R. W.; Thalib, L.; Rutherford,  S. (2001) Associations between outdoor air pollution
20             and hospital admissions in Brisbane, Australia. Arch. Environ. Health 56: 37-52.
21       Plunkett, L. M.; Turnbull, D.; Rodricks, J. V. (1992) Differences between adults and children affecting exposure
22             assessment. In: Guzelian, P. S.; Henry, D. J.; Olin, S. S., eds. Similarities and differences between children
23             and adults: implications for risk assessment. Washington, DC: ILSI Press, pp. 79-96.
24       Poloniecki, J. D.; Atkinson, R. W.; Ponce de Leon, A.;  Anderson, H. R. (1997) Daily time series for cardiovascular
25             hospital admissions and previous day's air pollution in London, UK. Occup. Environ. Med. 54: 535-540.
26       Ponce de Leon, A.; Anderson, H. R.; Bland, J. M.; Strachan, D. P.; Bower, J. (1996) Effects of air pollution on daily
27             hospital admissions for respiratory disease in London between 1987-88 and 1991-92. In: StLeger, S., ed.
28             The APHEA project.  Short term effects of air pollution on health: a European approach using epidemiological
29             time series data. J. Epidemiol. Commun. Health  50(suppl. 1): S63-S70.
30       Ponka, A.; Virtanen, M. (1996) Asthma and ambient air pollution in Helsinki. In: St Leger, S., ed. The APHEA
31             project. Short term effects of air pollution on health: a European approach using epidemiological time series
32             data. J. Epidemiol. Community Health 50(suppl. 1): S59-S62.
33       Ponka, A.; Savela, M.; Virtanen, M. (1998) Mortality and air pollution in Helsinki. Arch. Environ. Health 53:
34             281-286.
35       Pope, C. A., Ill; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
36             cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA J. Am. Med.
37             Assoc. 287: 1132-1141.
38       Prescott,  G. J.; Cohen, G. R.; Elton, R. A.; Fowkes, F. G. R.; Agius, R. M. (1998)  Urban air pollution and
39             cardiopulmonary ill health: a 14.5 year time series study. Occup.  Environ. Med. 55: 697-704.
40       Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. (1987) Acute respiratory function and transported air pollutants:
41             observational studies. Presented at: 80th annual meeting of the Air Pollution Control Association; June;
42             New York, NY. Pittsburgh, PA: Air Pollution Control Association; paper no. 87-32.6.
43       Raizenne, M. E.; Burnett, R. T.; Stern, B.; Franklin, C.  A.; Spengler, J. D. (1989) Acute lung function responses to
44             ambient acid aerosol exposures in children. Environ. Health Perspect. 79: 179-185.
45       Ramadour, M.; Burel, C.; Lanteaume, A.; Vervloet, D.; Charpin, D.; Brisse,  F.;  Dutau, H.; Charpin, D. (2000)
46             Prevalence of asthma and rhinitis in relation to long-term exposure to gaseous air pollutants. Allergy
47             (Copenhagen)  55: 1163-1169.
48       Ramsay, T. O.; Burnett, R. T.; Krewski, D. (2003) The effect of concurvity in generalized additive models linking
49             mortality to ambient paniculate matter. Epidemiology 14: 18-23.
50       Rich, K. E.; Petkau, J.; Vedal, S.;Brauer,M. (2004) A  case-crossover analysis of paniculate air pollution and
51             cardiac arrhythmia in patients with implantable cardioverter defibrillators. Inhalation Toxicol.  16: 363-372.
52       Rich, D. Q.; Schwartz, J.; Mittleman, M. A.; Link, M.;  Luttmann-Gibson, H.; Catalano, P. J.; Speizer, F. E.;
53             Dockery, D. W. (2005) Association of short-term ambient air pollution concentrations and ventricular
54             arrhythmias. Am. J. Epidemiol. 161: 1123-1132.
5 5       Ritz, B.; Yu, F. (1999) The effect of ambient carbon monoxide on low birth weight among children born in southern
56             California between 1989 and 1993. Environ. Health Perspect. 107: 17-25.
         August 2005                                   7-164         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Ritz, B.; Yu, F.; Chapa, G.; Fruin, S. (2000) Effect of air pollution on preterm birth among children born in Southern
  2             California between 1989 and 1993. Epidemiology 11: 502-511.
  3       Ritz, B.; Yu, F.; Fruin, S.; Chapa, G.; Shaw, G. M; Harris, J. A. (2002) Ambient air pollution and risk of birth
  4             defects in Southern California. Am. J. Epidemiol. 155: 17-25.
  5       Roemer, W. H.; Van Wijnen, J. H . (2001) Daily mortality and air pollution along busy streets in Amsterdam,
  6             1987-1998. Epidemiology 12: 649-653.
  7       Romieu, I.; Meneses, F.; Sienra-Monge, J. J. L.; Huerta, J.; Velasco, S. R.; White, M. C.; Etzel, R.  A.;
  8             Hernandez-Avila, M. (1995) Effects of urban air pollutants on emergency visits for childhood asthma in
  9             Mexico City. Am. J. Epidemiol.  141: 546-553.
10       Romieu, L; Meneses, F.; Ruiz, S.; Sienra, J. J.; Huerta, J.; White, M. C.; Etzel, R. A. (1996) Effects of air pollution
11             on the respiratory health of asthmatic children living in Mexico City. Am. J.  Respir. Crit. Care Med.
12             154:300-307.
13       Romieu, L; Meneses, F.; Ruiz, S.; Huerta, J.; Sienra,  J. J.; White, M.; Etzel, R.; Hernandez, M. (1997) Effects of
14             intermittent ozone exposure on peak expiratory flow and respiratory symptoms among asthmatic children in
15             Mexico City. Arch. Environ. Health 52: 368-376.
16       Romieu, L; Meneses, F.; Ramirez, M.; Ruiz, S.; Padilla, R. P.; Sienra, J. J.; Gerber, M.; Grievink, L.; Dekker, R.;
17             Walda, L; Brunekreef, B. (1998) Antioxidant supplementation and respiratory functions among workers
18             exposed to high levels of ozone.  Am. J. Respir. Crit. Care Med.  158: 226-232.
19       Romieu, L; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Tellez-Rojo, M. M.; Moreno-Macias, H.; Reyes-Ruiz, N. L;
20             Del Rio-Navarro, B. E.; Ruiz-Navarro, M.  X.; Hatch, G.; Slade, R.; Hernandez-Avila, M. (2002) Antioxidant
21             supplementation and lung functions among children with asthma exposed to  high levels of air pollutants. Am.
22             J. Respir. Crit. Care Med. 166: 703-709.
23       Romieu, L; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. L; Estela del
24             Rio-Navarro, B.; Hernandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and
25             antioxidant supplementation influence lung function in relation to ozone exposure in asthmatic children in
26             Mexico City. Thorax 59: 8-10.
27       Rona, R. J.; Gulliford, M. C.; Chinn, S. (1993) Effects of prematurity and intrauterine growth on respiratory health
28             and lung function in childhood. Br. Med. J. 306: 817-820.
29       Ross, M. A.; Persky, V. W.; Scheff, P. A.; Chung, J.; Curtis, L.; Ramakrishnan, V.; Wadden, R. A.; Hryhorczuk,
30             D. O. (2002) Effect of ozone and aeroallergens on the respiratory health of asthmatics. Arch. Environ. Health
31             57:568-578.
32       Ruidavets, J.-B.; Cournot, M.; Cassadou, S.; Giroux, M.; Meybeck, M.; Ferrieres, J. (2005) Ozone  air pollution is
33             associated with acute myocardial infarction. Circulation 111: 563-569.
34       Saez, M.; Tobias, A.; Munoz, P.; Campbell, M. J. (1999) A GEE  moving average analysis of the relationship
3 5             between air pollution and mortality for asthma in Barcelona, Spain. Stat. Med.  18: 2077-2086.
36       Saez, M.; Ballester, F.; Barcelo, M. A.; Perez-Hoyos, S.; Bellido, J.; Tenias, J. M.; Ocafia, R.; Figueiras, A.;
37             Arribas, F.; Aragones, N.; Tobias, A.; Cirera, L.;  Canada, A.; on behalf of the EMECAM Group. (2002) A
3 8             combined analysis of the short-term effects of photochemical air pollutants on mortality within the EMECAM
39             project. Environ. HealthPerspect. 110: 221-228.
40       Saldiva, P. H. N.; Lichtenfels, A. J. F. C.; Paiva, P. S. O.; Barone, I. A.; Martins, M. A.; Massad, E.; Pereira, J.  C.
41             R.; Xavier, V. P.; Singer, J. M.; Bohm, G. M. (1994) Association between air pollution and mortality due to
42             respiratory diseases in children in Sao Paulo, Brazil: a preliminary report. Environ. Res. 65:  218-225.
43       Saldiva, P. H. N.; Pope, C. A., Ill; Schwartz, J.; Dockery, D. W.;  Lichtenfels, A. J.; Salge, J. M.; Barone, I.; Bohm,
44             G. M. (1995) Air pollution and mortality in elderly people: a time-series study in Sao Paulo, Brazil. Arch.
45             Environ. Health 50: 159-163.
46       Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, L; Dockery, D. W.;  Schwartz, J.; Zanobetti, A.
47             (2000) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
48             pollution in the United States. Cambridge,  MA: Health Effects Institute; research report no. 94, part II.
49       Sarnat, J. A.; Schwartz, J.; Catalano, P. J.; Suh, H. H. (2001) Gaseous pollutants in  paniculate matter epidemiology:
50             confounders or surrogates? Environ. Health Perspect.  109: 1053-1061.
51       Sartor, F.; Snacken, R.; Demuth, C.; Walckiers, D. (1995) Temperature, ambient ozone levels, and  mortality during
52             summer, 1994, in Belgium. Environ. Res. 70: 105-113.
53       Scarlett, J. F.; Abbott, K. J.; Peacock, J. L.; Strachan, D. P.; Anderson, H. R. (1996) Acute effects of summer air
54             pollution on respiratory function in primary school children in southern England. Thorax 51:1109-1114.
55       Schildcrout, J. S.; Heagerty, P. J. (2005) Regressions analysis of longitudinal binary data with time-dependent
56             environmental covariates: bias and efficiency. Biostatistics: doi: 10.1093/biostatistics/kxi033.
         August 2005                                   7-165         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Schindler, C.; Ktinzli, N.; Bongard, J.-P.; Leuenberger, P.; Karrer, W.; Rapp, R.; Monn, C.; Ackermann-Liebrich,
  2             U.; Swiss Study on Air Pollution and Lung Diseases in Adults Investigators. (2001) Short-term variation in
  3             air pollution and in average lung function among never-smokers. Am. J. Respir. Crit. Care Med.
  4             163:356-361.
  5       Schouten, J. P.; Vonk, J.  M; de Graaf, A. (1996) Short term effects of air pollution on emergency hospital
  6             admissions for respiratory disease: results of the APHEA project in two major cities in The Netherlands,
  7             1977-89. In: St Leger,  S., ed. The APHEA project. Short term effects of air pollution on health: a European
  8             approach using epidemiological time series data. J. Epidemiol.  Community Health 50(suppl. 1): S22-S29.
  9       Schwartz, J. (1991) Paniculate air pollution and daily mortality in Detroit. Environ. Res. 56: 204-213.
10       Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease. Epidemiology 7: 20-28.
11       Schwartz, J. (2005) How sensitive is the  association between ozone and daily deaths to control for temperature?
12             Am. J. Respir. Crit. Care Med. 171: 627-631.
13       Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.; Ponce de
14             Leon, A.; Ponka, A.; Rossi, G.; Saez, M.; Schouten, J. P. (1996) Methodological issues in studies  of air
15             pollution and daily counts of deaths or hospital admissions. In: St Leger, S., ed. The APHEA project.
16             Short term effects of air pollution on health: a European approach using epidemiological time series data.
17             J. Epidemiol. Commun. Health 50(suppl. 1): S3-SI 1.
18       Schwartz, J.; Litonjua, A.; Suh, H.; Verrier, M.; Zanobetti, A.; Syring, M.; Nearing, B.; Verrier, R.; Stone, P.;
19             MacCallum, G.; Speizer, F. E.; Gold, D. R. (2005) Traffic  related pollution and heart rate variability in a
20             panel of elderly subjects. Thorax 60: 455-461.
21       Selwyn, B. J.; Stock, T. H.; Hardy, R. J.; Chan, F. A.;  Jenkins, D. E.; Kotchmar, D. J.; Chapman, R. S. (1985)
22             Health effects of ambient ozone exposure in vigorously exercising adults. In: Lee, S. D., ed. Evaluation of the
23             scientific basis for ozone/oxidants standards: proceedings of an APCA international specialty conference;
24             November 1984; Houston, TX. Pittsburgh, PA: Air Pollution Control Association; pp. 281-296. (APCA
25             international specialty  conference transactions:  TR-4).
26       Sheppard, L. (2003) Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington,
27             1987-1994. In: Revised analyses of time-series  studies of air pollution and health. Special report. Boston,
28             MA: Health Effects Institute; pp. 227-230. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
29             [18 October, 2004].
3 0       Sheppard, L. (2005) Acute air pollution effects: consequences of exposure distribution and measurements. J.
31             Toxicol. Environ.  Health Part A 68: 1127-1135.
32       Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air pollution on
33             nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
34       Sheppard, L.; Slaughter,  J. C.; Schildcrout, J.; Liu, L.-J. S.; Lumley, T. (2005) Exposure and measurement
35             contributions to estimates of acute air pollution effects. J. Exposure Anal. Environ. Epidemiol. 15: 366-376.
36       Sherwin, R. P.; Richters, V.; Kraft, P.; Richters, A. (2000) Centriacinar region inflammatory disease in young
37             individuals: a comparative study of Miami and Los Angeles residents. Virchows Arch. 437: 422-428.
38       Shumway, R. H.; Azari, A. S.; Pawitan, Y. (1988) Modeling mortality fluctuations in Los Angeles as functions of
39             pollution and weather effects. Environ. Res. 45: 224-241.
40       Silverman, B. W. (1986) Density estimation. London,  England: Chapman & Hall.
41       Simpson, R. W.; Williams, G.; Petroeschevsky, A.; Morgan, G.; Rutherford, S. (1997) Associations between outdoor
42             air pollution and daily  mortality in Brisbane, Australia. Arch. Environ. Health 52: 442-454.
43       Spektor, D. M.; Lippmann, M.; Lioy, P. J.; Thurston, G. D.; Citak, K.; James, D. J.; Bock, N.; Speizer, F. E.;
44             Hayes, C. (1988a) Effects of ambient ozone on respiratory function in active, normal children. Am. Rev.
45             Respir. Dis. 137: 313-320.
46       Spektor, D. M.; Lippmann, M.; Thurston, G. D.; Lioy, P. J.; Stecko, J.; O'Connor, G.; Garshick, E.; Speizer, F. E.;
47             Hayes, C. (1988b) Effects of ambient ozone on respiratory function in healthy adults exercising outdoors.
48             Am. Rev. Respir. Dis.  138: 821-828.
49       Spektor, D. M.; Lippmann, M. (1991) Health effects of ambient ozone on healthy children at a summer camp. In:
50             Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
51             international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
52             Association; pp. 83-89. (A&WMA transactions series no. TR-19).
53       Stieb, D. M.; Burnett, R. T.; Beveridge, R. C.; Brook,  J. R. (1996) Association between ozone and asthma
54             emergency  department visits in Saint John, New Brunswick, Canada. Environ. Health Perspect.
55             104: 1354-1360.
         August 2005                                   7-166         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Stieb, D. M; Judek, S.; Burnett, R. T. (2002) Meta-analysis of time-series studies of air pollution and mortality:
  2             effects of gases and particles and the influence of cause of death, age, and season. J. Air Waste Manage.
  3             Assoc. 52: 470-484.
  4       Stieb, D. M.; Judek, S.; Burnett, R. T. (2003) Meta-analysis of time-series studies of air pollution and mortality:
  5             update in relation to the use of generalized additive models. J. Air Waste Manage. 53: 258-261.
  6       Sunyer, J.; Basagana, X. (2001) Particles, and not gases, are associated with the risk of death in patients with chronic
  7             obstructive pulmonary disease. Int. J. Epidemiol. 30: 1138-1140.
  8       Sunyer, J.; Castellsague, J.; Saez, M.; Tobias, A.; Anto, J. M. (1996) Air pollution and mortality in Barcelona. In:
  9             St Leger, S., ed. The APHEA project. Short term effects of air pollution on health: a European approach using
10             epidemiological time series data. J. Epidemiol. Community Health 50(suppl. 1): S76-S80.
11       Sunyer, J.; Spix, C.; Quenel, P.; Ponce-de-Leon, A.; Ponka, A.; Barumandzadeh, T.; Touloumi, G.; Bacharova, L.;
12             Wojtyniak, B.; Vonk, J.; Bisanti, L.; Schwartz, J.; Katsouyanni, K. (1997) Urban air pollution and emergency
13             admissions for asthma in four European cities: the APHEA project. Thorax 52: 760-765.
14       Sunyer, J.; Basagana, X.; Belmonte, J.; Anto, J. M. (2002) Effect of nitrogen dioxide and ozone on the risk of dying
15             in patients with severe asthma. Thorax 57: 687-693.
16       Tager, I. B. (1993) Introduction to working group on tropospheric ozone, Health Effects Institute Environmental
17             Epidemiology Planning Project. Environ. Health Perspect. 101(suppl. 4): 205-207.
18       Tager, I. B. (1999) Air pollution and lung function growth. Is it ozone? [editorial]. Am. J. Respir. Crit. Care Med.
19             160:387-389.
20       Tager, I. B.; Kunzli, N.; Lurmann, F.; Ngo, L.; Segal, M.; Balmes, J. (1998) Methods development for
21             epidemiologic investigations of the health effects of prolonged ozone exposure. Part II:  an approach to
22             retrospective estimation of lifetime ozone exposure using a questionnaire and ambient monitoring data
23             (California sites). Cambridge, MA: Health Effects Institute; research report no. 81; pp. 27-78.
24       Taggart, S. C. O.; Custovic, A.; Francis, H. C.; Faragher, E. B.; Yates, C. J.; Higgins, B. G.; Woodcock, A. (1996)
25             Asthmatic bronchial hyperresponsiveness varies with ambient levels of summertime air pollution. Eur.
26             Respir. J. 9: 1146-1154.
27       Tellez-Rojo, M. M.; Romieu, L; Ruiz-Velasco, S.; Lezana, M.-A.; Hernandez-Avila, M. M. (2000) Daily respiratory
28             mortality and PMio pollution in Mexico City: importance of considering place of death. Eur. Respir. J.
29             16:391-396.
30       Tenias, J. M.; Ballester, F.; Rivera, M. L. (1998) Association between hospital emergency visits for asthma and air
31             pollution in Valencia, Spain. Occup. Environ. Med. 55: 541-547.
32       Tenias, J. M.; Ballester, F.; Perez-Hoyos, S.; Rivera, M. L. (2002) Air pollution and hospital emergency room
33             admissions for chronic obstructive pulmonary disease in Valencia, Spain. Arch. Environ. Health 57: 41-47.
34       Thompson, A. J.; Shields, M. D.; Patterson, C. C. (2001) Acute asthma exacerbations and air pollutants in children
3 5             living in Belfast, Northern Ireland. Arch. Environ. Health 56: 234-241.
36       Thurston, G. D.; Ito, K. (2001) Epidemiological studies of acute ozone exposures and mortality.  J. Exposure Anal.
37             Environ. Epidemiol.  11: 286-294.
38       Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
39             hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers.
40             J. Exposure Anal. Environ.  Epidemiol. 2: 429-450.
41       Thurston, G. D.; Ito, K.; Hayes, C. G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital admissions and
42             summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ. Res.
43             65:271-290.
44       Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. (1997)  Summertime haze air pollution and children with
45             asthma. Am. J. Respir. Crit. Care Med. 155:  654-660.
46       Tobias, A.; Campbell, M. J.; Saez, M. (1999) Modelling asthma epidemics on the relationship between air pollution
47             and asthma emergency visits in Barcelona, Spain. Eur. J. Epidemiol. 15: 799-803.
48       Tolbert, P. E.; Mulholland, J. A.; Macintosh, D. L.; Xu, F.; Daniels, D.; Devine, O. J.; Carlin, B. P.; Klein, M.;
49             Dorley, J.; Butler, A. J.; Nordenberg, D. F.; Frumkin, H.; Ryan, P. B.; White, M. C.  (2000) Air quality and
50             pediatric emergency room visits for asthma in Atlanta, Georgia. Am. J. Epidemiol. 151: 798-810.
51       Touloumi, G.; Katsouyanni, K.; Zmirou, D.; Schwartz, J.; Spix, C.; Ponce de Leon, A.; Tobias, A.; Quennel, P.;
52             Rabczenko, D.; Bacharova, L.; Bisanti, L.; Vonk, J. M.; Ponka, A. (1997) Short-term effects of ambient
53             oxidant exposure on mortality: a combined analysis within the APHEA project. Am. J. Epidemiol.
54             146: 177-185.
55       Tsai, S.-S.; Goggins, W. B.; Chiu, H.-F.; Yang, C.-Y. (2003a) Evidence for an association between air pollution and
56             daily stroke admissions inKaohsiung, Taiwan. Stroke  34: 2612-2616.
         August 2005                                    7-167        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Tsai, S.-S.; Huang, C.-H.; Goggins, W. B.; Wu, T.-N.; Yang, C.-Y. (2003b) Relationship between air pollution and
  2             daily mortality in a tropical city: Kaohsiung, Taiwan. J. Toxicol. Environ. Health Part A 66: 1341-1349.
  3       Tyler, W. S.; Tyler, N. K.; Last, J. A.; Gillespie, M. J.; Barstow, T. J. (1988) Comparison of daily and seasonal
  4             exposures of young monkeys to ozone. Toxicology 50: 131-144.
  5       U.S. Environmental Protection Agency.  (1996a) Air quality criteria for ozone and related photochemical oxidants.
  6             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
  7             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608.  Available:
  8             www.epa.gov/ncea/ozone.htm.
  9       U.S. Environmental Protection Agency.  (1996b) Air quality criteria for paniculate matter. Research Triangle Park,
10             NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
11       U.S. Environmental Protection Agency.  (2004) Air quality criteria for paniculate matter. Research Triangle Park,
12             NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
13             http://cfpub.epa.gov/ncea/ [9 November, 2004].
14       Ulmer, C.; Kopp, M.; Ihorst, G.; Frischer, T.; Forster, J.; Kuehr, J. (1997) Effects of ambient ozone exposures during
15             the spring and summer of 1994 on pulmonary function of schoolchildren. Pediatr. Pulmonol. 23: 344-353.
16       Vaughan, T. R.; Weber, R. W.; Tipton, W. R.; Nelson, H. S. (1989) Comparison of PEFR and FEVj in patients with
17             varying degrees of airway obstruction: effect of modest altitude. Chest 95: 558-562.
18       Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
19             pollution. Environ. HealthPerspect. Ill: 45-51.
20       Vedal, S.; Rich, K.; Brauer, M.; White, R.; Petkau, J.  (2004) Air pollution and cardiac arrhythmias in patients with
21             implantable cardiovascular defibrillators. Inhalation Toxicol. 16: 353-362.
22       Verhoeff, A. P.; Hoek, G.; Schwartz, J.; Van Wijnen,  J.  H. (1996) Air pollution and daily mortality in Amsterdam.
23             Epidemiology 7: 225-230.
24       Virnig, B.  A.; McBean, M. (2001) Administrative data for public health surveillance and planning. Annu. Rev.
25             Public Health 22: 213-230.
26       Wang, X.; Ding, H.; Ryan, L.; Xu, X. (1997) Association between air pollution and low birth weight: a
27             community-based study. Environ. Health Perspect. 105: 514-520.
28       Ward, D. J.; Roberts, K. T.; Jones, N.; Harrison, R. M.;  Ayres, J. G.; Hussain, S.; Walters, S.  (2002) Effects of daily
29             variation in outdoor particulates and ambient acid species in normal and asthmatic children. Thorax
30             57:489-502.
31       Weisel, C. P.; Cody, R. P.; Lioy, P. J. (1995) Relationship between summertime ambient ozone levels and
32             emergency department visits for asthma in central New Jersey. Environ. Health Perspect.
33             103(suppl. 2): 97-102.
34       Weisel, C. P.; Cody, R. P.; Georgopoulos, P. G.; Purushothaman, V; Weiss, S. H.; Bielory, L.; Gregory, P.; Stern,
35             A. H. (2002) Concepts in developing  health-based indicators for ozone. Int. Arch. Occup. Environ. Health
36             75:415-422.
37       White, M.  C.; Etzel, R. A.; Wilcox, W. D.; Lloyd, C.  (1994) Exacerbations of childhood asthma and ozone pollution
38             in Atlanta. Environ. Res. 65: 56-68.
39       Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. (199la) Activity
40             patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
41             ARB/R93/487. Available from: NTIS, Springfield, VA.; PB94-108719.
42       Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b) Study of children's activity
43             patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
44       Wilson, A. M.; Wake, C. P.; Kelly, T.; Salloway, J. C. (2005) Air pollution, weather, and respiratory emergency
45             room visits in two northern New England cities: an ecological time-series study. Environ. Res. 97: 312-321.
46       Wong, T. W.; Lau, T. S.; Yu, T. S.; Neller, A.; Wong, S. L.; Tarn, W.; Pang, S. W. (1999a) Air pollution and
47             hospital admissions for respiratory and cardiovascular diseases in Hong Kong. Occup.  Environ. Med.
48             56:679-683.
49       Wong, C.-M; Ma, S.; Hedley, A. J. Lam, T.-H. (1999b) Does ozone have any effect on daily hospital admissions for
50             circulatory diseases? J. Epidemiol. Community Health 53: 580-581.
51       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
52             two cities: effects of air pollution on hospital admissions in Hong Kong and London compared. Environ.
53             Health Perspect. 110:67-77.
         August 2005                                    7-168         DRAFT-DO NOT QUOTE OR CITE

-------
  1       World Health Organization. (2004) Meta-analysis of time-series studies and panel studies of paniculate matter
  2             (PM) and ozone (O3): report of a WHO task group. Copenhagen, Denmark: WHO Regional Office for
  3             Europe; document no. EUR/04/5042688. Available: http://www.euro.who.int/document/E82792.pdf
  4             [18 November, 2004].
  5       Woodruff, T. I; Grille, I; Schoendorf, K. C.  (1997) The relationship between selected causes of postneonatal infant
  6             mortality and paniculate air pollution in the United States. Environ. Health Perspect. 105: 608-612.
  7       Xu, X.; Ding, H.; Wang, X. (1995) Acute effects of total suspended particles and sulfur dioxides on preterm
  8             delivery: a community-based cohort study. Arch. Environ. Health 50: 407-415.
  9       Yang, Q.; Chen, Y.; Shi, Y.; Burnett, R. T.; McGrail, K. M; Krewski, D. (2003) Association between ozone and
10             respiratory admissions among children and the elderly in Vancouver, Canada. Inhalation Toxicol.
11             15:1297-1308.
12       Yang, C.-Y.; Chen, Y.-S.; Yang, C.-H.; Ho, S.-C. (2004a) Relationship between ambient air pollution and hospital
13             admissions for cardiovascular diseases inKaohsiung, Taiwan. J. Toxicol. Environ. Health Part A 67: 483-493.
14       Yang, C.-Y.; Chang, C.-C; Chuang, H.-Y.; Tsai, S.-S.;  Wu, T.-N.; Ho, C.-K. (2004b) Relationship between air
15             pollution and daily mortality in a subtropical city: Taipei, Taiwan. Environ. Int. 30: 519-523.
16       Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A. (2000) Exposure
17             measurement error in time-series studies of air pollution: concepts and consequences. Environ. Health
18             Perspect. 108: 419-426.
19       Zhu, L.; Carlin, B. P.; Gelfand, A. E. (2003) Hierarchical regression with misaligned spatial data: relating ambient
20             ozone and pediatric asthma ER visits in Atlanta. Environmetrics 14: 537-557.
21       Zidek, J. V. (1997)  Interpolating air pollution for health impact assessment. In: Barnett, E. V.; Turkman, K. F., eds.
22             Pollution Assessment and Control. New York, NY: John Wiley & Sons. (Statistics for the Environment,
23             no. 3).
24       Zidek, J. V.; Wong, H.; Le, N.  D.; Burnett, R. (1996) Causality, measurement error and multicollinearity in
2 5             epidemiology. Environmetrics 7: 441 -451.
26       Zidek, J. V.; White, R.; Le, N.  D.; Sun, W.; Burnett, R.  T. (1998) Imputing unmeasured explanatory variables in
27             environmental epidemiology with application to health impact analysis of air pollution. Environ. Ecol. Stat.
28             5:99-115.
29       Zmirou, D.; Barumandzadeh, T.; Balducci, F.; Ritter, P.; Laham, G.; Ghilardi, J.-P. (1996) Short term effects of air
30             pollution on mortality in the city of Lyon, France, 1985-90. In: St Leger, S., ed. The APHEA project. Short
31             term effects of air pollution on health: a European approach using epidemiological time series data. J.
32             Epidemiol. Community  Health 50(suppl. 1): S30-S35.
33       Zmirou, D.; Schwartz, J.; Saez, M.; Zanobetti, A.; Wojtyniak, B.; Touloumi, G.; Spix, C.; Ponce de Leon, A.;
34             Le Moullec,  Y.; Bacharova, L.; Schouten, J.; Ponka, A.; Katsouyanni, K. (1998) Time-series analysis of air
35             pollution and cause-specific mortality.  Epidemiology 9: 495-503.
36
         August 2005                                    7-169        DRAFT-DO NOT QUOTE OR CITE

-------
 i        8.  INTEGRATIVE SYNTHESIS:  OZONE EXPOSURE
 2                             AND HEALTH EFFECTS
 3
 4
 5     8.1  INTRODUCTION
 6          This integrative synthesis is structured to provide a coherent framework for the assessment
 7     of health risks associated with human exposures to ambient surface-level (tropospheric) ozone
 8     (O3) in the United States. The main goal of the chapter is to integrate newly available scientific
 9     information with key findings and conclusions from the 1996 O3 AQCD (U.S. Environmental
10     Protection Agency, 1996a), so as to address issues central to the EPA's assessment of evidence
11     needed to support the current review of the primary O3 NAAQS.  The integrated assessment of
12     key findings and conclusions provided here and elsewhere in this document with regard to O3
13     exposure and health effects will be drawn upon and their policy implications  considered in an
14     Ozone Staff Paper prepared by EPA's Office of Air Quality Planning and Standards (OAQPS).
15     The analyses provided in that Staff Paper aim to "bridge the gap" between scientific assessments
16     in this criteria document and judgments required of the EPA administrator in evaluating whether
17     to retain or, possibly, to revise the current primary O3 NAAQS. Other types of scientific
18     information concerning ambient O3 welfare effects (i.e., tropospheric O3 effects on vegetation
19     and ecosystems, relationships to surface-level solar UV flux/climate changes, and effects on
20     man-made materials) are assessed in ensuing Chapters 9, 10, and  11. That information will also
21     be considered in the OAQPS staff paper in posing options regarding the secondary O3 NAAQS.
22          As discussed in Chapter 2 of this document, O3 found in the earth's troposphere generally
23     originates from photochemical reactions that are predominantly catalyzed by the interaction of
24     sunlight with precursor pollutants, especially nitrogen oxides (NOX) and hydrocarbons such as
25     volatile organic compounds (VOCs), emitted by surface-level mobile and stationary sources.
26     Other photochemical oxidants, such as peroxyacetyl nitrate (PAN) and hydrogen peroxide
27     (H2O2), are also generated along with O3 by such atmospheric interactions. In addition to the
28     tropospheric O3 generated by these interactions, some O3 is found near the earth's surface as the
29     result of its downward transport from the stratosphere.  However, in contrast to stratospheric O3,
30     which plays an important role in maintaining the habitability of the planet by shielding the
       August 2005                              8-1        DRAFT -DO NOT QUOTE OR CITE

-------
 1      surface from harmful solar ultraviolet (UV) radiation, tropospheric O3 at the surface can exert
 2      adverse effects on humans, nonhuman animal species, and vegetation. As was the case for
 3      previous O3-related NAAQS criteria revisions, the present criteria document focuses mainly on
 4      the assessment of health and welfare effects resulting from exposures to surface-level
 5      concentrations of tropospheric O3, whereas less attention is accorded to the distinctly much more
 6      limited available information on other photochemical oxidants, e.g., PAN or H2O2.
 7           Based on the criteria review completed in 1978, the original primary and secondary
 8      NAAQS set in 1971 for total photochemical oxidants were revised in 1979 to focus on O3  as the
 9      indicator for new primary and secondary standards that were attained when the expected number
10      of days per calender year with maximum 1-h average O3 concentrations >0.12 ppm did not
11      exceed one. The  NAAQS for ambient O3 were revised in 1997 by replacing the 1-h standards
12      with an 8-h primary standard that is met when the 3-year average of the annual fourth highest
13      daily maximum 8-h average concentration is <0.08 ppm.  The new 1997 primary standard was
14      based on various  scientific supportive data from experimental human exposure, animal
15      toxicological and epidemiological studies, as assessed in the 1996 O3 AQCD and in the  1996 O3
16      Staff Paper (U.S.  Environmental Protection Agency, 1996b).
17
18      8.1.1  Chapter Organization
19           In addition to providing the above brief background information regarding prior O3
20      NAAQS reviews  (including the 1997 EPA revision of the O3 NAAQS), this first section
21      (8.1 Introduction) of the integrative synthesis chapter aims to orient the reader to the
22      organization and  content of the chapter. The next section (Section 8.2) focuses on air quality
23      trends and current ambient O3levels to help provide context for the ensuing discussions of O3
24      exposures and associated health effects. The subsequent sections (8.3, 8.4, and 8.5)
25      then integrate newly available key scientific information assessed in  Chapters 4 through 7 of
26      this document, including integration of information on O3 dosimetry, toxicological information
27      derived from controlled human exposure and laboratory animal studies, and epidemiologic
28      evidence.
29           These sections collectively  address the following topics:  (1) ambient O3 exposures,
30      personal exposures, and dosimetric considerations; (2) experimental  studies on toxicological
31      responses to acute O3 exposures in humans (clinical studies) and both acute and chronic effects

        August 2005                               8-2        DRAFT -DO NOT QUOTE OR CITE

-------
 1     in animals; (3) epidemiological evidence for associations between O3 exposure of human
 2     populations and health effects and the strength and robustness of these associations;
 3     (4) integration of the experimental and epidemiological evidence; (5) biological mechanisms and
 4     other evidence useful in judging the plausibility of adverse health effects being associated with
 5     human exposures to ambient O3 levels encountered in the United States; and (6) identification of
 6     susceptible and vulnerable populations likely at increased risk for O3-related health effects and
 7     numbers of people potentially falling in such categories in the United States.
 8           The present chapter mainly focuses on discussion of new scientific information that has
 9     become available since the 1996 O3 criteria review that supported EPA's revision of the O3
10     NAAQS in 1997. This includes assessment of information published or accepted for publication
11     in peer-reviewed open literature mainly through December 2004, with a few particularly
12     pertinent and important studies published beyond that point also being considered.
13           Important data gaps and uncertainties that still exist with regard to various important issues
14     and research needs are also briefly noted for some key areas. Detailed discussion of such
15     research needs is beyond the scope of this document; however, such discussion is typically
16     undertaken later as part of EPA efforts focused on identification of O3 research needs and
17     development of associated research planning documents.
18
19
20     8.2  AMBIENT OZONE AIR QUALITY IN UNITED  STATES
21     8.2.1  Current Ozone Concentrations and Spatial Patterns
22           Ambient air O3 is monitored in the United States during 'ozone seasons', which vary in
23     length depending on location.  The ozone season extends all year in the Southwest. In most
24     other areas of the country, O3 is monitored typically from April to October. However, O3 is
25     monitored throughout the year in many urban areas, in as much as O3 is present the year round
26     not only in polluted areas but in clean areas as well. The median of the daily maximum 8-h
27     average O3 concentration  in the United States, averaged over May to September from 2000 to
28     2004 for all U.S. counties, was 0.049 ppm.  In 95% of all counties, the median of the daily
29     maximum 8-h average O3 concentration was less than 0.057 ppm. However, it should be noted
30     that most monitors are located in the East.  The daily maximum  1-hour concentrations were
       August 2005                              8-3        DRAFT -DO NOT QUOTE OR CITE

-------
 1     typically much higher in large urban areas or in areas downwind of them. For example, in
 2     Houston, TX they approached 0.20 ppm during this period.  Daily 1-hour maximum ozone
 3     concentrations were lower in the rest of the country, but were still above 0.12 ppm in many
 4     locations. Eight hour daily maximum concentrations were not as high, but tend to be highly
 5     correlated with 1-hour daily maximums.
 6          Within individual MSAs, O3 concentrations tend to be well  correlated across monitoring
 7     sites, although spatial variations in concentrations can be substantial.  In many city centers, O3
 8     concentrations tend to be lower than in  either upwind or downwind areas, largely due to reaction
 9     of O3 with NO emitted by motor vehicles. For example, much lower O3 concentrations overall
10     are found in downtown Los Angeles (e.g., in Lynwood) than at sites located further downwind
11     (e.g., in San Bernadino). The much higher downwind levels are formed from photochemical
12     reactions involving the urban emissions, including products formed as the result of reactions
13     titrating O3 in the urban core.  Thus, O3 concentrations tend to be higher downwind of urban
14     centers, and they decrease again in going to areas that are more remote from precursor sources.
15     Likewise, surface-level O3 can be depleted in rural areas close to NO sources, such as highways
16     and powerplants.
17
18     8.2.2   Diurnal and Seasonal  Variations
19          Ozone concentrations typically tend to peak in early to mid-afternoon in areas where there
20     is strong photochemical activity and to  peak later in the afternoon or during early evening in
21     areas where transport is more important in determining the O3 abundance.  Summertime maxima
22     in O3 concentrations occur in those U.S. areas where substantial photochemical activity acts
23     on O3 precursors emitted as the result of human activities. Monthly maxima can occur anytime
24     from June through August. However, springtime maxima are observed in some National  Parks,
25     mainly in the western United States, and at a number of other relatively unpolluted monitoring
26     sites throughout the Northern Hemisphere. For example, the highest O3 concentrations at
27     Yellowstone National Park tend to occur during April and May. Typically, monthly minima
28     tend to occur from November through February at polluted sites and during the fall at relatively
29     remote sites.
30
31

       August 2005                              8-4       DRAFT -DO NOT QUOTE OR CITE

-------
 1      8.2.3  Long-Term Trends
 2           National attention started to be focused in the 1940s on O3 and associated photochemical
 3      smog in the Los Angeles area.  Prior to the adoption of stringent emissions controls, peak levels
 4      of O3 were consistently higher in the Los Angeles area than are currently observed.  For
 5      example, in 1958, peak O3 concentrations measured in Los Angeles were about 0.6 ppm but have
 6      declined since then, although not at a steady rate. Peak O3 levels of 0.2 to 0.5 ppm were still
 7      found at some locations in the Los Angeles basin during the 1970s. For example, on two  days
 8      (October 13 and 14) during a 1978 episode, Tuazon et al. (1981) observed peak 1-h averaged
 9      values of O3 of nearly 0.4 ppm and nearly 0.5 ppm. Currently, peak 1-h and 8-h average O3
10      concentrations are about 0.17 and 0.15 ppm in the Los Angeles basin (cf. Figures  3-10 and 3-11).
11      High O3 levels were also earlier found throughout the rest of the United Sates as well, but peak
12      O3 levels have also gradually declined across the country during the 1980s. However, during
13      one particularly hot summer (of 1988) in the East, peak 1-h O3 concentrations of about 0.2 ppm
14      were observed in many eastern U.S. cities (U.S. Environmental Protection Agency,  1990).
15           Historically high O3 concentrations, as noted above, have not only been observed in the
16      United  States. For example, during an episode in Great Britain in 1976, peak O3 levels exceeded
17      0.25 ppm and daily  maximum 8-h O3 concentrations were above 0.1 ppm for 18 consecutive
18      days at one rural site (Wayne, 1991). Also, concentrations of O3 in the range found in
19      Los Angeles during the 1970s are still found in Mexico City.
20           Nationwide, 2nd highest 1-h ozone concentrations in the United States have decreased
21      dramatically during the past several decades, i.e., by approximately 29 percent from 1980 to
22      2003 and 16 percent from 1990 to 2003. Also, 4th highest 8-h O3 concentrations decreased by
23      approximately 21 percent since 1980 and 9 percent since 1990 (U.S. Environmental Protection
24      Agency, 2003). Trends in metrics for evaluating compliance with the O3 NAAQS (i.e., changes
25      in the 4th highest O3 concentration) can be found in EPA's "National Air Quality  and Emissions
26      Trends  Reports".  These reports indicate that the 4th highest O3 concentrations are still
27      decreasing nationwide, but the rate of decrease has  slowed since 1990.  However, such trends
28      have not been uniform across the United States. In general, reductions in the O3 metrics given
29      above have been largest in New England and in states along the West Coast and smallest in
30      midwestern states. Downward trends in California  O3 concentrations have been driven mainly
31      by notable decreases in Southern California, with reductions in other areas not being as large.

        August 2005                               8-5        DRAFT -DO NOT QUOTE OR CITE

-------
 1      Trends in peak O3 metrics do not necessarily reflect changes in O3 values across the middle of
 2      the distribution of O3 concentrations.  Of note, O3 concentrations towards the center of its
 3      nationwide distribution have not shown much change, and there are some indications that O3
 4      concentrations at the lower end of the distribution may even be increasing.
 5
 6      8.2.4   Interrelationships Between Ozone and Other Ambient Pollutants
 7           Data on ambient concentrations of other oxidants (e.g., H2O2, PAN) and oxidation products
 8      (e.g., HNO3, H2SO4) in the atmosphere are not nearly as abundant as they are for O3. Because
 9      data for such species are usually obtained only as part of specialized field studies, it is difficult to
10      relate observed ambient O3 concentrations to ambient levels of other oxidant species or oxidation
11      products. In general, such secondary species are expected to be at least moderately positively
12      correlated with O3.  On the other hand, primary species are expected to be more highly correlated
13      with each other than with secondary species, provided that the primary species originate from
14      common sources in given areas. Measurements of gas phase oxidants conducted as part of the
15      Southern Oxidants Study (SOS) indicated combined hydroperoxide (H2O2, CH3OOH, and
16      HOCH2OOH) concentrations typically in the range of several ppb. Concentrations of PAN, PPN
17      and MPAN also observed during the SOS likewise  indicated combined concentrations in the
18      range of several ppb. Oxidants are also present in airborne cloud droplets,  rain drops, and
19      particulate matter (PM). A few measurements of reactive oxygen species (expressed as
20      equivalent H2O2) in ambient fine PM indicated levels of less than 1% of those for ambient O3 on
21      a molar basis. However, it should be noted that these measurements are potentially subject to
22      both positive and negative artifacts.
23           Because PM is not a single distinct chemical species, but rather a mix of primary and
24      secondary species, relationships between ambient O3 and PM concentrations can be quite
25      complex. As an example of this complexity, PM25  concentrations positively correlated with O3
26      during the summer, but negatively correlated with O3 during the winter at Ft. Meade, MD. Also,
27      Ito et al. (2005) examined relationships between PM10 and O3 on a seasonal basis in several
28      urban areas (cf. Figure 7-24).  Seasonal relationships with ambient O3 similar to those at
29      Ft. Meade were found, reflecting the dominant contribution of PM2 5 to PM10 in the urban areas
30      studied (although PM10 generally contains a higher  fraction than does PM25 of primary [mainly
31      crustal] material). Possibly contributing to the higher correlations observed between fine PM

        August 2005                               8-6        DRAFT -DO NOT QUOTE OR CITE

-------
 1      and O3 in the summer is the fact that O3 can contribute to formation of submicron particles via
 2      interactions with various other atmospheric constituents present, such as terpenes, and other
 3      biogenically derived hydrocarbons from trees, other vegetation, and wood products. Formation
 4      of ultrafine particles by this mechanism is most likely to occur during afternoons of summer
 5      days when temperatures and O3 concentrations are sufficiently elevated to facilitate O3 reactions
 6      with increased amounts of terpenes emitted from vegetation. Bursts of ultrafine particle
 7      formation have been observed repeatedly in both urban and rural air. Woo et al. (2001), for
 8      example, reported rapid formation of ultrafine particles in the ambient air of Atlanta typically
 9      around noon in both summer and winter.  The mechanisms underlying such ultrafine particle
10      formation events may also involve other atmospheric reactions that are related to O3 formation,
11      such as the nucleation of H2SO4 (produced by oxidation of SO2) and, probably, NH3.
12
13      8.2.5    Policy Relevant Background (PRB) Ozone Concentrations
14          Background O3 concentrations used for NAAQS-setting purposes  are referred to as Policy
15      Relevant Background (PRB) O3 concentrations.  Policy Relevant Background concentrations are
16      those that would occur in the United States in the absence of anthropogenic emissions in
17      continental North America (defined here as the United States, Canada, and Mexico).  Such
18      PRB O3 concentrations include contributions from natural sources everywhere in the world and
19      from anthropogenic sources outside these three countries. For the purpose of informing O3
20      NAAQS decisions, EPA focuses on assessing risks to human health and environmental effects
21      from O3 levels in excess of PRB concentrations. Issues  concerning the methodology for
22      estimating PRB O3 concentrations are discussed in detail in Section AX3.9 of Annex AX3.
23          Contributions to PRB  O3 include photochemical reactions involving natural emissions of
24      VOCs, NOX, and CO, as well as the long-range transport of O3 and its precursors from outside
25      North America and the stratospheric-tropospheric exchange (STE) of O3. Processes involved in
26      STE are  described in detail in Section AX2.3 of Annex AX2.  Natural sources of O3 precursors
27      include biogenic emissions, wildfires, and lightning. Biogenic emissions from agricultural
28      activities are not considered in the formation of PRB O3.
29          Currently, estimates of PRB O3 concentrations are based on predictions generated by the
30      global scale, three dimensional, chemical transport model GEOS-CHEM (Fiore et al., 2003).
31      Estimates of PRB O3 concentrations cannot be derived solely from measurements of O3 at

        August 2005                               8-7        DRAFT -DO NOT QUOTE OR CITE

-------
 1      relatively unpolluted sites because of long-range transport from anthropogenic source regions
 2      within North America. It is impossible to determine sources of O3 at a particular location
 3      without ancillary data that could be used as tracers of sources or to calculate photochemical
 4      production and loss rates for O3. Policy relevant background O3 concentrations vary as a
 5      function of season, altitude, and total surface O3 concentration, with PRB O3 concentrations at
 6      the surface generally falling in the range of 0.015 to 0.035 ppm from 1300 to 1700 local time
 7      and tending to decline under conditions conducive to O3 episodes. The PRB concentrations are
 8      highest during spring and decline into summer; and higher values tend also to occur at higher
 9      elevations during the spring due to contributions from  hemispheric pollution and stratospheric
10      intrusions. The contribution to surface O3 by stratospheric intrusions is typically well below
11      0.020 ppm. Stohl (2001) and Sprenger et al. (2003) found that the maximum probability of
12      stratospheric intrusions reaching the 800 hPa level (-1800 m) was less than 1% and that higher
13      probabilities (1 to 2%, and 10%) applied for stratospheric intrusions penetrating to the 600 hPa
14      level (-4100 m) and 500 hPa level (-5400 m), respectively. Thus, stratospheric intrusions only
15      rarely contribute to elevated  surface-level O3 concentrations at low altitude sites but have a
16      higher (albeit still low) probability of elevating them at high-altitude sites.
17
18
19      8.3  FACTORS AFFECTING HUMAN EXPOSURE TO AMBIENT OZONE
20           Exposure to O3 and related photochemical oxidants varies over time due to changes in their
21      ambient  concentrations and because people move between locations having notably different
22      concentrations. The amount of O3 delivered to the lung is not only influenced by the ambient
23      concentration but also by the individual's breathing route and rate. Thus, activity level is an
24      important consideration in determining the potential O3 exposure  and dose received.
25           The use of data from ambient air monitoring  stations is still  the most common surrogate for
26      assigning exposure estimates in epidemiologic studies. Since the primary source of O3 exposure
27      is the ambient air, O3 concentration data from outdoor community monitoring sites  should
28      provide a relative assignment of exposure with time, if:  concentrations are relatively uniform
29      across the region; time-activities pattern are roughly the same across the study population; and
30      housing  characteristics (such as ventilation rates and O3 sinks contributing to indoor O3 decay
31      rates) are relatively constant for the study area. However, because these types of factors  often do

        August 2005                               8-8       DRAFT -DO NOT  QUOTE OR CITE

-------
 1      vary across populations and locations, some error tends to be associated not only with estimates
 2      of the magnitude of O3 exposure but, also, potentially with relative exposure assignments based
 3      solely on ambient monitoring data. Nevertheless, ambient O3 monitoring data appear to provide
 4      the most useful index of human O3 exposure currently available to help characterize health
 5      outcomes associated with O3 exposures of large population groups.
 6
 7      8.3.1   Personal Exposure
 8           Personal O3 concentrations have been measured for children, outdoor workers,  and
 9      individuals with COPD, all being populations potentially susceptible to O3 or other respiratory
10      irritants. Outdoor workers can be  expected to have somewhat higher O3 exposures than other
11      individuals, because they typically spend more time outdoors and often engage in prolonged
12      moderate and heavy  exertion activities.  Children also tend to be more active outside and,
13      therefore, often manifest a higher breathing rate than most adults. However, available exposure
14      measurement studies are not sufficient to allow for highly confident broad quantitative
15      generalization about the "typical"  magnitudes of observed differences in exposure between the
16      general population and such potentially susceptible subpopulations.
17
18      8.3.2   Indoor Concentrations
19           Apart from only a few specific indoor sources  such as photo-copying machines, O3 indoors
20      is derived from the infiltration of ambient air from outdoors.  Generally, O3 enters indoor
21      environments through infiltration from outdoors and through building components, such as
22      windows, doors, and ventilation systems. Ozone concentrations in indoor environments depend
23      primarily on the outdoor O3 concentration, outdoor/indoor infiltration and the air exchange rate
24      (AER). Once indoors,  O3 reacts on various surfaces and with airborne components of either
25      indoor or outdoor origin.
26           Indoor O3 concentrations tend to reflect outdoor concentrations and, hence, are higher
27      when outdoor O3 is higher. However, because O3 reacts indoors with surfaces and other
28      contaminants, O3 concentrations are typically lower indoors than outdoors. Gas phase reactions
29      occurring outdoors also produce other oxidants analogous to the production of photochemical
30      smog. The extent and rate of production of these other species indoors is a function of indoor O3
        August 2005                               8-9        DRAFT -DO NOT QUOTE OR CITE

-------
 1      concentrations and the presence of other necessary precursors (i.e., VOCs), along with an
 2      optimal AER.
 3           Several studies have measured O3 concentrations in residences, schools, office buildings
 4      and museums; and typical concentrations varied across all such locations.  However, indoor
 5      concentrations generally varied in relationship to the AER in the indoor environment (increasing
 6      with higher AER) and generally tended to be notably lower than outdoor ambient O3 levels.
 7      For example, one study examining the relationship between O3 concentrations indoors and
 8      outside of a school in New England reported average O3 concentrations of 20 ppb (0.020 ppm)
 9      indoors and 40 ppb (0.040 ppm) outdoors. With regard to mobile source microenvironments, as
10      is the case for other enclosed environments,  O3 exposures depend on the extent of mixing of
11      outdoor air into the vehicle cabin. Thus, if windows are kept open, O3 concentrations inside the
12      vehicle may be expected to approach outdoor values; but, if windows are kept closed and there is
13      air conditioning, then interior values can be much lower than those outside, especially if
14      recirculated air is used. For example, in one N.C. study involving police cars with air
15      conditioning  and recirculated air, O3 concentrations in the vehicle cabin (11.7 ppb average) were
16      less than half those outside (28.3 ppb average at outdoor monitoring sites in the area).
17           Although concentrations  of O3 may be reduced to lower levels once ambient O3 enters
18      indoor environments, it should be kept in mind that the indoor O3 may interact with other
19      airborne substances of indoor or outdoor origin that may be present indoors. For example,
20      Wainman et al. (2000) showed that O3 reacts with d-limonene, a common component of air
21      fresheners to produce submicron particles. These particles are found mainly in the size range
22      from 0.1 to 0.3 jim. Wainman  et al. noted that terpenes such as limonene are emitted by wood
23      products; that they are used as  solvents, as odorants in cleaning products, and as air fresheners;
24      and, because of their widespread uses, their concentrations are often higher indoors than they are
25      outdoors. In addition to particle formation, Weschler (2004) points out that gas phase products,
26      such as aldehydes and hydroperoxides, produced by reactions of O3 with terpenes and other
27      unsaturated carbon compounds may also be of concern.  During the formation of these products,
28      OH radicals are also produced  which can react with compounds that do not react with O3. To the
29      extent that building ventilation rates etc. remain constant between days characterized by high
30      and low O3, the concentrations of these other secondary pollutants formed indoors will tend to be
31      correlated with ambient O3. Thus, ambient O3 concentrations measured outdoors at community

        August 2005                               8-10       DRAFT -DO NOT QUOTE OR CITE

-------
 1      monitoring sites and/or personal O3 exposure monitor measurements may serve not only as
 2      indices of direct human exposure to O3 per se, but also as surrogate indices of exposures to
 3      broader O3-containing mixtures of ambient or indoor air contaminants.
 4
 5
 6      8.4  SYNTHESIS OF AVAILABLE INFORMATION ON OZONE-
 7           RELATED HEALTH EFFECTS
 8           The integrated synthesis of the latest available information on O3-related health effects
 9      poses large challenges, especially in view of the emergence of certain important new information
10      since the 1996 O3 AQCD, which adds greatly to the complexity of an integrative assessment.
11      Such information includes new findings from:
12        •  Dosimetry studies that clarify further factors potentially affecting regional distribution
             of O3 in the respiratory tract of humans and laboratory animals and providing improved
             bases by which to attempt animal-to-human extrapolations of experimentally-observed
             O3-induced health effects.
13        •  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;
14        •  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.
15
16           Previous criteria assessments, including the 1996 O3 AQCD, found that experimental
17      studies of controlled human and laboratory animal exposures to O3 provided the most clear cut
18      and compelling evidence with regard to characterizing O3-related health effects.  This section
19      first summarizes key dosimetry and health related findings derived from the 1996 O3 AQCD and
20      then integrates those findings with new information obtained since 1996 from human and animal
21      experimental studies. Ozone-induced physiological, pathological, cellular and biochemical
22      alterations are evaluated in order to assess human health effects due to ambient O3 exposures.
23      Also, the influence of O3-induced changes at cellular and molecular levels are integrated to
24      elucidate scientific bases for the observed physiological and pathological alterations. These
25      research results are evaluated in order both to help assess the biological plausibility of health
        August 2005                              8-11        DRAFT -DO NOT QUOTE OR CITE

-------
 1      outcome associations observed in epidemiologic studies and to assess the coherence of the
 2      overall body of evidence relevant to O3-related health outcome conclusions.
 3
 4      8.4.1   Key Health-Related Findings and Conclusions from the 1996
 5             Ozone Air Quality Criteria Document
 6          Based on extensive dosimetric and experimental data as well as growing epidemiologic
 7      evidence available at the time, the 1996 O3 AQCD arrived at a  set of findings and conclusions
 8      stated in relation to answering five key questions regarding potential health effects of ambient O3
 9      exposure. In general, the existing evidence was such to warrant a high degree of confidence in
10      those conclusions derived from experimental (controlled exposure) studies. Considerable
11      confidence could also be placed in the emerging field/panel studies providing observational
12      study results substantiating and extending the controlled exposure study findings. Other
13      epidemiologic studies provided highly suggestive, although less conclusive, indications of
14      increased morbidity (e.g., as indexed by emergency department visits, hospital admissions, etc.)
15      and, possibly, mortality being associated with exposure of human populations to ambient O3.
16      The main findings and conclusions derived from the 1996 ozone criteria review are recapitulated
17      (largely verbatim) below in relation to the five key questions addressed in the summary and
18      conclusions of the Integrative Synthesis in the 1996 O3 AQCD.
19
20      1.  What are the effects of short-term (<8-h) exposures to ozone?
21          Short-term O3 exposure of laboratory animals and humans causes changes in pulmonary
22     function, including tachypnea (rapid, shallow breathing), decreased lung volumes and flows,
23      and increased airway responsiveness to nonspecific stimuli. Increased airway resistance occurs
24      in both humans and laboratory animals, but typically at higher exposure levels than other
25     functional endpoints. In addition, adult human subjects experience O3 induced symptoms of
26      airway irritation such as cough or pain on deep inspiration.  The changes in pulmonary function
27      and respiratory symptoms occur as a function of exposure concentration, duration,  and level of
28      exercise. Adult human subjects with  mild asthma have qualitatively similar responses in lung
29      volume and airway responsiveness to bronchoconstrictor drugs as nonasthmatics. Respiratory
30      symptoms are also similar, but wheezing is a prevalent symptom in O3-exposed asthmatics in
31      addition to the other demonstrated symptoms of airway irritation. Airway resistance, however,

        August 2005                             8-12       DRAFT-DO NOT QUOTE OR CITE

-------
 1     increases relatively more in asthmatics from an already higher baseline. Recovery from the

 2     Affects ofO3 on pulmonary function and symptoms is usually complete within 24 h of the end of

 3     exposure, although other responses may persist somewhat longer.

 4       •  Increased O3 levels are associated with increased hospital admissions and emergency
            department visits for respiratory causes. Analyses from data in the Northeastern
             United States suggest that O3 air pollution is associated with a substantial portion (on the
            order of 10 to 20%) of all summertime respiratory hospital visits and admissions.

 5       •  Pulmonary function in children at summer camps in southern Ontario, Canada,  in the
            northeastern United States, and in Southern California is associated with O3
            concentration. Meta-analysis indicates that a 0.50-mL decrease in FEV, is associated
            with a 1 ppb increase in O3 concentration. For preadolescent children exposed to
            120ppb (0.12 ppm) ambient O3, this amounts to an average decrement of 2.4 to 3.0% in
            FEV,.  Similar responses are reported for children and adolescents exposed to O3 in
            ambient air or O3 in purified air for 1 to 2 h while exercising.

 g       •  Pulmonary function decrements are generally observed in healthy subjects (8 to 45 years
            of age) after 1  to 3 h of exposure as a function of the level of exercise performed and the
            O3 concentration inhaled during the exposure. Group mean data from numerous
            controlled human exposure and field studies indicate that, in general, statistically
            significant pulmonary function decrements beyond the range of normal measurement
            variability (e.g., 3 to 5% for FEVJ occur

 7            -at > 0.50 ppm O3 when at rest,

 g            -at > 0.37 ppm O3with light exercise (slow walking),

 9            -at >0.30 ppm O3 with moderate exercise (brisk walking),

10            -at >0.18 ppm O3 with heavy exercise (easy jogging), and

\\            -at >0.16 ppm O3with very heavy exercise (running).

12       •  Smaller group mean changes (e.g., <5%) in FEV, have been observed at lower
            O3 concentrations than those listed above. For example, FEV, decrements have been
            shown to occur with very heavy exercise in healthy adults at 0.15 to 0.16ppm O3, and
            such effects may occur in healthy young adults at levels as low as 0.12 ppm.  Also,
            pulmonary function decrements have been observed in children and adolescents at
            concentrations ofO. 12 and 0.14 ppm O3 with heavy exercise. Some individuals within a
            study may experience FEV, decrements in excess of 15% under these exposure
            conditions, even when the group mean decrement is less than 5%.

13       •  For exposures of healthy subjects performing moderate exercise during longer duration
            exposures (6 to 8 h), 5% group mean decrements in FEV, were observed at
14            -0.08 ppm O3 after 5.6 h,

15               0.10 ppm O3 after 4.6 h, and
       August 2005                              8-13       DRAFT -DO NOT QUOTE OR CITE

-------
 I                0.12 ppm O3 after 3 h.

 2       •   For these same subjects, 10% group mean FEVj decrements were observed at 0.12 ppm
             O3 after 5.6 and 6.6 h.  As in the shorter duration studies, some individuals experience
             changes larger than those represented by the group mean changes.

 3       •   An increase in the incidence of cough has been reported at O3 concentrations as low as
             0.12 ppm in healthy adults during 1 to 3 h of exposure with very heavy exercise. Other
             respiratory symptoms, such as pain on deep inspiration, shortness of breath, and lower
             respiratory scores (a combination of several symptoms), have been observed at 0.16 to
             0.18 ppm O3 with heavy and very heavy exercise. Respiratory symptoms also have been
             observed following exposure to 0.08, 0.10, and 0.12 ppm O3for 6.6 h with moderate
             levels of exercise.

 4       •   Increases in nonspecific airway responsiveness in healthy adults have been observed
             after 1 to 3 h of exposure to 0.40 ppm,  but not 0.20 ppm, O3 at rest and have been
             observed at concentrations as low as 0.18 ppm, but not to 0.12 ppm,  O3 during exposure
             with very heavy exercise. Increases in nonspecific airway responsiveness during 6.6-h
             exposures with moderate levels of exercise have been observed at 0.08, 0.10, and
             0.12 ppm O3.

 5           Short-term O3 exposure of laboratory animals and humans disrupts the barrier function of

 6     the lung epithelium, permitting materials in the airspaces to enter lung tissue,  allowing cells and

 1     serum proteins to enter the airspaces (inflammation), and setting off a cascade of responses.

 g       •   Increased levels ofPMNs and protein in lung lavage fluid have been observed following
             exposure of healthy adults to 0.20, 0.30, and 0.40 ppm with very heavy exercise and have
             not been studied at lower concentrations for 1- to 3-h exposures. Increases in lung
             lavage protein andPMNs also have been observed at 0.08 and 0.10 ppm O3 during 6.6-h
             exposures with moderate exercise; lower concentrations have not been tested.

 9           Short-term O3 exposure of laboratory animals and humans impairs alveolar macrophage

10     clearance of viable and nonviable particles from the lungs and decreases the effectiveness of

11     host defenses against bacterial lung infections in animals and perhaps humans.  The ability of

12     alveolar macrophages to engulf microorganisms is decreased in humans exposed to 0.08 and

13     0.10 ppm  O3for 6.6 h with moderate exercise.

14           Recent epidemiology studies addressing the  effects of short-term ambient exposure to O3 in

15     the population have yielded significant associations with a wide range of health outcomes,

16     including lung function decrements, aggravation of preexisting respiratory disease, increases in

17     daily hospital admissions and emergency department visits for respiratory causes, and increased

18     mortality. Results from lung function epidemiology studies are generally consistent with the
       August 2005                              8-14       DRAFT-DO NOT QUOTE OR CITE

-------
 1      experimental studies in laboratory animals and humans. An association between daily mortality
 2      and O3 concentration for areas with high O3 levels (e.g., Los Angeles) has been suggested,
 3      although the magnitude of such an effect is unclear.
 4
 5      2.  What are the effects of repeated, short-term exposures to ozone?
 6           During repeated short-term exposures, some of the O3-induced responses are partially or
 1      completely attenuated. Over a 5-day exposure, pulmonary function changes are  typically
 8      greatest on the second day, but return to control levels by the fifth day of exposure. Most of the
 9      inflammatory markers (e.g., PMN influx) also attenuate  by the fifth day of exposure, but markers
10      of cell damage (e.g., lactate dehydrogenase enzyme activity) do not attenuate and continue to
11      increase. Attenuation of lung function decrements is reversed following 7 to 10 days without O3.
12      Some inflammatory markers are also reversed during this time period, but others still show
13      attenuation even after 20 days without O3.  The mechanisms and impacts involved in attenuation
14      are not known, although animal studies show that the underlying cell damage continues
15      throughout the attenuation process.  In addition, attenuation may alter the normal distribution
16      ofO3 within the lung, allowing more O3 to reach sensitive regions, possibly affecting normal
17      lung defenses (e.g., PMN influx in response to inhaled microorganisms).
18
19      3.  What are the effects of long-term exposures to ozone?
20           Available data indicate that exposure to O3for months and years causes structural changes
21      in several regions of the respiratory tract, but effects may be of the greatest importance in the
22      centriacinar regions (where the alveoli and conducting airways meet); this region typically is
23      affected in most chronic airway diseases of the human lung.  This information on O3 effects in
24      the distal lung is extrapolated from animal toxicological studies because, to date, comparable
25      data are not available from humans. The apparent lack of reversal of effects during periods of
26      clean air exposure raises concern that seasonal exposures may have a cumulative impact over
27      many years.  The role of adaptive processes in this response is unknown but may be critically
28      dependent on the temporal frequency or profile of exposure. Furthermore, the inter species
29      diversity in apparent sensitivity to the chronic effects ofO3 is notable, with the rat representing
30      the lower limit of response, and the monkey the upper limit.  Epidemiological studies attempting
        August 2005                              8-15        DRAFT -DO NOT QUOTE OR CITE

-------
 1      to associate chronic health effects in humans with long-term O3 exposure provide only
 2      suggestive evidence that such a linkage exists.
 3           Long-term exposure in the females of one strain of mice to high O3 levels (1 ppm) caused a
 4      small, but statistically significant increase in lung tumors. There was no concentration-response
 5      relationship, and rats were not affected. Genotoxicity data are either negative or weak. Given
 6      the nature of the database, potential carcinogenicity in animals is uncertain. Ozone did not
 1      show tumor-promoting activity in a chronic rat study (at 0.5 ppm O^.
 8
 9      4.  What are the effects of binary pollutant mixtures containing ozone?
10           Combined data from laboratory animal and controlled human exposure studies of
11      O3 support the hypothesis that coexposure to pollutants, each at low-effect levels, may result in
12      Affects of significance. The data from human studies ofO3 in combination with NO2, SO2, H2SO4,
13      HNO3, or CO show no more than an additive response for lung spirometry or respiratory
14      symptoms. The larger number of laboratory animal studies with O3 in mixture with NO2 and
15      H2SO4 show that effects can be additive, synergistic, or even antagonistic, depending on the
16      exposure regimen and the endpoint studied. This issue of exposure to copollutants remains
17      poorly understood,  especially with regard to potential chronic effects.
18
19      5.  What population groups are at risk as a result of exposure to ozone?
20           Identification of population groups that may show increased sensitivity to O3 is based on
21      their (1) biological responses to O3, (2) preexisting lung disease (e.g., asthma), (3) activity
22      patterns, (4) personal exposure history, and (5) personal factors (e.g., age, nutritional status).
23           The predominant information on the health effects ofO3 noted above comes from clinical
24      and field studies on healthy, nonsmoking, exercising subjects, 8 to 45 years of age.  These studies
25      demonstrate that, among this group, there is a large variation in sensitivity and responsiveness
26      to O3, with at least a 10-fold difference between the most and least responsive individuals.
27      Individual sensitivity to O3 also may vary throughout the year, related to seasonal variations in
28      ambient O3 exposure.  The specific factors that contribute to this large intersubject variability,
29      however, remain undefined. Although differences may be due to the dosimetry ofO3 in the
30      respiratory tract, available data show little difference on O3 deposition in the lungs for
31      inhalation through  the nose or mouth.

        August 2005                              8-16        DRAFT -DO NOT QUOTE OR CITE

-------
 1           Daily life studies reporting an exacerbation of asthma and decrease in peak expiratory
 2     flow rates, particularly in asthmatic children, appear to support the controlled studies; however,
 3      those studies may be confounded by temperature, particle or aeroallergen exposure, and asthma
 4      severity of the subjects or their medication use. In addition, field studies of summertime daily
 5      hospital admissions for respiratory causes show a consistent relationship between asthma and
 6      ambient levels ofO3 in various locations in the Northeastern United States, even after
 1      controlling for independent contributing factors.  Controlled studies on mild asthmatics suggest
 8      that they have similar lung volume responses but greater airway resistance changes to O3 than
 9      nonasthmatics. Furthermore, limited data from studies of'moderate asthmatics suggest that this
10      group may have greater lung volume responses than nonasthmatics.
11           Other population groups with preexisting limitations in pulmonary function and exercise
12      capacity (e.g., chronic obstructive pulmonary disease, chronic bronchitis, ischemic heart
13      disease) would be of primary concern in evaluating the health effects ofO3.  Unfortunately, not
14      enough is known about the responses of these individuals to make definitive conclusions
15      regarding their relative responsiveness to O3.  Indeed, functional effects in these individuals with
16      reduced lung function may have greater clinical significance than comparable changes in
17      healthy individuals.
18           Currently available data on personal factors or personal exposure history known or
19      suspected of influencing responses to O3 follow.
20        •   Human studies have identified a decrease in pulmonary function responsiveness to O3
             with increasing age, although symptom rates remain similar.  Toxicological studies
             are not easily interpreted but suggest that young animals are  not more responsive
             than adults.
21        •   Available toxicological and human data have not conclusively demonstrated that males
             and females respond differently to O3. If gender differences exist for lung function
             responsiveness to O3, they are not based on differences in baseline pulmonary junction.
22        •   Data are not adequate to determine whether any ethnic or racial group has a different
             distribution of responsiveness to O3.  In particular, the responses ofnonwhite asthmatics
             have not been investigated.
23        •   Information derived from O3 exposure of smokers is limited.  The general trend is that
             smokers are less responsive than nonsmokers.  This reduced responsiveness may wane
             after smoking cessation.
        August 2005                              8-17        DRAFT -DO NOT QUOTE OR CITE

-------
 I        •   Although nutritional status (e.g., vitamin E deficiency) makes laboratory rats more
             susceptible to O3-induced effects, it is not clear if vitamin E supplementation has an effect
             in human populations. Such supplementation has no or minimal effects in animals.  The
             role of such antioxidant vitamins in O3 responsiveness, especially their deficiency, has
             not been well studied.
 2           Based on information presented in this document, the population groups that have
 3      demonstrated increased responsiveness to ambient concentrations ofO3 consist of exercising,
 4      healthy and asthmatic individuals, including children, adolescents, and adults.
 5
 6           Since the 1996 O3 AQCD evaluations, a distinctly more extensive database of air pollution
 7      epidemiologic studies has become available.  A subset of these studies which examined O3
 8      health effects  have reported a variety of O3-related health effects associations.  Based on the
 9      physiological, biochemical and molecular changes observed in controlled human exposure
10      studies and animal toxicological studies, new evidence is now available by which to evaluate the
11      biological plausibility and extent of coherence for various health outcomes (such as respiratory
12      and cardiovascular effects, fetal and infant development effects, and mortality) reported in the
13      epidemiologic studies as discussed in ensuing sections. Biological observations pointing
14      towards putative mechanisms of action in developing hypotheses to interpret associated
15      pathological symptoms reported in epidemiologic studies are also critically evaluated in
16      subsequent sections, as are in vitro and in vivo experimental studies using novel molecular
17      technologies to address potential mechanisms of action.
18
19      8.4.2  Assessment and Integration of New Experimental Evidence
20      8.4.2.1  Background on Cross-Cutting Issues
21           Discussion of several cross-cutting issues that will facilitate a clearer understanding of the
22      ensuing assessment is provided here to enhance an integrated and comprehensive understanding
23      of the experimental and epidemiologic studies on O3 health  effects. An important issue to
24      be considered is the extrapolation of observed effects in animals to humans,  from the perspective
25      of dosimetry and the strength and weaknesses of such extrapolation models. The most
26      challenging issue is the integration of (a) epidemiologic (observational) findings that suggest a
27      potential causative role of ambient O3 (with adjustments for other copollutants) in producing
        August 2005                              8-18        DRAFT -DO NOT QUOTE OR CITE

-------
 1      health effects with (b) physiological, biochemical and toxicological findings from experimental
 2      studies.
 3
 4      8.4.2.2  Approaches to Experimental Evaluation of Ozone Health Effects
 5           Three chapters in the current document provide detailed discussion of various experimental
 6      approaches utilized to evaluate O3-related health effects.  Chapter 4 discusses dosimetry issues
 7      pertinent to both animal and human exposure scenarios. Chapter 5 discusses the experimental
 8      studies of physiological, biochemical (cellular and molecular changes) and pathological
 9      observations in laboratory animals (including nonhuman primates, dogs, and rodent species) and
10      in vitro studies using cell culture systems (in certain cases, on humans cells recovered from
11      BALF postexposure to O3).  Chapter 6 evaluates studies on human volunteers exposed to O3
12      which have investigated a variety of physiological and biochemical endpoints.
13           In interpreting the results from the experimental approaches, one must consider the
14      following three issues: (1) exposure/dose considerations; (2) role of confounders; and
15      (3) interpretation of results from high dose exposures and animal to human extrapolations.
16      Earlier animal toxicology studies were carried out using relatively high O3 exposure
17      concentrations/doses that do not necessarily reflect "real-world" exposure scenarios.  Those
18      experiments were primarily aimed at understanding the pathophysiology associated with O3
19      exposure in healthy animals, to help understand potential mechanisms(s) of action, and to help
20      validate health outcomes reported in epidemiologic studies.  Since the 1996 O3 AQCD, the
21      majority of human and animal studies have used ambient and/or near ambient doses.  Earlier
22      controlled  chamber exposure studies on human volunteers mainly limited exposures to O3 alone
23      in comparison to sham (clean air) exposures, thus providing evidence concerning direct effects
24      of O3 per se versus more closely mimicking real-world atmospheric exposures to multipollutant
25      mixes.  Some newer air pollution clinical studies are utilizing various co-exposure regimens to
26      simulate more closely ambient exposure to air pollution mixtures; and the results from these
27      studies will be highly useful in developing better models to interpret the toxicological effects
28      associated with O3-containing ambient air pollutants mixes.
29           Interpretations of experimental studies of air pollution, as in the case of environmental
30      comparative toxicology studies, are affected by limitations associated with animal extrapolation
31      models.  The differences between humans and rodents with regard to O3 inhalability,  absorption

        August 2005                               8-19       DRAFT -DO NOT QUOTE OR CITE

-------
 1      and distribution profiles based on breathing pattern, exposure dose and differences in lung
 2      structure and anatomy (see Chapter 4 and 5 for details) have to be taken into consideration.
 3      Also, in spite of a high degree of homology and the existence of a high percentage of
 4      orthologous genes across human and rodents, particularly mice, extrapolation of molecular
 5      alterations at the gene level suffers from the regulatory control of various signaling units as
 6      simple as cis and trans activating transcription factor units.  Given these molecular differences,
 7      extrapolation of physiological parameters (which are under the control of various biochemical,
 8      endocrine and neuronal controls) observed between human and rodents represents a difficult
 9      task.
10
11      8.4.2.3  Interspecies Comparison of Experimental Results: Dosimetric Considerations
12           In this section, a brief overview of the experimental results  obtained from studies on
13      human and laboratory animals are comparatively analyzed and presented to provide background
14      for assessing biological plausibility and coherence discussed in detail in the following section.
15      Each subsection starts off with an introduction to what was known at the time of the publication
16      of the previous O3 AQCD (U.S. Environmental Protection Agency, 1996), followed by
17      discussion of new information.
18
19      Dosimetry Considerations
20           Dosimetric studies demonstrate fundamental relationships between ambient exposures and
21      doses to target tissues.  While experimental and theoretical dosimetry (modeling) studies of O3
22      have  proved to be valuable in the assessment of toxicity, they are most useful when conducted as
23      part of an integrated approach to determining the distribution  of inhaled O3 along the upper and
24      lower respiratory tract. Derivation of credible dosimetry estimates greatly facilitates the
25      development of useful extrapolation models by which to compare doses and effects across
26      species and subpopulations.
27           The state-of-the-art of O3 dosimetry, as described in 1996 O3 AQCD, indicated consistency
28      across data and models derived from in vivo human and animal studies, thus increasing the level
29      of confidence in the development of dosimetric extrapolation  models. Earlier dosimetry models
30      predicted that the tissue dose  of inhaled O3 was greatest at the bronchoalveolar junction, the
31      region experimentally shown to be most impacted by O3. Ozone  bolus inhalation studies in

        August 2005                              8-20       DRAFT -DO NOT QUOTE OR CITE

-------
 1      humans have indicated that inspired O3 reaches the distal airways and alveoli of resting humans;
 2      and, with increased inspiratory flow rates due to exercise, O3 penetrates deeper and in greater
 3      quantity to the distal regions of the lung.  These findings have been corroborated by observations
 4      of 18O3 (oxygen-18-labeled ozone) in the BALF of humans and rats (Hatch et al.,1994).
 5           Some acute responses to O3 have compared well across species when controlled for dose,
 6      indicating that animals and humans (a) respond to O3 in a dose-dependent manner, i.e., they
 7      exhibit increasing breathing frequency with an accompanying decrease in tidal volume
 8      (tachypnea), and (b) show similar changes in alveolar permeability as measured by protein in the
 9      bronchoalveolar lavage fluid (BALF).  These parallel changes in humans and animals were
10      sufficiently homologous to suggest a common mode of action.  It has also been recognized
11      that O3-induced spirometric changes, the hallmark of response in humans, also occur in exposed
12      rats when hyperventilated with CO2 stimulation.  However, the effect of anesthesia in the rodent
13      model in contrast to the awake human remains uncertain; but, as will be discussed, activity level
14      differences between species appear to strongly influence dose.  Nevertheless, most lung function
15      decrements subside with repeated exposures in both humans and animals, with analogous
16      attenuation of certain (but not all) parameters measured  in the BALF. The mechanisms
17      associated with attenuation are unclear but may involve endogenous antioxidants. The
18      significance of non-attenuated markers in BALF has been interpreted to relate to potential
19      chronicity of O3 effects.  Studies on long-term exposure in monkeys and rats do show long term
20      changes in the distal lung that appear to be represented by a near-linear dose-response pattern.
21      More thorough analysis of this dose response is needed, however.
22           As discussed in the 1996 O3 AQCD, Hatch  et al. (1994) compared responses of exercising
23      humans (15-min intervals of rest and exercise at 60 L/min for 2 h) to those of resting rats also
24      exposed to 0.4 ppm 18O3 (oxygen-18-labeled ozone) for  2 h. They observed 4 or 5 times the 18O3
25      dose (as adduct) in BALF constituents of humans as compared to those of F344 male rats.  This
26      4- to 5-fold difference appeared to be due to the exercise-stimulated hyperventilation of the
27      humans when compared to the rats. Only when the resting rats were exposed to 2 ppm O3for 2 h
28      at rest did the 18O3 labeling of BALF constituents and indices of effect (i.e., BAL cells and
29     protein at 24 h) compare favorably with those of the exercising humans exposed to 0.4 ppm 18O3
30     for 2 h with intermittent exercise. Thus, the rat and human appear to have similar sensitivities
31      to O3 when exercise is taken into  account as a dose modifier. It was further concluded in the

        August 2005                              8-21       DRAFT -DO NOT QUOTE OR CITE

-------
 1      1996 O3 AQCD that attempts to compare resting animal data to exercising human data obtained
 2      at similar O3 concentrations would likely underestimate the dose to the lung and, presumably, the
 3      resultant risk of effect.
 4           In the past decade, no further reports have been published on O3 uptake studies in animals,
 5      although several controlled human bolus and/or general O3 uptake studies have provided refined
 6      data. The bolus uptake studies suggest that prior exposure to O3 diminishes bolus uptake. In the
 7      earlier document, the effect of mode of breathing (oral or nasal) on O3 uptake was thought to be
 8      minimal, with approximately equal uptake via the nose or mouth. Newer bolus dose studies
 9      have demonstrated that the uptake and regional respiratory tract distribution of O3 is sensitive to
10      mode of breathing (nasal uptake greater than oral) and to air flow rate (uptake decreases with
11      increasing flow). Similarly, the change in breathing with exercise vs rest causes a shift in
12      regional O3 distribution, allowing deeper respiratory tract penetration, with resultant greater dose
13      and damage to respiratory bronchiolar and alveolar tissues (as predicted by the models described
14      in the 1996 O3 AQCD).
15           The efficiency of O3 uptake is chemically rate dependent.  The resultant reaction products
16      (hydrogen peroxide, aldehydes, and hydroxyhydroperoxides) created by ozonolysis of lipids in
17      airway  and  epithelial lining fluid are thought to mediate O3 toxicity. The dependence of O3
18      absorption on chemical-reaction rates is consistent with the observation of Bush et al. (2001),
19      that the rate of O3 uptake is lower than for C12 despite the similar gas-phase diffusion coefficients
20      of these two gases. The slower uptake rate of O3 relative to C12 appears due to the limiting
21      reaction rate of O3 in the epithelial lining fluid.  The work by Rigas et al. (1997) using the O3
22      bolus technique in humans, showing uptake to be increased by continuous  exposure to NO2
23      and SO2 and decreased by continuous O3 exposure,  suggests an important role for copollutant
24      exposures.  Thus, an inflammatory response may magnify the production of O3-reactive
25      substrates in the epithelial lining fluid when other oxidants are present.
26           New uptake studies (Ultman et al., 2004) carried out in controlled human clinical studies
27      have observed gender-specific differences in the uptake of O3, but these differences do not
28      correlate well with spirometric responses.  Rather, they appear to be related to breathing pattern
29      and lung size, with females having smaller lungs than males.  Other uptake studies carried out in
30      humans using environmentally relevant O3 concentrations have demonstrated the significance of
31      incorporating inter-subject variability in dose-response relationship prediction and extrapolation.

        August 2005                               8-22        DRAFT -DO NOT QUOTE OR CITE

-------
 1      Thus, a number of variables seem to have a degree of impact on O3 uptake, notably including
 2      age, route of breathing, breathing pattern, gender and certain pre-exposure conditions.  These
 3      differences are important in order to interrelate biological effect and risk assessment estimates.
 4           The general consistency observed in O3 uptake in animal and human experimental
 5      exposure studies provides increased confidence in the use of theoretical dosimetry modeling and
 6      the use of animal toxicological data (see Chapter 4 for detailed discussion).  Models have taken
 7      into consideration various factors such as age, as well as anatomical, physiological, and
 8      biochemical alterations. Incorporation of novel information, such as (a) the identification of
 9      primary site of acute cell injury, (b) the site of O3 reaction/diffusion in the epithelial lining fluid,
10      (c) the roles of intermediate reactive oxygen species (ROS) and lipid ozonation products in
11      oxidative injury, and (d) the roles of metabolic enzyme profiles in developing lung tissue, can be
12      expected to lead to refined novel models and better extrapolation.
13
14      8.4.2.4   Critical Analysis of Toxicological Effects of O3 Exposure
15           In the following subsections, research results generated from experimental studies on
16      humans and animals during the past decade are assessed (keeping in view the interspecies
17      differences discussed in the preceding section) in evaluating experimental evidence for
18      biological plausibility and coherence for O3 health effects discussed in the later sections.
19
20      8.4.2.4.1   Pulmonary Function
21           A number of controlled human exposure, animal, and epidemiological studies assessed in
22      the 1996 O3 AQCD demonstrated alterations in various measurements of pulmonary function.
23      Inhalation of O3 for several hours while physically active elicits both acute pathophysiologic
24      changes and subjective respiratory tract symptoms.  The pulmonary responses observed in
25      healthy human subjects exposed to ambient O3 concentrations include decreased inspiratory
26      capacity; mild bronchoconstriction; rapid, shallow breathing pattern during exercise; and
27      subjective symptoms of tracheobronchial airway irritation, including cough and pain during
28      inspiration. Acute O3 exposures also cause decreases in forced vital capacity (FVC), forced
29      expiratory volume in 1 s (FEVj), and increased airways resistance (SRaw). The severity of
30      symptoms and the magnitude of response depends on inhaled dose, individual O3 sensitivity, and
31      the extent of tolerance  resulting from previous exposures.

        August 2005                               8-23        DRAFT -DO NOT QUOTE OR CITE

-------
 1           A progressive decrease in tidal volume and a "compensatory" increase in frequency of
 2      breathing to maintain steady minute ventilation during exposure suggests a direct modulation of
 3      ventilatory control.  These changes in humans parallel responses of many animal species
 4      exposed to O3 and other lower airway irritants (Tepper et al., 1990). Pulmonary function
 5      evaluations carried out in several animal species on acute exposure to O3 generally show
 6      responses similar to those observed in humans, such as increased breathing frequency, decreased
 7      tidal volume, increased resistance and decreased FVC. These effects are observed at relatively
 8      low O3 concentrations (0.25 to 0.4 ppm) following several hours of exposure in many species.
 9      The alterations in breathing pattern return to normal within hours after exposure and the pattern
10      of attenuation in responses following repeated exposures is similar to that observed in humans.
11      When rats were exposed to concentrations > 1 ppm even breathing mechanics were found to
12      be affected.
13           The time course of spirometry responses to O3 exposure depends on the specific exposure
14      conditions. Early controlled human studies reviewed in the 1986 and 1996 O3 AQCD typically
15      reported statistically significant pulmonary responses in exercising (intermittent or continuous)
16      subjects exposed for 2 h to a concentration in the range of 0.12 to 0.4 ppm O3 (mimicking
17      midday ambient O3 peaks reported in Los Angeles, CA). Significant effects were not observed
18      following 2 h exposures in sedentary subjects below 0.5 ppm O3. Some later human studies
19      reviewed in the 1996 O3 AQCD utilized 6-8 h exposures with exercise in order to better mimick
20      longer exposures to ambient O3 (recognizing the more prolonged elevated ambient O3 levels
21      often observed in some urban areas in the northeastern states) and provided  some of the strongest
22      and most quantifiable concentration-response data on the acute health effects of O3 based on
23      pulmonary function tests.
24           All evaluations have indicated that there exists considerable interindividual differences in
25      the magnitude of responses to O3. However, an individual's lung function and  to a lesser extent,
26      respiratory symptom responses to O3 are reproducible over a period of time, indicating that some
27      individuals are consistently more responsive than  others to O3. Figure 8-1 illustrates the
28      variability in FEVj responses in young healthy adults following a prolonged (6.6 h) exposure
29      to O3, as summarized in the 1996 O3 AQCD.  Referring to this figure, the average FEVj response
30      following exposure to 0.08 ppm O3 is small (between a 5 and 10%  decrement).  However,  -18%
31      of the exposed subjects had moderate FEVj decrements of 10 to 20% and -8%  experienced

        August 2005                              8-24       DRAFT -DO NOT QUOTE OR CITE

-------
43 au •
O
.2. 40-
.Q
W 30-
limn
° 20-
c
8 10-
<5
n n
0 ppm





nn
__





n = 87
— i




0%



n

.

-
.

•
0.08 ppm
n = 60
26%



n
_
I
PI



•-I






OH nr,n

.

-
.

.
0.10 ppm
n = 32
31%










_






n





n n

.

•
.

•
0.12 ppm
n = 49
46%




1
In






n
niirin n n
               -10  0  10 20 30 40
                                  -10  0  10 20 30  40    -10  0  10  20  30  40
                                             FEV, (%Deerement)
                                                                          -10 0  10  20  30  40
       Figure 8-1.   Frequency distributions of FEVt decrements following 6.6-h exposures to 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. The percentage in each panel indicates
                    the portion of subjects having a FEVt decrement in excess of 10%.
       Source: McDonnell (1996).
 1     large FEVj decrements of greater than 20%.  This serves to emphasize that while average
 2     responses may be small and seem physiologically insignificant, some individuals typically
 3     experience distinctly more severe effects. As a further example of intersubject variability,
 4     Figure 8-2 illustrates the portion of young healthy adult males (24 yr old) predicted to have FEVj
 5     decrements of greater than 5, 10, and 15% when exposed to O3 during moderate exercise, as also
 6     presented in the 1996 O3 AQCD.
 7          New studies (assessed in Chapter 6 and Annex 6 of this document) which evaluated
 8     responses in hundreds of subjects clearly indicate that FEVj decrements and symptom responses
 9     decrease with age beyond young adulthood (18 to 20 years).  Hazucha et al. (2003), for example,
10     examined gender and age differences in O3 responsiveness and found that young females lose O3
11     sensitivity faster than young males, but the rate is about the same for both genders by middle age
12     (see Figure 8-3).
13          The development of effects is time-dependent during both exposure and recovery periods,
14     with considerable overlap of evolving and receding effects. In healthy human subjects exposed
15     to typical ambient concentrations (i.e., <0.2 ppm O3), spirometric responses largely resolve
16     within a few hours (4 to 6 h) postexposure; but cellular effects persist for longer periods (-24 h).
17     Persisting small residual lung function effects are almost completely resolved within 24 hours.
       August 2005
8-25
DRAFT -DO NOT QUOTE OR CITE

-------
   1.0H
T3
O  0.8 H
                          i  0.6 H
                          Q.
                          O
                             0.4-
                          c
                          g
                          |  0.2 H
                             0.0-
                                                                  5% Change
                                        10% Change
                                        15% Change
                                0.0      0.2       0.4      0.6      0.8
                                             Dose(C[ppm]xT[h])
                                                                          1.0
       Figure 8-2.  Proportion of moderately exercising healthy adults (24 yrs old) predicted to
                    have 5,10, or 15% decrements in FEVt as a function of concentration (0 to
                    0.12 ppm O3) times exposure duration (1 to 6.6 h).
       Source: McDonnell etal. (1995).
 1     In hyperresponsive individuals, the recovery takes longer (as much as 48 h) to return to baseline
 2     values. The majority of these responses are attenuated after repeated exposure, but such
 3     tolerance to O3 is lost within a week postexposure. The biochemical indicators of lung injury
 4     and associated morphological changes were not found to be attenuated in the majority of
 5     laboratory animals. Unfortunately, no data are available on pulmonary function changes in
 6     animals upon chronic exposure to O3.  However, earlier work of repeated exposure of rats to an
 7     episodic profile of O3 demonstrated small but significant decrements in lung function that were
 8     consistent with early indicators of focal fibrinogenesis in the proximal alveolar region.
 9           In the 1996 O3 AQCD, O3-induced decrease in inspiratory capacity was hypothesized to be
10     the result of neurogenic inhibition of maximal inspiration due to stimulation of C-fiber afferents
11     either directly or from O3-induced inflammatory mediators. Earlier human studies (Coleridge
12     et al., 1993; Hazucha and Sanf Ambrogio, 1993) reported a role for bronchial C-fibers and
13     rapidly adapting receptors as primary vagal afferents responsible for O3- induced changes in
14     ventilatory rate and depth. As discussed in Chapter 6, the newer results of Passannante et al.
       August 2005
                       8-26
DRAFT -DO NOT QUOTE OR CITE

-------
           110-
           100-
         0)  80-
         (0
         •°  70H
>  6(H
LJJ
"-  501
            40-
            30-1
                                               110-
                                               100-
                                            -r- 90-
                                             <8 80-
                                             tn
                                             TO
                                             -0 70H
                                             s?
                                                     LU
                                                        50-
                                                        40-
                                               30-I
  o°5 g   o     oo_»_ .
0 o 08° °°°  J>'S
 Oo°io,8o-Mo0
 ^ »"o o
-f 0  O O  o ° <=
                                                                           0 o
              15   20  25  30  35  40   45   50   55  60
                           Age (years)
                                                 15   20   25   30  35  40  45  50   55   60
                                                               Age (years)
       Figure 8-3.  Effect of age on FEVt responses to O3 exposure (0.42 ppm for 1.5 h with
                    intermittent exercise). Left and right panels are data for males (n = 146; 19 to
                    60 yrs old) and females (n = 94; 18 to 59 yrs old), respectively. On average,
                    FEVj responses to O3 exposure decrease with increasing age. However, there
                    is a large amount of intersubject variability in responses, e.g., responses of
                    20 to 25 year olds range from a small increase to greater than a 50%
                    decrement in FEVt following O3 exposure.
       Source: Adapted from Hazucha et al. (2003).
 1     (1998) also support C-fiber stimulation as a primary mechanism of the O3-induced reduction in
 2     inspiratory capacity and suggest a role for nociceptive mechanisms. This neurogenic mechanism
 3     also likely has an effect on airway responsiveness and lung inflammation.
 4          Lung function changes due to O3 exposure have been evaluated in patients with preexisting
 5     respiratory diseases under experimental controlled exposure regimens, with or without physical
 6     exertion in the form of intermittent exercise. These new studies found minimal O3-induced
 7     effects in  COPD patients. For example, Gong et al. (1997a) exposed nine COPD patients
 8     (0.24 ppm O3 for 4 h with intermittent exercise) and observed a nonsignificant FEVj decrement
 9     of - 8% in COPD patients that was not statistically different from the - 3% decrement seen in
10     healthy subjects.  Augmenting observations discussed in the 1996 O3 AQCD, newer studies  of
11     asthmatics (see Chapter 6) continue to indicate that pulmonary function deficiencies detected by
12     spirometric analyses are somewhat increased relative to healthy  controls. A tendency for
13     increased  O3-induced pulmonary function responses were reported in asthmatics relative to
       August 2005
                                        8-27
DRAFT -DO NOT QUOTE OR CITE

-------
 1      healthy subjects exposed to O3 concentrations of <0.2 ppm for 4- 8 h duration (Scannell et al.,
 2      1996).  Similarly, Alexis et al. (2000) observed statistically significant O3-induced decreases
 3      in FEVj in mild atopic asthmatics that tended to be greater than experienced by healthy control
 4      subjects.  In a longer exposure duration (7.6 h) study, Horstman et al. (1995) reported that mild-
 5      to-moderate asthmatics exposed to 0.16 ppm O3 had FEVj decrements that were significantly
 6      greater than in healthy subjects (19% versus  10% respectively). Moreover, Horstman et al.
 7      (1995) found that responses of asthmatics were more severe in patients with lower baseline lung
 8      function.  Though most controlled human exposure studies may not provide the required
 9      statistical power (due to the limited number of subjects compared to panel or field studies),
10      they do suggest that asthmatics are at least as sensitive, if not more, than healthy subjects.
11           In addition to effects of O3 exposure on the large airways as indicated by spirometric
12      responses, O3 exposure also affects the function of the small airways and parenchymal lung.
13      Studies reported by Foster et al. (1993, 1997) that examined the effect of O3 on ventilation
14      distribution in healthy adult males suggest a prolonged O3 effect on the small airways and
15      ventilation distribution in some individuals.  Animal  toxicology studies have shown the
16      centriacinar region (CAR) of the lung (the segment between the last conducting airway and the
17      gas exchange region) to be a region highly susceptible to O3-induced damage (epithelial cell
18      necrosis and remodeling of respiratory bronchioles) and seem to be reasonably predictive of
19      similar morphological changes as  being likely to occur in humans.  Unfortunately, common
20      pulmonary function tests do not measure acute changes in the small airways of the CAR.
21      Identification of acute effects of O3 in small airways, if any, would lend additional support for
22      concerns about long-term effects of repeated O3 exposures.
23
24      8.4.2.4.2   Airway Responsiveness
25           Increased airway responsiveness, also referred to as airway hyperresponsiveness (AHR) or
26      bronchial hyperreactivity, is an indicator of enhanced reactivity of airways to
27      bronchoconstriction induced by a  variety of stimuli (exposure to cold air, allergens or exercise).
28      AHR is assessed by airway function (either spirometry or plethysmography) after inhalation
29      exposure to specific (antigen, allergen) or nonspecific (methacholine, histamine)
30      bronchoconstrictor  stimuli. It was recognized in the  1996 O3 AQCD that exposure to O3 to
31      induce AHR in humans usually resolves in 18-24 h after exposure in a majority of subjects, but

        August 2005                               8-28       DRAFT -DO NOT QUOTE OR CITE

-------
 1      may persist in some individuals for longer periods. Gong et al. (1997b) found that subjects with
 2      asthma developed tolerance to repeated O3 exposures in a manner similar to normal subjects;
 3      however, there were more persistent effects of O3 on airway responsiveness, which only partially
 4      attenuated when compared to filtered air controls.  Such an occurrence and duration of increased
 5      nonspecific airway responsiveness following O3 exposure could have clinical  implications in
 6      asthmatics, possibly putting them at potential increased risk for more prolonged bouts of
 7      bronchoconstriction in response to various triggering stimuli (e.g., allergens, cold air, etc.).
 8           Studies examining the effects of O3 on exacerbations of antigen-induced asthma suggested
 9      that allergen-specific increased airway responsiveness indeed occurs in mild asthmatics upon
10      exposure to O3. Torres et al. (1996) confirmed that higher O3 concentrations cause increased
11      airway reactivity to specific antigens in subjects with mild allergic asthma and, to a lesser extent,
12      in subjects with allergic rhinitis, after exposure to 0.25 ppm O3 for 3  h. This enhancement of
13      allergen responsiveness after O3 exposure appears to be time dependent, suggesting that the
14      timing of allergen challenge in O3-exposed subjects with allergic asthma is important.
15      Significant, clinically relevant decreases in pulmonary function have been observed in the
16      early phase allergen response in subjects with rhinitis after consecutive (4-day) exposure to
17      0.125 ppm O3 (Holz et al. 2002).  Similar increased airway responsiveness to  house dust mite
18      antigen 16-18 h postexposure to a single dose of O3 (0.16 ppm for 7.6 h) was also observed in
19      asthmatics. These observations suggest that O3 exposure may be a clinically important factor
20      that can exacerbate the response to ambient bronchoconstrictor substances in individuals with
21      preexisting allergic asthma and that its influence may be both immediate and persist for
22      relatively long periods of time.
23           An extensive laboratory animal study database (using rats, mice, guinea pigs, and rabbits),
24      exploring the effects of acute, long-term,  and repeated exposures to O3, indicates that induction
25      of AHR occurs at relatively high O3 concentrations.  These studies provide clues to the roles of
26      physiological and biochemical components involved in this process,  but one has to exercise
27      caution in the interpretation of these results, as different mechanisms may be involved in
28      mediating high-dose and low-dose responses.  Some of these studies indicated differences
29      in O3-induced AHR between immature and adult rats and also between obese  and lean mice
30      strains.  Some of the ex-vivo studies carried out in New Zealand white rabbits using
31      environmentally relevant O3 concentrations indicated O3-induced  alterations in tracheal epithelial

        August 2005                               8-29       DRAFT -DO NOT QUOTE OR CITE

-------
 1      functions and potential O3-induced direct vascular constriction. As observed in humans, the
 2      acute changes in AHR do not persist upon long-term exposure in animals exposed to near-
 3      ambient concentrations of O3; and attenuation has been observed. Both human and animal
 4      studies indicate that airway responses are not associated with inflammation, but they do suggest
 5      a likely role for neuronal involvement.
 6
 7      8.4.2.4.3   Morphological and Biochemical Abnormalities
 8           Most of the research results alluded to the ensuing discussion come from toxicology
 9      studies using various laboratory animal species that were usually exposed to relatively high,
10      non-ambient concentrations of O3. However, these exploratory and mechanistic studies may
11      provide important and useful hypotheses to consider in integrating various health outcomes
12      observed or predicted by epidemiologic studies.  A limited number of controlled human
13      exposure studies evaluated cellular and biochemical parameters in the BALF.  These studies
14      have yielded limited evidence supporting the observations made in animal toxicology studies.
15      Keeping in view the species- specific differences, the morphological and biochemical alterations
16      in humans and animals are integrated in the following paragraphs to develop working hypotheses
17      to interpret human health outcomes.
18
19      Lung Injury and Morphological Changes
20           The 1996 O3 AQCD stated that short-term O3 exposure causes similar types of alterations
21      in lung morphology in all laboratory animal species studied, including primates. The cells in the
22      CAR have been recognized as a primary target, possibly because it receives the greatest dose of
23      O3 delivered to the lower respiratory tract. The ciliated cells in the nasal cavity and airways and
24      Type I epithelial cells in the gas-exchange region are also identified as targets. Differences in
25      the distribution of antioxidants in the CAR of the lung were responsible  for the differences in
26      injury and morphological changes observed between nonhuman primates and rodents. Though
27      acute O3 exposure induces structural changes such as fibrosis in the CAR, these structural
28      alterations appear to be partially transient, with recovery shortly postexposure; but the time for
29      recovery is dependent on species and the dose of O3.
30           New studies reviewed in the 1996 O3 AQCD of lung morphological changes or damage
31      due to long-term or prolonged exposure to O3 found chronic lesions similar to early lesions of

        August 2005                              8-30       DRAFT -DO NOT QUOTE OR CITE

-------
 1      respiratory bronchiolitis, which have the potential to progress to fibrotic lung disease. Some of
 2      the morphological changes associated with long-term exposures, such as increases in
 3      hyperplastic epithelial cells, appear to reverse following cessation of O3 exposure. However,
 4      in the underlying interstitium of the CAR, proliferation of fibroblasts creates excess noncellular
 5      matrices. These processes are only partially reversible and may progress following cessation of
 6      exposure. This suggests initiation of focal interstitial fibrosis, which can progress to chronic
 7      degenerative lung disease.  Another important observation reported in the 1996 O3 AQCD was
 8      that of greater injury observed in the monkey's lung upon intermittent exposure (simulated
 9      ambient) compared to continuous exposure, suggesting a role for loss of tolerance in this
10      process.
11           Reports of morphological changes following chronic O3 exposures in animal studies
12      (rodents and primates) published since the 1996 AQCD allude to the earlier findings assessed in
13      that document. In rats, the effects of chronic -0.5 ppm O3 exposure included mucous cell
14      metaplasia,  hyperplasia of the nasal epithelium, increased mucosubstances, and increased Bcl-2
15      protein levels. In mice, lifetime exposures of 0.5 ppm O3 were linked to similar outcomes.
16      Taken together, the rodent studies suggest that O3 exposure may have the potential to induce
17      similar long-lasting alterations in human airways.  A series of new studies that utilized infant
18      rhesus monkeys and simulated seasonal ambient exposure (0.5 ppm  8 h/day for 5 days, every
19      14 days for  11 episodes) reported remodeling in the distal airways; abnormalities in tracheal
20      basement membrane; eosinophil accumulation in conducting airways; and decrements in airway
21      innervation, again confirming the potential greater injury due to seasonal exposure compared to
22      continuous exposure alluded to in the 1996 O3 AQCD.
23           One epidemiologic report by Sherwin et al. (2000) compared results for autopsy of the
24      lungs of Los Angeles and Miami residents and observed a significantly greater extent and
25      severity of centriacinar region alterations in the lungs of Los Angeles residents  independent of a
26      smoking effect.  These results suggest that the severity of CAR alterations may be related to the
27      higher O3 levels in Los Angeles. Similar observations of CAR thickening and deposition of
28      collagen seen with chronic O3 exposure in rat also  suggest progressive structural lung injury that
29      can  evolve into a more chronic form, such as fibrosis.  Again, however, one must be cautious in
30      extrapolating these laboratory animal observations to humans, given the exposure regimens and
31      doses used.

        August 2005                              8-31        DRAFT -DO NOT QUOTE OR CITE

-------
 1     Lung Inflammation and Permeability
 2           The 1996 O3 AQCD recognized respiratory tract inflammation and increased cellular
 3     permeability as two important biological markers of ozone exposure in both animals and
 4     humans. These distinct, independent biological events have been observed in all species studied
 5     in response to acute exposure to O3. Increased epithelial permeability and inflammation in the
 6     lower respiratory tract are measured by increases in bronchoalveolar lavage fluid (BALF)
 7     protein and/or albumin and neutrophils (PMNs), respectively. Nasal lavage (NL) fluid and cells
 8     from O3-exposed humans were used to assess the inflammatory and permeability changes in the
 9     upper respiratory tract. Structural changes in nasal mucosa have been demonstrated after O3
10     exposure in animals and humans.  The presence of PMNs in the lung has long been accepted as a
11     hallmark of inflammation  and as an important indicator that O3 causes inflammation in the lungs.
12     Importantly, respiratory tract inflammation may lead to significant health effects, including
13     impaired host defenses and irreversible structural alterations (as discussed earlier). Ozone-
14     induced mucous membrane cell metaplasia observed in rodents appears to be mediated by
15     inflammation.
16           Laboratory animals exhibit varying degrees of sensitivity to O3 exposure (see Chapter 5 for
17     detailed discussion); and this is evident even for the induction of pulmonary inflammation and
18     permeability. Newer animal toxicology studies on O3-induced inflammation reviewed in
19     Chapter 5 indicate that the lowest  ozone  concentration that had an effect on mouse lung
20     inflammation was also 0.11 ppm for 24 hours. Shorter durations (8 h) required greater
21     concentrations of ozone (0.26 ppm) for effects on epithelial permeability but had no effect on
22     inflammation. The lowest concentration of ozone that had an effect on epithelial permeability or
23     inflammation in the rat was 0.5 ppm for 3 hours. Subchronic exposures in animals suggest that
24     permeability changes are transient (and species-dependent) and return to control levels even with
25     continuing exposure.  Chronic animal O3 exposure studies suggest a role for persistent
26     inflammation in O3-induced alterations in lung structure and function.  Significant remodeling of
27     epithelium and underlying connective tissues in distal airways have been reported in rat exposed
28     to 0.25 ppm O3 (12h/day for 6 wk) and in monkeys exposed to 0.2 ppm O3 (8h/day for 90d).
29     Various factors such as viral infection, chemotactants and oxidized matrix fragments are also
30     implicated in the establishment and persistence of O3-induced inflammation.
        August 2005                              8-32        DRAFT -DO NOT QUOTE OR CITE

-------
 1           A number of controlled human exposure studies reviewed in the 1996 O3 AQCD clearly
 2      indicated that a single acute exposure (1-4 h) of humans to moderate O3 concentrations
 3      (0.2-0.6 ppm) while exercising at moderate to heavy levels results in a number of cellular
 4      and biochemical changes suggesting pulmonary inflammation and increased lung permeability.
 5      Both the inflammatory response and increased lung permeability have been observed as early as
 6      1 h and persisted for at least 18 h.  Devlin et al. (1991) reported these changes (increased
 7      neutrophils, inflammatory mediators such as PGE2 and IL-6) to occur in humans exposed to 0.08
 8      to 0.12 ppm O3 with moderate exercise for 6.6 h.  The newer studies reviewed in this document
 9      (see Chapter 6 for details) have provided additional information on three different aspects of O3-
10      induced inflammatory responses, such as (1) intersubject variability; (2) differential attenuation
11      profile for various inflammatory markers; and (3) effects of repeated exposures.
12           Mean changes in inflammatory markers seen with exposure to ambient levels of O3 (Devlin
13      et al., 1991) exhibited interindividual differences; and, in some individuals, the changes were
14      comparable to those observed in subjects exposed to 0.4 ppm (as reported by Koren et al., 1989),
15      suggesting that some individuals in the population may be quite sensitive at ambient levels of O3.
16      Mudway and Kelly (2004) examined O3-induced inflammatory responses (PMN influx) and
17      altered epithelial permeability (protein leakage) via a meta-analysis of 21 controlled human
18      exposure studies.  Their analysis of PMN responses is illustrated in Figure 8-4.  Tentatively,
19      Mudway and Kelly (2004) suggested that the O3 dose predicted to produce an average PMN
20      influx exceeding the 95% confidence interval for PMN levels following filtered air (FA)
21      exposures may, in essence, represent a threshold dose. For a 1  h exposure to 0.12 ppm O3, the
22      threshold dose for early phase PMN responses would not be exceeded unless an individual was
23      engaged in very heavy exercise (VE = 90 L/min).  However, a longer 8 h exposure to 0.08 ppm
24      O3 could reach the early phase PMN dose threshold during relatively light sustained activity
25      (VE =17 L/min).  For these same 1- and 8-h exposure scenarios, BALF protein levels would be
26      predicted to increase by about 1.1-fold.  Regarding late phase PMN responses, their threshold
27      dose was 26% greater than the early phase responses. Mudway and Kelly (2004) did note that
28      their "threshold" doses were for average PMN responses in healthy adults and that some
29      individuals would respond at lower doses. Indeed, Krishna et al. (1998) and Stenfors et al.
30      (2002) observed significant early and late phase PMN responses, respectively,  at doses below the
31      levels tentatively referred to as threshold doses by Mudway and Kelly (2004).  Additionally,

        August 2005                              8-33       DRAFT -DO NOT QUOTE OR CITE

-------
          35
          30
          20
          15
             0-6 h Postexposure
0                1000
        CVT (mg/m2 body surface area)
                                              2000
                                                               i
                                             18-24H Postexposure
                                                     10
                                                           , I  ,,'f , I  , I  , I  , I  , I
                                                                 800          1600
                                                              CVT (mg/m2 body surface area)
                                                                                         2400
      Figure 8-4.  Neutrophilia response in the distal airways postexposure (PE) to O3 or filtered
                   air. Early (0-6 h PE) and delayed (18-24 h PE) responses illustrated in the left
                   and right panel, respectively. Inhaled dose (CVT) is the product of O3
                   concentration, minute ventilation per body surface area, and exposure
                   duration. Data are mean (bars are standard errors) of neutrophils (% in
                   BAL) after ozone (v) or air (•) from 21 studies where subjects (18-40 yrs of
                   age) were exposed to between 0.08 and 0.6 ppm O3 for 1 to 6.6 h. The  dashed
                   lines (	) are average percent neutrophils following air exposures and the
                   shaded area is the 95% confidence interval (CI). Solid lines (—) illustrate the
                   linear relationship between neutrophil response and O3 dose with a 95% CI
                   illustrated by dotted lines (—)•

      Source: Adapted from Mudway and Kelly (2004).
1     significant inflammatory responses to O3 exposures that did not elicit significant spirometric

2     responses have been reported (Holz et al., 2005; McBride et al., 1994).

3           Soluble mediators of inflammation (e.g., the cytokines IL-6 and IL-8) as well as

4     arachidonic acid metabolites (e.g., PGE2, PGF2a, thromboxane, and leukotrienes [LTs] such

5     as LTB4) have been measured in the BAL fluid of humans exposed to O3. In addition to their

6     role in inflammation, many of these compounds have bronchoconstrictive properties and may be
      August 2005
                                    8-34
DRAFT -DO NOT QUOTE OR CITE

-------
 1      involved in increased airway responsiveness following O3 exposure. The time course for the
 2      inflammatory responses (including recruitment of neutrophils and other soluble mediators) is not
 3      clearly established, but a differential attenuation profile for many of these parameters is evident
 4      from the meta-analysis of 21 controlled human exposure studies reviewed by Mudway and
 5      Kelly (2004).
 6          Repeated exposures in humans also indicate ongoing cellular damage irrespective of
 7      attenuation of the inflammatory responses and lung function (Devlin et al. 1997; Torres et al.
 8      2000). Devlin et al. (1997) examined the inflammatory responses of humans repeatedly exposed
 9      to 0.4 ppm O3 for 5 consecutive days.  Several indicators of inflammation (e.g., PMN influx,
10      IL-6, PGE2, BAL protein, fibronectin) were attenuated after 5 days of exposure (i.e., values were
11      not different from FA).  Several other markers (LDH,  IL-8, total protein, epithelial cells) did not
12      show attenuation, indicating that tissue damage probably continues to occur during repeated
13      exposure.  The recovery of the inflammatory response occurred for some markers after 10 days,
14      but some responses did not return to normal even after 20 days. When re-exposed 2 weeks later,
15      changes in BALF indicated that epithelial cells appeared to be fully repaired (Devlin et al.,
16      1997). Kopp et al. (1999) observed inflammatory responses only after the first O3 peak in
17      summer; and its absence late in summer (even after exposure to higher levels of O3) may be
18      due to attenuation of the inflammatory response in the subjects.
19          Numerous studies reported acute O3-induced changes in lung epithelial permeability
20      assessed by indirect assay (increased levels of albumin and protein in BALF).  Few other studies
21      demonstrated O3-induced epithelial cell permeability through direct assessment of clearance
22      of 99mTc-DTPA (technetium-99m labeled diethylene triamine pentaacetic acid). For example,
23      Kehrl et al. (1987) showed increased 99mTc-DTPA clearance in healthy young adults at
24      75 minutes postexposure to 0.4 ppm O3 for 2 h. More recently, Foster and Stetkiewicz (1996)
25      have shown that increased 99mTc-DTPA clearance persists for at least 18-20 h post-O3 exposure
26      (130 min to average O3 concentration of 0.24 ppm), and the effect is greater at the lung apices
27      than at the base.
28          Interaction of O3 with the constituents of the extracellular lining fluid and the induction of
29      oxidative stress is implicated in injury and inflammation.  Animal toxicology and human in vitro
30      studies that evaluated biochemical mediators implicated in injury and inflammation found
31      alterations in the expression of cytokines, chemokines, and adhesion molecules, indicative of an

        August 2005                               8-35       DRAFT  -DO NOT  QUOTE OR CITE

-------
 1      ongoing active stress response as well as injury repair and regeneration processes.  Both animal
 2      and human studies indicate cellular and biochemical changes associated with inflammation and
 3      increased permeability, but the relationship between these changes and their role in lung function
 4      and airway responses is not known.
 5
 6      Host Defense
 1          Evidence for O3-induced dysfunction of host defense components and for subsequent
 8      enhanced susceptibility to bacterial lung infection stems from studies carried out in laboratory
 9      animals. Acute exposures of 0.08 ppm (3 h) O3 have been implicated in the mortality of mice
10      due to Streptococcal bacterial infection. Changes in antibacterial defenses are dependent on
11      exposure regimens, species and strain of test animal, species of bacteria, and age of animal (with
12      young mice being more susceptible to the effects of O3, for example).  Animal toxicology studies
13      indicated that acute O3-induced suppression of alveolar phagocytosis and immune functions
14      observed in animals appeared to be transient and were attenuated with continuous or repeated
15      exposures. A single study reviewed in the 1996 O3 AQCD reported  decrements in the ability of
16      alveolar macrophages (AMs) to phagocytose microorganisms upon exposure to 0.08-0.1 ppm O3
17      for 6.6 h (Devin et al., 1991). It has also been reported that O3 exposures can interfere with
18      AM-mediated clearance in the respiratory region of the lung and with mucociliary  clearance of
19      the tracheobronchial airways. Ozone-induced perturbations in the clearance process have been
20      found to be dose-dependent, with low dose exposures accelerating clearance and high doses
21      slowing the clearance process. Some respiratory tract regional- and  species-specific differences
22      have also been observed.
23          In vitro cultures of epithelial cells obtained from nonatopic and mild atopic asthmatics,
24      when exposed to O3 (0.01-0.1 ppm), exhibited significantly increased permeability compared to
25      cells from normal persons, thus indicating a potential inherent susceptibility of cells from
26      asthmatics for O3-induced permeability. New animal toxicology studies reported O3-induced
27      modulation of cell-mediated immune responses affecting the onset and persistence of infection
28      in rats (Cohen et al. 2001, 2002). However, there is no compelling evidence from animal
29      lexicological, human clinical or epidemiologic studies that O3 enhances the incidence of
30      respiratory viral infection in humans.
        August 2005                              8-36        DRAFT -DO NOT QUOTE OR CITE

-------
 1           The available data at this time indicate that acute O3 exposure has a potential to impair the
 2      host defense capability, primarily by interfering with the functions of alveolar macrophages.
 3      Any impairment in macrophage function may lead to decreased clearance of microorganisms or
 4      nonviable particles.  Compromised alveolar macrophage functions in asthmatics may increase
 5      their susceptibility to other O3 effects or to the effects of particles.
 6
 7      Biochemical A Iterations
 8           An extensive experimental database, including research assessed in 1996 O3 AQCD,
 9      suggests that potential biochemical alterations in various metabolic pathways (including
10      xenobiotic metabolism) are involved in lung injury,  inflammation, and functional alterations.
11      Interaction of O3 with the lipid constituents of pulmonary surfactant has been proposed as one of
12      the key mechanisms by which O3 exerts its toxic effects. Experimental evidence clearly
13      indicates a role for the initial interaction of O3 with lipid constituents of the ELF and generation
14      of lipid ozonation products and secondary redox mediators in the initiation of site-specific cell-
15      injury response cascades.  One such lipid ozonation  product, 4-hydroxynonenal, has been found
16      to bind to proteins and increased protein adducts in human alveolar macrophages, suggesting a
17      role for 4-hydroxynonenal in acute cell toxicity.  Cholesterol, the most abundant neutral lipid in
18      pulmonary surfactant is susceptible to attack by O3 resulting in multiple oxidized cholesterol
19      products, including the formation of cholesterol epoxide.  A 20-fold increase in cholesterol
20      epoxide in the BALF from mice exposed to 0.5 ppm O3 for 3 h suggests a potential role for this
21      oxidation product in O3 toxicity (Pulfer et al., 2005). Species- and region-specific increases in
22      lung xenobiotic metabolism have been observed in response to both short- and long-term O3
23      exposure. It has been well recognized that antioxidants in the ELF confer protection against O3
24      toxicity. But the observation of O3 reactivity even with environmentally relevant exposures,
25      questions their ability to quench O3 reactivity.  Species-specific and age-dependent changes in
26      the antioxidant metabolism add another dimension to their role in this process.  Carefully
27      controlled studies  of dietary antioxidant supplementation (Samet et al., 2001; Trenga et al.,
28      2001) reported some protective effects of a-tocopherol and  ascorbate for O3-induced spirometric
29      lung function decrements but not for the intensity of subjective symptoms and inflammatory
30      responses (including cell recruitment, activation, and a release of mediators). Dietary
31      antioxidants have also afforded partial protection to  asthmatics by attenuating postexposure

        August 2005                                8-37       DRAFT -DO NOT QUOTE OR CITE

-------
 1      bronchial hyperresponsiveness (Trenga et al., 2001).  In addition, two epidemiologic studies of
 2      street workers and asthmatic children in Mexico City found that subjects taking antioxidant
 3      supplements containing vitamins E and C were protected from O3-induced changes in lung
 4      function (Romieu et al., 1998, 2002).
 5           Based on the above discussion, it is evident that a very extensive experimental database
 6      accumulated from animal toxicology studies (including nonhuman primate studies) and limited
 7      controlled human exposure studies, has provided important insights into various biochemical,
 8      cellular, and molecular alterations in lung tissue exposed to O3.  The majority of these studies,
 9      although using acute exposure regimens and relatively high concentrations at times, do provide
10      credible hypotheses  regarding potential molecular mechanisms implicated in O3 toxicity.
11      Utilizing this information in relevant rodent-to-human extrapolation models with appropriate
12      species-specific adjustments may well provide useful information on initial biochemical
13      alterations that may  aid in the development of suitable biomarkers for O3 exposures/effects.
14
15      Cardiovascular Effects
16           Ozone-induced lung injury and permeability changes, as well as O3-induced alterations in
17      the hemodynamics may lead to O3 effects on the cardiovascular system.  Also, the interaction of
18      O3 with ELF, lipids,  and surfactants,  and the lipid ozonation products and ROS  generated in this
19      process have the potential to penetrate the epithelial barrier and to initiate toxic effects on the
20      cardiovascular system.  An increasing body of animal toxicology evidence suggests that
21      hematological and thermoregulatory  alterations (in heart rate variability and/or core body
22      temperature) may mediate acute cardiovascular effects. Studies carried out using isolated
23      perfused rat lung model (Delaunois et al., 1998) indicate inhibition of pulmonary mechanical
24      reactivity to bronchoconstrictors and persistent vasoreactivity of the vascular bed upon exposure
25      to O3 (0.4 ppm for 4 h). Earlier studies in rats indicate a potential role for platelet activating
26      factor (PAF) in O3-inflammatory response.  Recent observations of O3-induced generation of
27      oxysterols and p-epoxides from cholesterol in surfactant suggest that these lipid ozonation
28      products like lysophospholipids may initiate PAF-like activity and initiate clotting and
29      thrombolytic effects in the cardiovascular system.
30           A few human experimental studies have examined the potential effects of O3 exposure on
31      cardiovascular functions. For example, Gong et al  (1998) evaluated various cardiac function and

        August 2005                               8-38        DRAFT -DO NOT QUOTE OR CITE

-------
 1      hemodynamic variables in healthy and hypertensive adult males and observed impairment of
 2      alveolar-arterial oxygen transfer, leading them to suggest that such an impairment could lead to
 3      decreased oxygen supply to the myocardium.  Also, Foster et al (1993; 1997) have reported
 4      O3-induced ventilation-perfusion mismatch. Such an altered ventilation distribution profile
 5      observed even in relatively young healthy adults could contribute to the alveolar-arterial oxygen
 6      transfer impairment reported by Gong et al. (1998). Taken together, these observations support
 7      the regional differences in ventilation and perfusion in severe COPD patients reported by King
 8      and Briscoe (1968) and Kronenberg et al. (1973).  Such preexisting compromised gas exchange
 9      abnormalities would likely make lungs of individuals with COPD more vulnerable to O3-induced
10      gas exchange inhibition and reduced oxygen saturation.
11
12      8.4.3   Assessment of Epidemiological Evidence
13           Based on the O3 epidemiologic evidence available at the time, the 1996 O3 AQCD arrived
14      at the following conclusions:
15
16                 An association between daily mortality and 03  concentration for areas with high
17                 03 levels (e.g., Los Angeles) has been suggested, although the magnitude of such
18                 an effect is unclear. Increased 03 levels are associated with increased hospital
19                 admissions and emergency department visits for respiratory causes. Analyses
20                 from data in the northeastern United States suggest that 03 air pollution is
21                 associated with a substantial portion (on the order of 10 to 20%) of all summertime
22                 respiratory hospital visits and admissions. Pulmonary function in children at
23                 summer camps in southern Ontario, Canada, in the northeastern United  States,
24                 and in Southern California is associated with 03 concentration." (U.S. EPA, 1996,
25                 pl-29).
26
27      The 1996 O3 AQCD further stated that only suggestive epidemiologic evidence existed for health
28      effects of chronic ambient O3 exposure in the  population, and this was partly due to an inability
29      to isolate potential effects related to O3 from those of other pollutants, especially PM (U.S.
30      Environmental Protection Agency,  1996).
31           The discussion in this section of scientific  strength and limitations of the growing body of
32      epidemiologic evidence for associations between ambient exposure to O3 and various health
33      effects discussed is based primarily on Chapter 7 evaluations.  The following criteria were
34      considered in assessing the relative scientific quality of the epidemiologic studies:  (1) the
35      strength of reported associations, in terms of magnitude, statistical significance and statistical
36      power of effects estimates; (2) robustness of reported associations (based on defined health
37      endpoint criteria), potential confounding by copollutants; (3) consistency and coherency of the

        August 2005                                8-39       DRAFT -DO NOT QUOTE OR CITE

-------
 1      effect associations (4) temporality, in terms of lag periods between exposure and observed
 2      effects; and (5) biological plausibility of the observed O3-related health effects assessed in terms
 3      of their coherence in relation to findings derived from controlled human exposure studies which,
 4      overall, provide insights into the plausibility of reported O3 human health effects reflecting
 5      causal relationships.
 6           Many newly available epidemiologic studies have provided additional evidence for
 7      O3-related health effects beyond that previously known. Significant statistical associations have
 8      been reported by various investigators between acute O3 exposure and several respiratory and
 9      cardiovascular health endpoints, including:  mortality; hospital admissions; emergency
10      department visits; respiratory illness and symptoms; and changes in pulmonary function.
11      Similarly, associations have been reported between long-term exposure to O3 and increased
12      morbidity; development of respiratory disease; and declines in lung function and lung function
13      growth.  The numerous new epidemiological studies that have been conducted in areas across the
14      United States and Canada, as well as in Europe, Latin America, Australia and Asia, are
15      summarized in the Annex to Chapter 7. Based on evidence extracted from the full body of
16      epidemiologic studies carried out and reviewed since the 1996 O3 AQCD, it has been well
17      demonstrated that deleterious human health outcomes are positively associated with acute
18      ambient O3 concentrations currently encountered in the United States.
19
20      8.4.3.1   Strength and Consistency of Epidemiological Associations
21           As quoted above, assessments in the 1996 O3 AQCD supported a consistent relationship
22      between O3 concentration and respiratory illness, hospital visits and reduced lung function.
23      However, due to insufficient evidence examining O3-mortality associations and uncertainties
24      regarding weather model specification,  the 1996 O3 AQCD was limited to only a very qualitative
25      assessment of O3-mortality associations.  Since then, Generalized Additive Models (GAMs) have
26      become widely utilized for epidemiologic time-series analysis of health effects attributable to air
27      pollution, increasing our confidence in quantitative estimation of O3-mortality risks. Some
28      concerns have been raised regarding the use of default convergence criteria in applications of
29      commercially available software employed for GAM analyses to estimate air pollution-related
30      health effects, as discussed in the 2004 PM AQCD (U.S. Environmental Protection Agency,
31      2004a).  However, reanalyses of a number of studies, comparing results using default GAM

        August 2005                               8-40       DRAFT -DO NOT QUOTE OR CITE

-------
 1      convergence criteria to results from analyses using stringent GAM convergence criteria and/or
 2      from GLM analyses, found little difference among the O3 effect estimates obtained (as discussed
 3      in Chapter 7 of this document). Overall, the magnitude of the effect-size estimates observed
 4      for O3-mortality relationships tend to be relatively consistent across the newly available studies.
 5      The effect estimates for O3-morbidity endpoints have greater variability, but consistent positive
 6      associations between ambient O3 and various health outcomes have also been observed.
 7
 8      8.4.3.1.1  Acute Exposure Studies
 9          Numerous epidemiological  studies carried out over the past decade have added evidence to
10      the knowledge base assessed in the 1996 O3 AQCD, which included both (a) individual-level
11      camp and exercise studies that established a relationship between ambient O3 exposure and
12      human lung function decline and (b) aggregate time-series studies that suggested positive
13      relationships for O3-related respiratory morbidity. The new studies reviewed in Chapter 7 in this
14      document include numerous field/panel studies and time-series studies from various regions.
15      In field/panel studies on the effects of air pollution exposure, the most common health outcomes
16      measured were lung function and respiratory symptoms. The time-series studies examined daily
17      emergency department visits, hospital admissions, and mortality data.
18
19      Field/Panel Studies of Acute Exposure Effects
20      Pulmonary Function And Respiratory Symptoms
21      Healthy Individuals
22          Many of the new field/panel studies reviewed in Chapter 7 and the controlled human
23      exposure studies reviewed in Chapter 6 of this document provide additional data supporting two
24      major  findings reported in the 1996 O3 AQCD. First, acute O3 exposure is associated with a
25      significant decline in lung function parameters. Ozone-related lung function decrements were
26      most notable in children and asthmatics. In addition, adults who work or exercise  outdoors also
27      were found  to be vulnerable to O3-associated declines in lung function due to their increased
28      exposure to O3. Second, acute exposure to O3 is associated with increased respiratory symptoms,
29      particularly cough, and increased as-needed medication use in asthmatic children.  Immediate
30      effects of O3 were observed on both lung function and respiratory symptoms, with the strongest
        August 2005                              8-41       DRAFT -DO NOT QUOTE OR CITE

-------
 1      associations often observed at a lag of 0- or 1-day.  The two health outcomes are further
 2      discussed below.
 3           Pulmonary function was determined by either spirometry (forced expiratory volume in
 4      Is [FEVJ and forced vital capacity [FVC]) or by peak expiratory flow (PEF) meters. The
 5      spirometric parameter, FEVj is a strong and consistent measure of lung function.  PEF is a
 6      closely related but different metric of lung function, and PEF is more feasibly performed in field
 7      studies, using inexpensive peak flow meters that produce similar results to PEF measured
 8      spirometrically.
 9           In a number of newly available field/panel studies, FEVj was measured in panels of
10      exercising children, outdoor workers, and adult hikers exposed to ambient O3 while experiencing
11      elevated exertion levels. Collectively, the results of the new studies (discussed in Section
12      7.2.3.1) confirm and extend those from analogous field/panel studies assessed in the 1996 O3
13      AQCD and findings from experimental controlled human exposure studies indicating that
14      acute O3 exposures prolonged over several hours and combined with elevated levels of exertion
15      or exercise magnify O3 effects on lung function, as  evaluated in terms of FEVj.
16           For example, six field studies by three different research groups of 7- to 17-year-old,
17      healthy (nonasthmatic) children exposed for several hours to ambient O3 during increased
18      physical exertion in summer camp activities were assessed in the 1996 O3 AQCD. When
19      analyzed together by consistent statistical methods, the data from those studies  showed an
20      average relationship between afternoon FEVj and concurrent 1-h O3 concentrations of
21      -0.50 mL/ppb (95% CI: -0.63, -0.36), with individual slopes ranging from -0.19 to
22      -1.29 mL/ppb (Kinney et al., 1996). Four new field/panel studies (assessed in  Section 7.2.3.1 of
23      this document) that  evaluated pulmonary function in healthy school-aged children exposed to
24      mean 1-h max O3 concentrations ranging from 20 to 112 ppb found exposure-response functions
25      of approximately -0.18 to -1.42 mL/ppb. Also, two  other studies assessed in the 1996 O3
26      AQCD that measured lung function before and after well-defined exercise events (1/2-h long) in
27      adults during exposures to ambient O3 across 4 to 135 ppb found exposure-response slopes of
28      -0.4  mL/ppb (95% CI: -0.7,-0.1) (Selwyn et al.,  1985) and-1.35 mL/ppb (95% CI: -2.04,
29      -0.66) (Spektor et al., 1988). In comparison, new studies of healthy adult workers (street
30      workers, berry pickers) and hikers engaged in prolonged (>6 h) strenuous physical exertion at
31      mean exposure levels ranging from 40 to 123 ppb 1-h max O3 reported exposure-response slopes

        August 2005                              8-42       DRAFT -DO NOT QUOTE OR CITE

-------
 1      of -1.40 to -3.8 mL/ppb (as assessed in Section 7.2.3.1 in Chapter 7 of this document). The
 2      most representative data are those of Korrick et al. (1998) from a U.S. study of adult hikers that
 3      provided outcome measures stratified by gender, age, smoking-status, and presence of asthma
 4      within a population capable of above-normal exertion.
 5          Pulmonary function changes or declines measured in susceptible populations suggest that
 6      juvenile asthmatics are at greater risk with the largest O3-related decline of FEVj of -2.08%
 7      (95% CI: -6.24, 2.08) per 40 ppb increase in 1/2- h max O3 at a 2-day lag (Hoppe et al., 2003).
 8      An extended analysis by Hoppe et al.  (2003) examined individual susceptibility to O3 effects.
 9      Ozone responders were regarded as those with a greater than 10% change in FEVj.  Compared to
10      athletes and the elderly, a greater percentage of children and asthmatics (20% versus 5%) were
11      found to be sensitive to O3 effects on lung function.  The small sample sizes in these studies limit
12      extrapolation to larger populations; however, the studies indicate a trend that should be evaluated
13      in larger epidemiologic studies.
14
15      Asthma Panels
16          Several studies assessed in the 1996 O3 AQCD that evaluated elevated respiratory
17      symptoms and/or pulmonary function decrements in asthmatic children showed greater
18      responses in asthmatic  than in nonasthmatic  subjects, suggesting that asthmatic individuals
19      might constitute a sensitive population group in O3 epidemiologic studies.
20          Additional panel  studies carried out over the past decade to understand the effect of acute
21      exposure to O3 in asthmatics evaluated lung  function by PEF and/or respiratory symptoms (i.e.,
22      cough, wheeze, shortness of breath and medication use) ascertained by questionnaire. Several
23      additional studies (see Figures 7-1 and 7-2),  both in the U.S. and in other countries, reported
24      decrements in PEF to be associated with O3 exposures among asthmatics. One large U.S.
25      multicity study (Mortimer et al., 2002) associated O3 concentrations with the incidence of > 10%
26      declines in morning PEF (odds ratio of 1.30  [95% CI:  1.04, 1.61] per 30 ppb increase in 8-h
27      avg O3 for a 5-day cumulative lag). In a group of adult hikers in Mount Washington, NH
28      (Korrick et al., 1998), asthmatic subjects experienced a  four-fold greater decline in FEVj
29      compared to healthy individuals with the same exposure to O3.  Asthmatic hikers experienced a
30      mean change of -4.47% (95% CI:  -7.65,  -1.29) per 30 ppb increase in 8-h avg O3 while other
31      hikers had a mean change of-1.08%  (95% CI:  -2.49, 0.33).

        August 2005                              8-43       DRAFT -DO NOT QUOTE OR CITE

-------
 1           Several single-city studies did not observe statistically significant declines in lung function
 2      parameters (e.g., Delfmo et al., 1997; Hilterman et al., 1998), which might be partially
 3      attributable to small sample sizes and/or low levels of O3. Although results were not always
 4      statistically significant in these single-city studies, effect estimates were consistently negative,
 5      providing suggestive evidence that O3 was associated with lung function declines among
 6      asthmatics. Collectively, results from the large multicities study by Mortimer et al. (2002), as
 7      well as those from the smaller single-city studies suggest that exposure to O3 may be associated
 8      with declines in lung function in this potentially susceptible population.
 9           The majority of studies that evaluated respiratory symptoms (i.e., cough, shortness of
10      breath, and wheeze) and the increased use of asthma medication related to O3 exposure are also
11      focused on asthmatic children.  Two large U.S. studies (Mortimer et al., 2002; Gent et al., 2003)
12      and some international studies (Hilterman et al., 1998; Desqueyroux et al., 2002a,b) suggest
13      positive associations between O3 ambient concentrations and increased symptoms or asthma
14      medication use. In the multicities study by Mortimer et al. (2002), the odds ratio for the
15      incidence of symptoms (including cough, chest tightness,  and wheeze) was 1.35 (95% CI: 1.04,
16      1.69) per 30 ppb increase in 8-h avg O3 for a 4-day cumulative lag. As in the case of studies
17      examining the O3 effect on lung function, not all studies on respiratory symptoms observed
18      consistent positive associations with O3. For example, Avol et al. (1998) studied symptoms in
19      asthmatic, wheezy, and healthy children aged 10 to 12 years in southern California.  Some
20      symptom associations were noted but they were inconsistent, possibly due to relatively low O3
21      concentrations during the study period. The authors also noted that the study children did not
22      spend substantial time outdoors engaged in physical activities. Once again, the strong evidence
23      from the large multicities study by Mortimer et al. (2002), along with less consistent but
24      generally supportive evidence from several single-city studies suggest that O3 exposure may be
25      associated with increased respiratory symptoms and medication use in asthmatic  children.
26
27      School Absenteeism
28           Two large U.S. studies (Chen  et al., 2000; Gilliland, 2001) and one study from Seoul,
29      Korea (Park et al., 2002) investigated the relationship between ambient O3 concentrations and
30      school absenteeism. Results from all these studies suggested a positive association between O3
31      and absences from school, with each one arriving at these associations using different lag

        August 2005                               8-44       DRAFT -DO NOT QUOTE OR CITE

-------
 1      periods. Chen et al. (2000) reported a 10.4% (95% CI:  2.7, 18.1) excess rate of total daily
 2      school absences for 40 ppb increase in daily 1-h max O3 with a distributed lag of 1 to 14 days.
 3      The 12 southern California communities study by Gilliland et al. (2001) also reported larger
 4      O3-related school absences due to respiratory causes, 147% (95% CI:  6, 478) per 30 ppb
 5      increase in 8-h -avg O3 with a 30-day lag, compared to nonrespiratory causes, 61% ( 95% CI:
 6      9, 137).  The studies reported by Park et al. (2002) were analyzed using GAM with default
 7      convergence criteria and indicated a positive association with same day O3 (16% [95% CI: 12,
 8      22] per 30 ppb increase in 8-h -avg O3).  Results from the three studies listed above suggest that
 9      ambient O3 concentrations may be associated with school absenteeism, particularly illness-
10      related absences. However, the associations observed during the long lag period of two to four
11      weeks may reflect confounding by other time-varying factors or be a chance finding from an
12      exploratory analysis.  Additional studies and analysis using similar lag periods are needed to
13      more clearly delineate quantitative relationships between ambient O3 and school absences.
14
15      Field Studies on Cardiovascular Effects
16           A limited number of studies evaluated potential short term effects of air pollution on
17      cardiovascular functions. Several of these studies evaluated the effects of PM, O3 and other
18      gaseous pollutants. Two major U.S. population-based studies (Liao et al., 2004; Park et al.,
19      2005) suggested an association between short-term O3 exposure and decreased heart rate
20      variability (HRV).  Park et al., (2005) reported stronger associations of HRV with PM2 5 and O3
21      in people with ischemic heart disease and hypertension. These results are consistent with a
22      Mexico City study that observed an O3-induced HRV effect in individuals with hypertension
23      (Holguin et al., 2003). Several other studies, on the other hand, did not find any such
24      relationship, but these studies might have had limited power (e.g., low O3 concentration ranges,
25      small sample sizes) to examine the subtle effects. Studies that evaluated the relationship
26      between air pollutants and the onset of myocardial infarction (Ruidavets et al., 2005; Peters
27      et al., 2001) suggested a positive association with O3. However, due to lack of information on
28      potential confounding by PM and the limited number of studies available, additional research is
29      needed to confirm these observations.
30           Only a limited number of epidemiologic studies examined cardiovascular outcomes in
31      relation to O3 exposures. Among them, the larger population-based studies (Liao et al., 2004;

        August 2005                               8-45        DRAFT -DO NOT QUOTE OR CITE

-------
 1      Park et al., 2005; Ruidavets et al., 2005) observed suggestive evidence of an association of O3
 2      exposure with decreased HRV and increased incidence of myocardial infarctions.
 3
 4      Time-Series Analyses of Acute Exposure Effects
 5      Emergency Department Visits and Hospital Admissions
 6           Many time-series studies reviewed in the 1996 O3 AQCD indicated positive associations
 7      between O3 air pollution and increased hospital admissions.  Strong evidence establishing a
 8      correlation between O3 exposure and increased exacerbations of preexisting respiratory disease
 9      in the general public were reported at 1 h-maximum O3 concentrations <0.12 ppm.
10           Several studies published during the past decade examined temporal associations between
11      O3 exposures and emergency  department visits or hospital admissions for respiratory diseases
12      (see Table AX7-2 in Chapter 7 Annex).  One of these studies by Peel et al. (2005) reported
13      stronger and more positive associations between O3 and emergency department visits due to
14      respiratory-related diseases in the warm season.  A 3.1% (95% CI: 0.2, 6.2) excess risk in
15      asthma visits was associated with a standardized increment of 30 ppb in 8-h max O3 (see
16      Section 7.1.3.2).  This risk was significantly associated with respiratory infections, when
17      adjusted for PM10, NO2, and CO in multipollutant models. Several other U.S. and Canadian
18      studies reported positive associations between O3 concentrations and emergency department
19      visits due to respiratory causes (Figure 8-5).  However, several of the European studies observed
20      no association between O3  concentrations and emergency department visits for respiratory
21      diseases.  These inconsistent results might be partially attributable to differences in model
22      specifications and statistical methods used to evaluate seasonal patterns and potential
23      confounding by copollutants. Overall, then, the current body of evidence remains inconclusive
24      regarding ambient O3 effects on risk of emergency departments visits. Additional studies are
25      needed to establish stronger and more convincing associations between increased concentrations
26      of ambient O3 and increased risk of emergency department visits.
27           Studies of acute O3 exposure effects on respiratory disease-related hospital admissions
28      (summarized in Section 7.3.3) have considered various factors in their analyses. The hospital
29      admission data were assessed based on the type of respiratory disease (such as asthma, COPD),
30      seasonal effects (summer vs. winter,  O3 concentration and temperature), age of the study
31      population,  studies carried out in single or multiple cities, effect of confounders and lag days

        August 2005                               8-46        DRAFT -DO NOT QUOTE OR CITE

-------
             Studies from U.S. & Canada

„

fin-

40-
-
20-

00-



X Alt year
• Warm
O Cool






4



5 1*


1


i
i
i

i
i
<*

t
i

1
l*i
°
JS


i

®
® |
f



EDV
it


(

'
ft &
iil '








j


t


1
sits for Asthma
X Ally
• War


O Cooi



o

«
- (


I




i

1









J


\ •
1 	 	


Respiratory Hospitalization






> ,1
1 ,. 1
1 » • :

	 -**'i' 	 


-------
 1      or two city studies for five or more years) also suggested a positive association between increase
 2      in O3 and increased risk of hospital admissions.
 3           Many other studies reported less consistent or no associations between increases in O3
 4      concentrations and hospital admissions. A few other studies raise questions and concern about
 5      other factors in this relationship.  Despite these inconsistencies noted across the studies, the
 6      collective evidence supports the findings of significant and robust effects of O3 on respiratory
 7      hospitalization outcomes. Large multicity studies as well as several individual city studies have
 8      reported positive O3 associations with total respiratory, asthma, and COPD hospitalizations,
 9      especially in those that analyzed the O3 effect during the summer or warm season.
10           A subset of hospital admissions studies examined the effect of O3 on cardiovascular
11      outcomes (see Figure 8-5). The evidence is inconclusive on the association between O3 exposure
12      and cardiovascular hospitalizations with regard to year-round data. However, in studies that
13      adjusted for seasonal or meteorological factors, there was suggestive evidence that O3 was
14      associated with increased risk of cardiovascular hospital admissions in the warm season.
15
16      Mortality-Related to Short-term O3 Exposures
17           Due to the limited number of studies and uncertainties regarding weather model
18      specifications, no meaningful  quantitative assessment of O3-mortality associations was possible
19      in the 1996 O3 AQCD. However, newly  available large  multicity studies designed specifically to
20      examine the effect of O3 on mortality have provided much more robust and credible information.
21           Two large multicity studies from the U.S. (Bell  et al., 2004; Schwartz et al., 2005) and one
22      from Europe (Gryparis et al., 2004) specifically evaluated O3 effects on mortality and indicated
23      positive associations between increased O3 levels and  mortality. Among the positive studies,
24      risk estimates for (U.S. and Canadian) single-city studies carried out using a single-pollutant
25      model are in the range of 0.8 to 3% excess deaths per 40 ppb increase in 1-h max O3 (Figure
26      8-6), while multicity studies and meta-analyses reported a risk estimate in the range of 0.5 to 2%
27      excess risk with identified heterogeneity  (due to model specifications) across cities and studies
28      (Figure 8-6).  Models examining different lag times, observed that there was an immediate effect
29      of O3 on mortality which persisted over several days, resulting in risk effect estimates at a
30      cumulative lag of 0 to 7 days.
        August 2005                               8-48        DRAFT -DO NOT QUOTE OR CITE

-------
            10
               Studies from U.S. & Canada
@ Multicity combined
• Single city

S ™ ~

1 1 1 1
en ift in  ° T Y
1

1 1 1 1 II
tf) tn (£> fl) CD (U
         >
         *^
         15

         o
         S
         15
         ^*-»
         o
         0)
         D)
         C
         TO
         £
         O
             5-
             o-
            -5-
           -10
                           o   o
                 o  ^  ^~
                    C*5  Ifi
                 CO  O  O
                 —TOO
                 —'  CN  CM
                 8
                 CM
2

O
                           Z-   ¥l  
                                      ^_-  =   01

                                             I
                                             o
                                                o
                                                O
                               a:
                               r
Figure 8-6.  A.  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.


             B.  Combined all cause (nonaccidental) O3 excess mortality risk estimates (95% CI)

                 from recent meta-analyses per standard increment of 40 ppb in 1-h max O3 or

                 equivalent. Note that all meta-analyses, except Stieb et al. (2003), included

                 studies which used Poisson GAM with default convergence criteria.
August 2005
          8-49
DRAFT -DO NOT QUOTE OR CITE

-------
 1           Some of the epidemiological studies and, particularly, meta-analyses examined the
 2      influence of season on O3 mortality associations and indicated larger O3-mortality risk estimates
 3      in the warm season compared to the colder season. These estimates appear to be consistent with
 4      a causal association when O3 levels are high in the warm season.  This seasonal dependence
 5      of O3-mortality effects complicates the evaluation and interpretation of risk estimates from year-
 6      round data without adjustment for temporal trends (Figure 8-7).  The confounding effects of
 7      copollutants were examined by three recent meta-analyses (Bell et al., 2005; Ito et al., 2005;
 8      Levy et al., 2005).  These, as well as other multicity and single-city studies, indicated that
 9      copollutants do not appear to substantially confound O3-mortality associations.  Meta-analyses
10      on the association between cause-specific mortality and O3 levels observed larger positive
11      associations with cardiovascular mortality compared to total mortality (Figure 8-8).
12      Additional analyses carried out to examine specific population groups potentially susceptible
13      to O3-mortality effects did not clearly identify any specific group, but they did suggest that
14      severe asthmatics and elderly populations might be relatively more susceptible to O3.
15
16      8.4.3.1.2 Chronic Ozone Exposure Studies
17           There were a limited number of studies reported in the 1996 O3 AQCD that addressed
18      potential health effects of long-term ambient O3 exposures.  Several longitudinal epidemiological
19      studies carried out in the past decade evaluated the potential effects of chronic (several weeks to
20      many years) O3 exposure on lung function, respiratory symptoms, lung inflammation, asthma
21      prevalence, cancer incidence, and mortality. Based on the available data at this time, no clear
22      conclusions can be drawn now regarding the relationship between chronic O3 exposure and such
23      health outcomes.  A limited number of studies also examined the potential effects of ambient O3
24      exposure on birth defects, and these also suggest the need for additional studies and  a larger
25      database before drawing any conclusions regarding possible associations.
26           Very few studies have investigated the effects of long-term O3 exposure on incidence of
27      cancer and mortality.  Uncertainties regarding the exposure  period of relevance and
28      inconsistencies across mortality outcomes and gender raise concerns regarding plausibility.
29      The largest and most representative U.S. study, by Pope et al. (2002), observed positive but
30      nonsignificant associations between O3 exposure and all cause, cardiopulmonary, and lung
        August 2005                               8-50       DRAFT -DO NOT QUOTE OR CITE

-------
          35
            Studies from U.S. & Canada
X All year
• Warm
O Cool














"9 *- c
*+ ** P
I





1 1
v> en a) a
QJ .g  Q^
~ o
o o
O Ii
ii a.
Q. *ft
to •£
1 i
s §j

tf) Q
Ii
i § ,.
|W UJ 0
0 » «
f | 0 |
T ' ™
^B (
_



1 1 1
>><; x  c
C
I







s
n
L









i





i
T)
i


<>
C
E


r T
(
; «
A.
















)



3
7)
D


O O3only
• O3 with PM















0 i i i !






II II
Iff te

















B* 6 1
T





1 1 1
i 5=? s
, t^ 1










o
«





>







rl
L5:


in
o
o
10 ii
° a
o „_
0.1
51
II
!£ «

!M K C..
^L jf c
_ra -^- "O)"O>A
^ o a g,

. LU LU - O
1 , ww I w
(I " 4 , t «
II O •»
()





1 1 1 1 1 1 1
n. ^j P =!" ^«>o ^" c
f il f! if? 1!
X All year 1 "•
• Warm
O Cool













OS O ^ CD
0 ™ T- o, D! OJ
0) 1 i g> ^ ™ -
A ** i II
T
o ::
CM )



Mill III 1
S.i'CL CO"^ (^ o Q. n^L OQ-
O c i- V £ f-t'£ fii r- n r i^ j^
 to
t
 o
43
p
          30 -
          25 -
          20 -
          15 -
       —  10 -
       0)
       O)
       c
       IS
       o
          5 -
          o -
          -5 -
         -10 -
         -15
                S?
                CD tD
                £
                ai
                 o
                 Q
(2
                      CO o5
                      C -2
                      ro cr
Gam
                               i^|
                                'E
                                       M
                                      :
               2
               O
                               -g
                               O
                               6
       Figure 8-7. All causes (nonaccidental) O3 excess mortality risk estimates (95% CI) per
                  standardized increment (see Section 7.1.3.2). (A) by season; (B) with
                  adjustment for PM indices in all year analysis and (C) with adjustment for
                  PM indices by season. Note:  Only U.S. and Canadian studies presented.
1
2
3
4
5
6
1
cancer mortality. Thus, the current evidence is inconclusive regarding a potential relationship
between chronic O3 exposure and increased mortality risk.

8.4.3.2  Robustness of Epidemiological Associations
     In evaluating the strength of the epidemiological evidence, the magnitude of observed O3
effect estimates and their statistical significance is important; however, consideration must be
given to the precision of the effect estimates and the robustness of the effects associations.
       August 2005
                                          8-51
DRAFT -DO NOT QUOTE OR CITE

-------
                   Studies from U.S. & Canada
tu -
^
•+••*
1 30-
o
1_
"5
8 2°-
(0
o
1
(0
o 10 -
_c
0
o>
c
5 n -
-E U
O
>p
0s-

1 n
" I U






X All year
• Warm








o
O) O
JS o>
JO
I
4 	 *. 	

I




Ii
i
g -8
| I
o" 2?
(0 "j~
1 o
CO S
§ i
o o
£S £i
— *jg
"

1 _ra i X ' ' ° T c Ii | 1 I >i== < .£• JA< —I ' S5 " 0 8° >; " SJ I « Sf I i -*J -*-j O ^ Q> O O "- »9 ° | I" ^ § !> S z £ 2 8 = "5 m >i .2 FS ° ! C 03 co OT CJ . . ; S < 3 ff g = • - jy ^ji- j= co ^ -- § fl _: ©CO C^ g o . . ro *- >- c a slS | c £ §. S S- J, I » 7T -1 >, S ' 1 1 s I ? 3 Q- 0 3 ° cv 2 o §,0 (0 in > ii? (U Xi *= CD t S o ±i « E 3 CD O CO 'i_ C 35- * D 3) 5 a-analysis of 25 estimates U.S. and non-U.S. studies ^ "OJ CO 6 . . ^ § §- * * -ffi Figure 8-8. Ozone-associated percent change (95% CI) in cardiovascular risk estimates per standardized increment (see Section 7.1.3.2). Analyses are for all ages. Only studies for U.S. and Canada presented. *Analyses using GAM default convergency criteria. 1 Examining the robustness of the associations includes evaluating the impact of alternative 2 models, model specifications for temporal trends and meteorological factors, and potential 3 confounding by copollutants. Also of interest are issues related to exposure assessment and 4 measurement error. A detailed discussion on each of these topics can be found in Chapter 7 5 (Sections 7-1 and 7-6). The following sections focus on the extent to which the current 6 epidemiological findings can be considered robust. August 2005 8-52 DRAFT -DO NOT QUOTE OR CITE


-------
 1      8.4.3.2.1  Exposure Issues: Ambient versus Personal
 2           In many air pollution epidemiologic studies, especially time-series studies utilizing
 3      administrative data on mortality and hospitalization outcomes, data from central ambient
 4      monitoring sites are generally used as the estimate of exposure.  Personal exposures of individual
 5      study participants generally are not directly observed in epidemiologic studies. The relationship
 6      between ambient O3 concentrations and personal O3 exposure levels varies, depending on factors
 7      such as time spent outdoors, ventilation conditions, personal factors, and air quality indices.
 8      There is suggestive evidence that ambient O3 concentrations from central monitors may serve as
 9      valid surrogate measures for aggregate personal O3 exposures in time-series studies. However,
10      ambient concentrations generally overestimate true personal O3 exposures.  Thus, use of ambient
11      concentrations in risk calculations will likely result in effect estimates that are biased towards the
12      null, resulting in biased descriptions of underlying concentration-response relationships. These
13      effect estimates, though conservative from a testing perspective, must be evaluated and used
14      with caution, as they may lead to  an underestimation of the overall impact of air pollution on
15      health effects.
16
17      8.4.3.2.2  Confounding by Temporal Trends and Meteorologic Effects
18           The effect of seasonal differences in the health outcomes and O3 exposure levels was
19      recognized in the 1996 O3 AQCD. This issue is discussed in detail in Section 7.6.5 of this
20      document. Two important factors, i.e., temporal trends and meteorological factors must be
21      considered in evaluating O3 health effects estimates. In the U.S. 95 communities study (Bell
22      et al., 2004), sensitivity analyses indicated that the O3 risk estimates were robust to tripling the
23      degrees of freedom for smoothing terms used to control for temporal trends. In a case-crossover
24      study by Schwartz (2004), the O3-mortality risk estimates from an analysis using nonlinear
25      regression splines to control for temperature were similar to those from an analysis that matched
26      on temperature, indicating that the effect estimates were not sensitive to methods used to control
27      confounding by temperature.
28           Analysis of O3 health effects is further complicated in view of the fact that the relationship
29      of O3 with temperature and with other pollutants appears to change across seasons. As shown in
30      Figures 8-5,  8-7 and 8-8, the O3 effect estimates from warm season data were consistently larger
31      compared to those calculated using all-year data and cool-season data.  In a study of daily

        August 2005                               8-53       DRAFT -DO NOT QUOTE OR CITE

-------
 1      hospital admissions (Burnett et al., 2001), season-stratified analyses appeared to effectively
 2      control confounding by season.
 3           In summary, adjusting for temporal trends and meteorological factors is critical to
 4      obtaining meaningful O3-effect estimates. Analyses have found that confounding by seasonal
 5      variability is controlled effectively by stratifying the data by season. Mortality and morbidity
 6      effect estimates computed using year-round data need to be interpreted with caution.
 7
 8      8.4.3.2.3  Assessment of Confounding by Copollutants
 9           The presence and influence of PM and other gaseous copollutants have to be considered in
10      assessing O3-health effects associations found by observational studies. The potential for
11      copollutant confounding in the epidemiological time-series  studies was assessed in some detail
12      in Section 7.6.6. Multipollutant modeling is the most common method used to test for potential
13      confounding in epidemiological studies; however, interpretation of the results is often
14      complicated by the high degree of correlation among air pollutants. Across the various health
15      outcomes, O3 effects were generally not confounded by PM and other gaseous copollutants.
16      The O3 effects on lung function, respiratory symptoms, respiratory hospitalizations,  and
17      mortality were robust to PM-adjustment in all year and warm season only analyses.
18           The O3 mortality risk estimates from two-pollutant models adjusted for PM are presented
19      in Figure 8-7 (U.S. and Canadian studies only). In the two multicity studies analyzed here, the
20      addition of PM10 did not substantially change the risk estimates (Samet et al., 2000; Dominci
21      et al., 2003; Schwartz, 2004).  The O3-mortality effects in single-city studies also were robust
22      after adjusting for PM10 indices, both in all-year and season-stratified analyses data.
23           In summary, assessing the health effects attributable to O3 is very challenging, given the
24      high covariation among the copollutants and the limitations in the statistical methodology to
25      assess independent effects of such correlated variables. Definitive partitioning out of the
26      individual pollutant-specific health  outcomes from among an ambient mixture of multiple
27      components is very difficult due to the dynamic nature of their interactions over time. However,
28      the new time-series studies that made an exhaustive survey using populations from multiple U.S.
29      cities do provide substantial epidemiologic evidence indicating that associations for O3 with
30      mortality and morbidity are robust to confounding by copollutants.
31

        August 2005                               8-54       DRAFT -DO NOT QUOTE OR CITE

-------
 1      8.4.3.3  Lag Period between Ozone Exposure and Health Response
 2           The lag times between cause and effect depend on underlying biological mechanisms
 3      involved in the processes. Different lag periods are appropriate for assessing different health
 4      outcomes. As discussed in Section 7.6.4, examining longer lag periods may be needed to
 5      understand more fully the O3-related health outcomes. The most significant associations
 6      between O3 concentrations and mortality and respiratory hospitalization were observed with
 7      0-day and 1-day lags. In the 95 U.S. communities study (Bell et al., 2004) and the related U.S.
 8      study of the 19 largest cities (Huang et al., 2005), the risk estimated over multiple days
 9      (cumulative lag of 0 to 6 days) using distributed lag models indicated a strong effect of O3 on
10      mortality.  It should be noted that when there is a pattern of effects across lag periods, selecting a
11      single-day lag effect estimate may underestimate the overall effect size and not fully capture the
12      risk distributed over adjacent days. Longer averaging periods may aid in characterizing
13      cumulative O3-related effects over several days; however, interpreting these results may not be
14      straightforward.
15
16      8.4.3.4  Concentration-Response Functions and Threshold
17           Ozone concentration-response relationships have been explored in several studies with
18      various health outcomes, including mortality, hospitalizations, emergency department visits,
19      lung function, and respiratory symptoms. While some studies found no threshold for O3 health
20      effects, others have found that a very low-level threshold may be present.  A study by Kim et al.
21      (2004) specifically  examined the presence of a threshold in O3-mortality effects  in Seoul, Korea
22      by analyzing data using a log linear GAM (linear model), a cubic natural spline  model (nonlinear
23      model), and a B-mode splined model (threshold model).  An estimated threshold value of 47 ppb
24      was observed for 1-h daily max O3. This study further observed that if a threshold truly exists,
25      the use of log-linear models may underestimate the O3 effect on mortality at levels above the
26      threshold.
27           It should be noted that exposure measurement error may reduce the ability to detect a
28      threshold in O3 population studies that used ambient O3 concentrations as an indicator of
29      personal ambient exposure.  In addition, due to the variability in individual  sensitivities, a
30      threshold may not be seen at the population level. The limited evidence suggests that if there is a
        August 2005                               8-55       DRAFT -DO NOT QUOTE OR CITE

-------
 1      threshold level in O3 health effects, it is likely near the lower limit of ambient O3 concentrations
 2      in the United States.
 3
 4      8.4.3.5  Consistency of Findings Across Epidemiologic Studies
 5           Most of the multicity and meta-analyses studies consistently found positive associations
 6      between O3 and mortality. Generally consistent O3 effects on hospitalizations and various
 7      respiratory health outcomes also were found.  Ozone concentrations tend to be spatially variable
 8      in urban areas.  The geographic variability in O3 concentrations and personal exposures may
 9      contribute to the heterogeneity in observed O3 health effects. The degree of influence of the
10      geographic variability on heterogeneity in effects tend to vary by  study, as study design affects
11      different aspects of exposure (e.g., time period and duration of exposure).  In addition, some of
12      the observed heterogeneity of O3 effects may be partially attributable to the use of centrally-
13      located ambient monitors to assess exposure.  There may be differences in relative personal
14      exposures to O3 by region due to varying factors, such as use of air conditioning and activity
15      patterns, that affect the relationship between personal exposure and ambient concentrations.
16           Among the field studies, various respiratory health outcomes were examined, including
17      PEF, other spirometric parameters, respiratory symptoms, and medication use.  One field study
18      investigated the O3 effect in asthmatic children living in eight urban cities in the U.S. (Mortimer
19      et al., 2002). In the analysis pooling data from all eight cities, O3  was associated with a
20      decrement in morning PEF for a 5-day cumulative lag period. The percent changes in PEF were
21      quite homogenous, with values ranging from -1.08% for Washington, DC to -1.71% for
22      St. Louis. Ozone also was associated with an increased incidence of morning symptoms in the
23      pooled analysis (Mortimer et al., 2002).
24           More than 80% of the O3-mortality estimates from the various studies conducted in North
25      America, South America, Europe, and Australia fell between 0.5 and 5% excess risk per 40 ppb
26      increase in 1-h max O3 using year-round data. In general, the O3-mortality estimates were
27      greater when using summer only data compared to year-round data. Though not all statistically
28      significant, most of the O3-mortality estimates were greater than zero, indicating a positive
29      relationship between O3 exposure and mortality.  Three recent mortality meta-analyses that
30      included both U.S. and non-U.S. studies found consistent all-year combined point estimates of
31      1.6 to 1.8% excess risk per 40 ppb increase in 1-h max O3. The O3 risk estimates from the

        August 2005                               8-56       DRAFT -DO NOT QUOTE OR CITE

-------
 1      numerous hospitalization and emergency department visit studies were generally larger in
 2      magnitude and more variable from study to study compared to the mortality studies.
 3           Because differences in study design, population, and data analysis may affect risk
 4      estimates, one study investigated the geographic heterogeneity of O3 effects in multiple cities
 5      using standardized methods. In the pooled analysis of 95 U.S. communities using all available
 6      data, intercity heterogeneity was observed among the 95 communities, which the authors noted
 7      as plausible given the city-specific differences in pollution characteristics, the use of air
 8      conditioning, time-activity patterns, and socioeconomic factors (Bell et al., 2004).
 9           Overall, the epidemiological studies indicate that there are associations between acute O3
10      exposures and morbidity and mortality outcomes in numerous locations across the United States.
11      In general, fairly consistent O3 effect estimates were observed for the various health outcomes,
12      including pulmonary function, symptoms, hospitalization, and mortality.
13
14      8.4.3.6   Summary and Conclusions for Epidemiology Findings
15           Discussions presented in the previous sections evaluated the merits of the epidemiologic
16      studies to derive judgments about potential causal relationships between O3 exposures and health
17      outcomes. These evaluations were carried out in the context of the criteria listed in Section
18      8.2.2. Information with regard to one of the criteria, i.e., coherence and biological  plausibility, is
19      discussed in the next section, which undertakes to provide an integrated analysis of the
20      biological evidence from human and animal toxicology studies with the epidemiologic evidence.
21           The results from the new field/panel studies evaluated in this document provide additional
22      evidence for likely causal relationships being reflected by significant associations between
23      acute O3 exposure and decrements in lung function. Several new studies also associated acute O3
24      exposure with increased respiratory symptoms and use of asthma medication in children and,
25      in some cases, adults.  New population based time-series studies also indicated a positive
26      association between acute O3 exposure and respiratory morbidity indexed by hospital admissions
27      and emergency visits, especially in season-stratified data. The results from large multicity
28      studies and several meta-analyses consistently suggest an elevated risk of mortality for acute
29      exposure to O3.  Additional analyses evaluating the potential susceptibility of individuals with
30      preexisting cardiovascular disease is rather limited. Analysis of the data from chronic mortality
31      and morbidity studies indicate possible associations between O3 and seasonal changes in lung

        August 2005                              8-57        DRAFT -DO NOT QUOTE OR CITE

-------
 1      function; but, overall, the strength of the evidence does not allow establishment of a likely causal
 2      relationship between chronic O3 exposure and these health outcomes.
 3           Issues regarding strengths of models used in air pollution epidemiology were carefully
 4      considered. There have been improvements in the modeling to adjust for potential confounding
 5      variables, including temporal trends, meteorological factors, and copollutants.  However, more
 6      sensitivity analyses would still be useful to examine the extent of adequate adjustment for
 7      confounding by these factors. Results from multipollutant models indicate that copollutants,
 8      e.g., PM, generally do not confound the association between O3 and acute health outcomes,
 9      suggesting an independent effect of O3. The limited evidence suggests that if there is a threshold
10      level in O3 health effects, it is likely near the lower limit of U.S. ambient O3 concentrations.
11           In conclusion, the epidemiological evidence continues to support likely causal associations
12      between acute ambient O3 exposures and increased risk of acute respiratory morbidity and
13      mortality, based on the assessment of strength, robustness, and consistency of results reported
14      from numerous studies reviewed in Chapter 7.  There is a lack, however, of sufficient evidence
15      by which to convincingly establish a positive association between chronic O3 exposure and
16      increased respiratory morbidity and mortality.  Additional investigations are needed to further
17      understand the health effects resulting from long-term O3  exposure.
18
19
20      8.5  BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE FOR
21           OZONE-RELATED HEALTH EFFECTS
22           This section is organized to integrate epidemiologic studies with toxicologic and
23      mechanistic information obtained from controlled human  exposure studies and animal
24      toxicology studies for the two major health endpoints, morbidity and mortality reported to be
25      associated with either short- or  long- term exposure to ambient O3. Morbidity associations have
26      been subdivided into (a) school absenteeism, (b) emergency department visits for asthma, and
27      (c) hospitalizations due to respiratory and cardiovascular illnesses. Mortality associations were
28      also critically evaluated to understand risk estimates for total nonaccidental mortality and
29      mortality in specific susceptible populations. The discussion in each subsection concisely
30      summarizes pertinent key information and then presents the plausibility of effects  being
31      reasonably attributed  to the conclusions derived for the endpoint assessed. To facilitate an easy
32      discussion and to recapitulate various biological endpoints that have been investigated in human

        August 2005                               8-58       DRAFT -DO NOT QUOTE OR CITE

-------
 1      and animal toxicology studies (see Chapters 4, 5, and 6), the first subsection addresses the
 2      plausible interpretative assessments from the salient observations from experimental studies.
 3           Several criteria listed in Section 8.4.2 are used in evaluating the available scientific support
 4      for conclusions regarding potential causal relationships between O3 exposure and specific types
 5      of health outcomes. In addition to those criteria addressed in the preceding discussion of
 6      epidemiological evidence, certain other critical evaluation measures must be considered to
 7      ensure that these observations are biologically relevant and consistent with experimentally
 8      demonstrated biological mechanisms of action. For this assessment, the ensuing discussion on
 9      biological plausibility  and coherence considers (a) the extent to which available epidemiological
10      evidence logically ties to a range of relevant health endpoints (from cardiopulmonary
11      physiological changes to morbidity to mortality) and (b) whether available toxicological and
12      biochemical evidence  supports plausible causal relationships for the observed epidemiological
13      associations.
14           The ensuing discussion on biological plausibility and  coherence also considers the
15      following criteria for integrated assessment:  (a) adequateness of the statistical power of the
16      epidemiological studies to establish evidence for associations, (b)  the location (urban vs. rural),
17      (c) seasonal pattern vs. all year O3 levels, (d) socioeconomic status of the  population, and (e) the
18      lag days (used  in the epidemiologic analysis), considered in relation to the time course of likely
19      biological mechanisms potentially implicated in the process.
20
21      Animal-to-Human Extrapolation Issues
22           The physiological and biochemical observations reported in  Table 8-1 represent the
23      knowledge base available from toxicological studies in humans and animals that underlie the
24      biological alterations that govern acute O3-induced health effects.  This table is generated from
25      the experimental database (see Annexes for Chapters 5 and 6 for experimental details) that
26      utilized exposure regimens of varied concentration and duration that are environmentally
27      relevant.  As noted in the  earlier section, most of the observed acute O3 effects are transient and
28      attenuate over time. However, the time-line for resolution of many of these physiological and
29      biological parameters in normal and human subjects with underlying cardiopulmonary diseases
30      follow different profiles as presented in Figure 8-9.  Alterations in the cellular and molecular
31      profiles observed in human airway epithelium upon acute exposure to O3  evolve over time

        August 2005                                8-59        DRAFT -DO NOT QUOTE OR CITE

-------
                  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:
 ROS
 Host Defense:
 Lung injury:
 Morphology

 Susceptibility:
 Cardiovascular Changes:
1 FEVj
T Frequency of breathing
(rapid, shallow)
i inspiratory capacity
(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,
Inter individual variability
Disease status
Polymorphism in certain genes
being recognized

Impairment in arterial O2 transfer
Ventilation-perfusion mismatch
(suggesting potential arterial
vasoconstriction)
       1 FEVj
       T Frequency of breathing
         (rapid, shallow)
       I inspiratory capacity
       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 variability (HRV)
      1 core body temperature
      T ANF
      Role for PAF indicated
      increased pulmonary vascular
      resistance
 1 Controlled chamber exposure studies in human volunteers were carried out for a duration of 1-6.6 h with O3
   concentration in the range of 0.08-0.4 ppm with intermittent exercise.
 2 Data on some of the biochemical parameters were obtained from in vitro studies using cells recovered
   from BALF.
 3 Responses were observed in animal toxicology studies with exposure for a duration of 2-72 h with O3
   concentration in the range of 0.1-2.0 ppm.
 4 Various species (mice, rat, guinea pigs and rabbit) and strains.
August 2005
                   8-60
DRAFT -DO NOT QUOTE OR CITE

-------
               Resolution Time-Line for Acute Ozone-Induced Physiological and Biochemical Responses in Humans
to
o
o
oo
H

6
o


o
H

O

O
H
W

O


O
HH
H
W
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
•4 	 ^ Complete
          Figure 8-9. Resolution time-line for the physiological and biochemical parameters are derived from studies

                    reported in Chapter 6 and Chapter 6 Annex.

-------
1
2
3
4
(Figure 8-10), and the knowledge of this profile is valuable in assessing biological plausibility to
integrate across evidence for various health endpoints.
                    Postulated Cellular and Molecular Changes in Human Airway
                             Epithelial Cells on Acute Exposure to Ozone
               Response Time

               Immediate 0-2 h
               Early 2-24 h
               (Neutrophil infiltration)
               Late (12-24 h)
               (Eosinophil/monocyte
               infiltration)
                                      Chemical reaction with ELF and epithelial cell membrane,
                                      Generation of ozonation products, lipid peroxides
                                      Lipid ozonation products
                                      Pro-inflammatory mediators (neutrophil chemotaxins,
                                      Anti-inflammatory mediators (prostanoids)
                                      Cytokines, proteases
                                      Lipid ozonation products
                                      Increase in pro-inflammatory mediators (monocyte chemotaxins,
                                      Decrease in anti-inflammatory mediators (prostanoids)
                                      Release of cytokines
                                      Increased expression of intracellular adhesion molecules
                                      Increased synthesis of collagen, fibronectin
                                      Release of leukocyte proteinase inhibitors
                                      Increased synthesis of antioxidants (SOD, GSH, catalase)
       Figure 8-10.  O3-induced cellular and molecular changes and their evolution depicted here
                     is derived from the data reported in Leikauf et al. (1995) and Mudway and
                     Kelly (2000).
1           Basic similarities in physiological, biochemical, and pathological processes that exist

2      between human and animal species are derived from the high degree of genome sequence

3      homology that exists across species. This homology reinforces the significance of knowledge

4      gained on the initiation, progression and treatment regimes for various disease processes across

5      animal species. This homology is also apparent in acute O3-induced effects, especially on the

6      respiratory tract of human and animal species as presented in Table 8-1 and Figures 8-9 and

7      8-10. The commonality of phenomenon observed in humans and rats with regard to respiratory

8      system effects (in terms of spirometry, ventilatory response,  host defense and inflammation) and

9      their attenuation adds strength to animal-human extrapolations. Such similarities observed at
       August 2005
                                             8-62
DRAFT -DO NOT QUOTE OR CITE

-------
 1      higher levels of cellular organization (neutrophilic inflammation, macrophage phagocytosis
 2      processes) have increased the value and importance of animal studies in generating important
 3      data that is impossible to collect in human studies but which may corroborate both clinical and
 4      epidemiologic studies.
 5           Quantitative extrapolation involves a combination of dosimetry, end point homology and
 6      species sensitivity, particularly in the case of exposure and health outcome analyses. However,
 7      extrapolation models have not been completely validated and, therefore, uncertainties do exist.
 8      Based on inflammatory markers in BALF, a 2 ppm O3 exposure in nonexercising rats
 9      approximates to a 0.4 ppm exposure in exercising humans despite species-specific differences
10      (Hatch et al., 1994).  This observation lends support to the use of some of the animal toxicology
11      data derived from relatively high O3 concentration exposure regimens in understanding putative
12      molecular changes associated with acute O3 exposure in humans. Similarly, the presence of
13      apparent O3-induced lesions in animals from chronic  O3 exposure studies (12 to 24 months)
14      indicate functional defects that may potentially provide a means to facilitate more direct
15      assessments of long-term health outcomes in humans.
16
17      8.5.1  Acute Ozone Exposure-Induced Health Effects
18           As noted  in Section 8.4.2, several new epidemiologic (field/panel) studies show positive
19      associations between short-term exposure to ambient O3 and human health effects.  These health
20      effects as evaluated include school absenteeism,  decline in lung function, increased use of
21      asthma medication, and increased hospitalization, especially among individuals with asthma or
22      certain known cardiopulmonary or cardiovascular diseases (see Chapter 7).  The patterns of
23      physiological and biochemical alterations reviewed earlier  (see Figures 8-9, 8-10 and Table 8-2)
24      tend to support certain hypotheses regarding underlying pathological mechanisms in the
25      development of respiratory effects reported in the epidemiologic studies. Some of these
26      mechanisms (see  Table 8-2) include (a) decrements in lung function (capacities and volume), (b)
27      bronchoconstriction, (c) increased airway responsiveness, (d) airway inflammation, (e) epithelial
28      injury, (f) immune system activation, (g) host defense and (h) sensitivity of an individual such as
29      age, genetic susceptibility and the extent of tolerance resulting from previous exposures.  The
30      time sequence,  magnitude, and overlap of these complex events, both in terms of development
        August 2005                              8-63       DRAFT -DO NOT QUOTE OR CITE

-------
 1      and recovery (see Figure 8-9 and 8-10), indicate the difficulties associated with the interpretation
 2      of biological plausibility associated with the cardiopulmonary health effects.
 3
 4      Respiratory Health Effects
 5           Controlled human exposure studies have clearly demonstrated the following three types of
 6      respiratory responses to acute O3 exposures: (1) irritative cough and substernal chest pain upon
 7      inspiration; (2) decrements in FVC and FEVj due to decreased inspiratory capacity rather than
 8      airways obstruction and (3) neutrophilic inflammation of the respiratory tract. Susceptibility or
 9      sensitivity to these effects were observed even among a carefully selected homogeneous study
10      population. The sources of this heterogeneity are uncertain. As discussed in the earlier section,
11      changes in baseline levels of various responses, the lag in the recovery phase and the role of
12      residual defects in these mechanisms in hyperresponsive individuals suggest potential for
13      increased health effects in cardiopulmonary compromised individuals such as people with
14      asthma, COPD and cardiovascular diseases. Recent research has emphasized further
15      characterization of the mechanisms and consequences of O3-induced pulmonary function and
16      inflammatory responses.  In addition, animal studies indicate morphological changes associated
17      with acute O3 exposures.
18           Ozone-induced altered breathing patterns (rapid shallow breathing) observed in controlled
19      human exposure studies and animals occur without significantly affecting minute ventilation,
20      suggesting compensatory changes in breathing pattern.  Such a shift in breathing pattern
21      diminishes deep lung penetration of O3.  Breathing pattern is modulated by changes in peripheral
22      mechanisms, such as direct or indirect stimulation of lung receptors and bronchial C-fibers. The
23      activity of these afferents is integrated with input from sensory pathways and thus determines the
24      depth and frequency of breathing. Stimulation of bronchial C-fibers along with inhibition of
25      inspiration through local axon reflexes can induce neurogenic inflammation via tachykinins and
26      other proinflammatory neuropetides. Ozone-induced increases in the levels of neuropeptide
27      substance P observed in the BALF of human subjects suggests potential neurogenic involvement
28      in vascular permeability, plasma protein extravasation, bronchoconstriction and mucus secretion
29      (Solway and Leff, 1991).  Similar neurogenic involvement due to vagally mediated stimulation
30      of C-fibers seen in animal toxicology studies support O3-induced bronchial hyperresponsiveness
31      observed in humans.

        August 2005                              8-64       DRAFT -DO NOT QUOTE OR CITE

-------
 1           An extensive database of animal, human, and in vitro studies supports the conclusion
 2      that O3 interacts with airway epithelial cell membranes and lining fluid to form lipid ozonation
 3      products and ROS.  These reactive products initiate a cascade of events leading to oxidative
 4      stress, injury, inflammation, airway epithelial damage and increased alveolar permeability to
 5      vascular fluids.  Inflammation is the outcome of host response to injury and usually resolves
 6      completely.  Continued irritant challenge may evolve into a chronic inflammatory state with
 7      simultaneous alterations in lung structure and function, leading to diseases  such as fibrosis and
 8      emphysema. Continued inflammation can also alter the lung's ability to respond to infectious
 9      agents, allergens, and toxins.  Acute inflammatory responses to O3 exposure are well documented
10      in humans and animals. As presented in Figure 8-10, the early inflammatory response
11      to O3-induced lung injury is apparent in human subjects within 3 h postexposure.  This initial
12      neutrophilic inflammatory response phase is characterized by increases in PMNs in the BALF
13      along with increased levels of inflammatory mediators such as interleukins, prostaglandins and
14      complement component C3a. In vitro studies using human and animal lung cell culture systems
15      have further examined  the involvement of various inflammatory mediators  and in some instances
16      their downstream signaling pathways. The late inflammatory phase in the lung is characterized
17      by increased levels of monocytes and eosinophils and respective mediators such as cytokines,
18      leukotrines, proteinases, and ROS.
19           Disruption of the lung's blood barrier by O3 resulting in vascular permeability changes
20      and plasma protein extravasation.  BALF analysis on plasma influx markers such as albumin,
21      proteins,  immunoglobulins, and epithelial cell damage markers such as LDH indicate
22      O3-induced lung epithelial injury.  Ozone-induced lung injury and subsequent disruption of the
23      airway epithelial barrier has been implicated in increased mucociliary clearance of particles
24      observed in controlled  human studies. Analogously, animal toxicology studies (see Chapter 5)
25      have reported increased mortality to bacterial and viral infections subsequent to O3 exposure and
26      also increased clearance of particles.
27           Controlled O3 exposure studies of healthy humans have indicated a large degree of
28      intersubject variability. The spirometric and symptomatic responses are highly reproducible
29      within the subject; but, within a group, pulmonary function measurements varied from -4% to
30      56%.  These are likely  also to be genetic function, but as yet this factor remains of uncertain
31      importance.  Analysis of personal  characteristics such as age, height, smoking history and

        August 2005                              8-65        DRAFT -DO NOT QUOTE OR CITE

-------
 1      allergies indicated age as an important contributor.  Based on FEVj measurements as criteria,
 2      young adults (18-25 yrs) were found to be more sensitive to O3 than children or adults.
 3      Sensitivity to O3 had been found to decline with older age (>60 yrs), but it should be noted that
 4      the baseline values for various measurable pulmonary functions are different in this group of
 5      population.
 6
 7      Cardiovascular Health Effects
 8           There exist few experimental studies in animals and humans that have investigated
 9      potential cardiovascular effects  with acute O3 exposures.  Ozone induces lung injury,
10      inflammation, and impaired mucociliary clearance with a host of associated biochemical changes
11      all leading to increased lung epithelial permeability. As discussed in the Section 5.4.2 the
12      generation of lipid ozonation products and ROS in lung tissue can influence the pulmonary
13      hemodynamics and ultimately cardiovascular system. Recent reports of interaction  of O3 with
14      cholesterol in the lung surfactant and the generation of highly reactive products such as
15      oxysterols and p-epoxide have indicated a role for cardiovascular effects and atherosclerosis
16      (Pulfer and Murphy, 2004).  Ozone-induced changes in heart rate variability, edema of heart
17      tissue, and increased tissue and  serum levels of ANF observed in animal toxicology studies lend
18      support to potential cardiovascular effects of acute O3 exposures. Such effects resulting from
19      stimulation of airway irritant receptors, c-fiber activation, may result from either local or central
20      nervous system involvement. The observation of O3 -induced changes in the alveolar-arterial
21      oxygen transfer in controlled human exposure studies on subjects with hypertension indicates
22      potential complex ANF effects that need to be investigated further.
23
24      Coherence Between Epidemiologic and Experimental Evidence for Acute Respiratory
25      and Cardiovascular Effects
26           Epidemiologic studies, indicate a positive association between exposure to ambient O3 and
27      declines in lung function in children and those with cardiopulmonary diseases such  as asthma.
28      Meta-analyses of children in summer camp studies (Kinney et al., 1996) and a multicity study by
29      Mortimer et al (2002) supports the earlier observations that children and asthmatics are
30      particularly susceptible to ambient O3.  This association based on decrements in lung function
31      and exacerbating pulmonary disease symptoms suggests that O3 exposures may result  in
        August 2005                               8-66       DRAFT -DO NOT QUOTE OR CITE

-------
 1      increased use of medication in children and asthmatics.  Studies that evaluated relationships
 2      between exposure to ambient O3 and school absences in children provide corroborative positive
 3      associations. An increased incidence of disease exacerbations in people with cardiovascular
 4      diseases could not clearly be established.
 5           Increased incidence of emergency department visits due to specific respiratory illness (e.g.,
 6      asthma) and hospitalization due to specific causes (e.g., respiratory or cardiovascular disease
 7      exacerbation) reported in various studies discussed in Chapter 7 (depicted in Figure 8-1 for
 8      studies from the United States and Canada) suggest a causal association supported by animal
 9      toxicology data.  This association becomes more apparent when the data are analyzed for the
10      influence of seasonal differences in ambient O3 levels.  Several controlled clinical studies
11      reviewed in the 1996 O3 AQCD  on atopic and asthmatic subjects have not shown enhanced
12      responsiveness to acute O3  exposure compared to healthy subjects. The majority of the newer
13      studies reviewed in Chapter 6 continue to suggest that asthmatics are as sensitive as, if not more
14      sensitive than, normal subjects in manifesting O3-induced pulmonary function decrements.
15           Ozone-induced increases in neutrophils, protein, and IL-8 were found to be significantly
16      higher in the BALF from asthmatics compared to healthy subjects. Similarly, subjects with
17      allergic asthma exhibited increased airway responsiveness to inhaled allergens upon acute O3
18      exposure. Consistent with these changes it is suggested that asthmatics will be more sensitive to
19      small airway effects of ambient O3 Asthmatics present a differential response profile for the
20      cellular, molecular, and biochemical parameters (Figure 8-10) in response to acute O3 exposure.
21      Increases in O3-induced nonspecific airway responsiveness incidence and duration could have
22      important clinical implications for asthmatics.
23           Bronchial constriction following provocation with allergens presents a two-phase response.
24      The early response is mediated by release of histamine and leukotrienes that leads to contraction
25      of smooth muscle cells in the bronchi, narrowing the lumen and decreasing the airflow.
26      In asthmatics, these mediators also attract accumulation eosinophils, followed by production
27      of mucus and a late-phase bronchial constriction and reduced airflow. Holz et al (2002) reported
28      an early phase response in subjects with rhinitis after a consecutive 4-day exposure to 0.125 ppm
29      O3 that resulted in a clinically relevant (>20%) decrease in FEVj. Allergen challenge in mild
30      asthmatics 24 h postexposure to  0.27 ppm O3 for 2 h had been found to significantly increase
        August 2005                               8-67       DRAFT -DO NOT QUOTE OR CITE

-------
 1      eosinophil counts in BALF compared to healthy subjects (Vagaggini et al., 2002). Epithelial
 2      cells from the mucosal biopsies of allergic asthmatics indicated significant increase in the
 3      expression of IL-5, IL-8 and GM-CSF suggesting increased neutrophilic inflammation compared
 4      to healthy subjects (Bosson et al., 2003).  In vitro exposure studies (0.1 ppm O3) of nasal
 5      epithelial cells from atopic asthma patients were found to release significantly greater amounts
 6      of neuropeptides, neurokinin A and Substance P, suggesting activation of neurogenic
 7      inflammation (Schierhorn et al., 1999). Collectively, these observations suggest that O3
 8      exposure may exacerbate pre-existing allergic asthma. People with allergic asthma may
 9      represent a segment of the population reported to have increased symptoms of respiratory illness
10      exacerbations, emergency department visits, and hospital admissions in epidemiologic studies.
11           Recent population time-series studies (Figure 8-5) have also indicated a potential
12      association between acute O3 exposure and cardiovascular hospitalization. Additional well-
13      designed time-series studies are needed to evaluate cardiovascular morbidity  more specifically
14      associated with acute O3 exposure. The intimate hemodynamic and neurohumoral relationships,
15      and potential cardiac consequences of pulmonary insults are well recognized. Two important
16      observations in human clinical studies:  (1) O3-induced impairment in  alveolar-arterial  oxygen
17      transfer (Gong et al., 1998) and (2) O3-induced ventilation-perfusion mismatch (Foster et al.,
18      1993,  1997) are consistent with potential cardiovascular impacts of O3. If such relationships are
19      validated, they will aid in our understanding the role of O3-induced reductions in gas exchange
20      and oxygen saturation in COPD patients with already compromised gas exchange process.
21      Cardiovascular disease conditions and COPD are most common among old age groups.
22      Thus,  age-associated pulmonary function deficiencies in older people would add an additional
23      burden with respect to O3-induced effects. The recent observations of air pollution-induced
24      vasoconstriction in controlled human exposure studies by Brook et al.  (2002) suggest a
25      possible role for O3.
26           Animal toxicology studies indicate acute O3-induced microvascular leakage and
27      subsequent edema of airways in guinea pigs (Inoue et al., 1997) and increased baseline values
28      for total vascular resistance in rabbit pulmonary vessels in ex vivo studies (Delaunois et al.,
29      1998). These observations support the possibility of potential cardiovascular effects.
30
31

        August 2005                               8-68       DRAFT -DO NOT QUOTE OR CITE

-------
 1      8.5.2   Chronic O3 Exposure-Induced Health Effects
 2           The effects of chronic O3 exposure in humans have been addressed primarily with cross-
 3      sectional epidemiologic studies. Due to lack of precise information on exposure, the possibility
 4      of selection bias and the difficulty of controlling for confounders, these findings are
 5      inconclusive.  However, several new longitudinal epidemiological studies have evaluated the
 6      potential associations between chronic exposure to O3 and morbidity and mortality (see Section
 7      7.5).  These studies suggest that long-term exposure may be related to changes in lung function,
 8      increased incidence of asthma, mortality, and, possibly, lung cancer.  However, based on
 9      available evidence, no definitive relationship could be established between chronic O3 exposure
10      and these health outcomes.  There are no data available from controlled human chamber studies
11      that evaluated chronic exposure regimens.
12           The lack of adequate data from epidemiologic and clinical studies in human has directed
13      attention to the results from chronic exposure studies in animals.  Earlier chronic animal studies
14      employed traditional exposure designs using chronic  stable exposures. Later studies have
15      attempted to incorporate design features that mimic diurnal and seasonal pattern of O3 exposure
16      and realistic exposure concentrations.  Studies on monkeys that compared these two designs
17      reported increased airway pathology with the latter design. Persistent and irreversible effects
18      observed in chronic animal toxicology studies indicate the need for complementary human data
19      from  epidemiologic studies.
20           Animal toxicology data provide a clearer picture indicating that long-term O3 exposure at
21      levels found in the ambient air may have lasting effects.  Chronic exposure studies in animals
22      have  reported biochemical and morphological changes suggestive of irreversible long-term O3
23      impacts  on the lung.  Some of the studies in rats (0.5-1.0  ppm O3 for 6 h/day) for 20 months and
24      monkeys (0.61 ppm) for one year noted increased deposition of collagen and thickening of the
25      CAR of the deep lung.  Differences in this degree of lung damage have been observed with
26      continuous exposure and seasonal pattern.  A long term study of infant rhesus monkeys exposed
27      to simulated seasonal O3 (0.5 ppm 8 h/day for 5 days  every 14 days for 11 episodes) resulted in
28      remodeling in the distal airways, abnormalities in tracheal basement membrane, accumulation of
29      eosinophils in conducting airways and decrements in airway innervation.  Earlier studies in rats
30      following seasonal episodic profiles also showed small, but significant, decrements in lung
31      function that were consistent with focal fibrinogenesis in the proximal alveolar region.  On the

        August 2005                              8-69       DRAFT -DO NOT QUOTE OR CITE

-------
 1      other hand, chronic O3 exposures in a range of 0.5 to 1.0 ppm induce epithelial hyperplasia that
 2      disappears in a few days, and the weight of evidence from new experimental animal studies
 3      (using non-lifetime exposures) does not support ambient O3 as being a pulmonary carcinogen.
 4           Collectively, the evidence from animal studies strongly suggest that O3 is capable of
 5      damaging the distal airways and proximal alveoli, resulting in lung tissue remodeling leading to
 6      apparent irreversible changes. Compromised pulmonary function and  structural changes due to
 7      persistent inflammation may exacerbate the progression and development of chronic lung
 8      disease.
 9
10      8.5.3  Mortality-Related Health Endpoints
11           An extensive analysis of population time-series studies that evaluated the air pollution
12      related mortality risk estimates presented in Section 7-4 utilized data from single and multicity
13      studies from around the world.  Mortality risk estimates derived from studies in U.S. and Canada
14      coupled with meta-analyses (Figures 8-6 and 8-7) all indicate an elevated risk  for mortality on
15      acute O3 exposure after adjustment for the influence of season and PM (Figure 8-7). Meta-
16      analyses of large U.S. multicity studies also suggest a positive association. Mortality risk
17      estimates derived from the studies that analyzed PM as a potential confounder suggest that the
18      reported estimates are not attributable to confounding by PM. Several single-city studies that
19      specifically evaluated the relationship between cardiovascular mortality and O3 exposure also
20      indicated a positive association.
21           The epidemiology results outlined above for mortality suggest a pattern of effects that may
22      be biologically germane to interpretation of its causality, but our knowledge about potential
23      underlying mechanisms remains very limited and suggests a need for further experimental
24      support.  The majority of the physiological and biochemical parameters evaluated both in human
25      clinical and animal toxicology studies (Table 8-1; Figure 8-9) suggest a relatively transient
26      nature for O3-induced biochemical perturbations.  Most effects attenuate over time, depending on
27      the preexisting pathophysiology. One can hypothesize a generic pathway of O3-induced lung
28      damage, potentially involving oxidative lung damage with subsequent inflammation and/or
29      decline in lung function leading to respiratory distress.
30           Recent analysis of third National Health and Nutrition Examination Followup study data
31      indicated that about 20% of the adult population have reduced FEVj values indicative of

        August 2005                               8-70        DRAFT -DO NOT QUOTE OR CITE

-------
 1      impaired lung function.  The majority of these individuals have COPD, asthma or fibrotic lung
 2      disease (Mannino et al., 2003). These cardiopulmonary disease conditions are associated with
 3      persistent low-grade systemic inflammation.  It has also been reported that patients with COPD
 4      are at increased risk for cardiovascular disease. Lung disease with underlying inflammation may
 5      also link to low-grade systemic inflammation associated with atherosclerosis. These effects in
 6      disease are independent of cigarette smoking (Sin et al., 2005). Lung function decrements in
 7      cardiopulmonary disease has also been associated with  inflammatory markers such as C-reactive
 8      protein (CRP) in blood.  In fact, at the population level, individuals with the lowest FEVj have
 9      the highest levels of CRP, while those with highest FEVj have the lowest values for CRP
10      (Mannino et al.,  2003; Sin and Man, 2003).  The complex, physiological and biochemical
11      perturbations that exist simultaneously (Figure 8-9 and  8-10) subsequent to acute exposure to O3
12      may tilt the biological homeostasis mechanisms leading to adverse health effects in people with
13      compromised cardiopulmonary systems. However, no experimental  data are available at this
14      time to support such a hypothesis as being operational in O3-induced cardiovascular mortality
15      observed in the epidemiological studies. It is possible that reevaluation of some of the
16      epidemiological  panel studies that reported changes in CRP in the context of air pollution
17      mortality evaluations focused on PM with correct adjustments for O3 may shed some light on
18      this potential relationship.
19
20
21      8.6  SUSCEPTIBILITY FACTORS
22          Many  factors  such as age, gender, disease, nutritional  status, smoking, and genetic
23      variability may contribute to the differential effects of environmental pollutants, including O3.
24      Genetic factors,  such as single nucleotide polymorphisms (SNPs) and developmental defects,
25      can contribute to innate susceptibility, while acquired susceptibility may develop due to personal
26      habits (smoking, diet, exercise) and other risk factors such as age, gender, pregnancy, and
27      copollutants. However, the available information from animal toxicology and epidemiologic
28      studies did not provide sufficiently clear scientific evidence by which to confidentially identify
29      and/or associate  any specific factor as contributing to adverse health  effects of O3 (U.S.
30      Environmental Protection Agency, 1996a). However, advances in available research results
        August 2005                              8-71        DRAFT -DO NOT QUOTE OR CITE

-------
 1      since then have improved our ability to delineate likely susceptible or vulnerable populations at
 2      increased risk for O3-induced health effects.
 3           New animal toxicology studies using various strains of mice and rats have identified
 4      O3-sensitive and resistant strains and illustrated the importance of genetic background in
 5      determining O3 susceptibility.  Using subacute low exposure regimen (0.3 ppm O3, 48h) studies
 6      on inbred strains that have been designated as inflammation prone or resistant, Kleeberger et al.,
 7      (1997) identified the pro-inflammatory cytokine gene, Tnf-a,, as a susceptibility gene. Further
 8      characterization of this model indicated a role for TNF receptors (TNFR1, TNFR2) in O3-
 9      induced pulmonary epithelial injury and inflammation (Cho et al., 2001). Studies on five inbred
10      strains of mouse with differing response to O3 exposure (acute high dose or low dose continuous
11      exposure for 3 days), reported a protective role for clara cell secretory protein (CCSP) against
12      O3-induced oxidative damage (Broeckaert et al., 2003; Wattiez et al., 2003). The role for these
13      genes and/or their orthologs in human susceptibility to O3 exposure is yet to be examined.
14           Apart from age at the time of exposure, controlled human exposure studies have also
15      indicated a high degree of interindividual variability in some of the pulmonary physiological
16      parameters. Recent studies by David et al. (2003) and Romieu et al. (2004) reported a role for
17      genetic polymorphism in antioxidant enzymes and genes involved in inflammation to modulate
18      pulmonary function and inflammatory responses to O3 exposure.  Similar to mouse studies
19      referred above, polymorphism in Tnf-a has been implicated in O3-induced lung function changes
20      in healthy, mild asthmatics and individuals with rhinitis.  These observations suggest a potential
21      role for these markers in the innate susceptibility to O3, however, the validity of these markers
22      and their relevance in the context of prediction to population studies need additional
23      experimentation.
24           Biochemical and molecular parameters extensively evaluated in these experiments were
25      used to identify specific loci on the chromosomes and, in some cases, to relate the differential
26      expression of specific genes to biochemical and physiological differences observed among these
27      species.  Utilizing O3-sensitive and O3-resistant species, it has been possible to identify the
28      involvement of AHR and inflammation processes in O3 susceptibility. However, most of these
29      studies were carried out using relatively high doses of O3, making the relevance of these studies
30      questionable in human health effects assessment.  No doubt, the molecular parameters identified
31      in these studies may serve as useful biomarkers with the availability of suitable technologies and,

        August 2005                               8-72       DRAFT -DO NOT QUOTE OR CITE

-------
 1      ultimately, can likely be integrated with epidemiological studies. Interindividual differences in
 2      O3 responsiveness have been observed across a spectrum of symptoms and lung function
 3      responses do not yet allow identification of important underlying factors, except a significant
 4      role for age.
 5
 6      8.6.1   Preexisting Disease as a Potential Risk Factor
 7           People with preexisting pulmonary disease may be at increased risk from O3 exposure.
 8      Altered physiological, morphological and biochemical states typical of respiratory diseases like
 9      asthma, COPD and chronic bronchitis may render people sensitive to additional oxidative burden
10      induced by O3 exposure.  Based on studies assessed in the 1996 O3 AQCD  (U.S. Environmental
11      Protection Agency, 1996a), asthmatics appear to be at least as, or more, sensitive to the acute
12      effects of O3 as healthy nonasthmatic subjects.  The new results reviewed in Chapters 6 and 7 of
13      this document from controlled exposure and epidemiologic studies also suggest that asthmatics
14      are a potentially sensitive subpopulation for O3 health effects.
15           A number of time-series  epidemiologic studies have reported increased risk in study
16      subsets of individuals with preexisting lung diseases, among which tend to implicate asthmatics
17      as potentially susceptible individuals.  The epidemiologic studies of acute exposure to O3
18      discussed in Section 8.4.2 indicate increased risk for exacerbation of disease symptoms during
19      the warm season.
20           Newly available human exposure studies by Stenfors and coworkers have shown
21      differences regarding PMN influx in BALF between asthmatics and healthy human subjects.
22      In vitro studies (Stenfors et al., 2002) using nasal mucosal biopsies from atopic and nonatopic
23      subjects exposed to 0.1 ppm O3 found significant differences in the release of IL-4, IL-6, IL-8,
24      and TNF-a. A subsequent study by the same group (Schierhorn et al.,  2002) found a significant
25      difference in the O3-induced release of the neuropeptides neurokinin A and substance P from
26      allergic patients, compared to nonallergic controls, suggesting increased activation of sensory
27      nerves by O3 in the allergic tissues.  Another report from Bayram et al. (2002) using in vitro
28      culture of bronchial epithelial cells recovered from atopic and nonatopic asthmatics indicated the
29      existence of a significant difference in permeability (by measuring the  paracellular flux
30      of 14C-BSA). Additional controlled O3 exposure studies in human subjects with intermittent
31      asthma (Hiltermann et al., 1999), and asthmatics (Basha et al., 1994; Scannell  et al., 1996)

        August 2005                               8-73        DRAFT -DO NOT QUOTE OR CITE

-------
 1      reported increased secretion of IL-8 suggesting increased neutrophilic inflammation in those
 2      subjects.  Two studies (Torres et al. 1996; Holz et al. 2002) observed increased airway
 3      responsiveness to repeated daily O3 exposure to bronchial allergen challenge in subjects with
 4      preexisting allergic airway disease.
 5           Newly available reports from controlled human exposure studies (see Chapter 6) utilized
 6      subjects with preexisting cardiopulmonary diseases such as COPD, asthma, allergic rhinitis, and
 7      hypertension. The data generated from these studies that evaluated pulmonary function changes
 8      in spirometry did not clearly find differences between filtered air and O3 exposure in COPD and
 9      asthmatic subjects.  However, the data on airway responsiveness, inflammation and various
10      molecular markers of inflammation and bronchoconstriction indicate that people with atopic
11      asthma and allergic rhinitis are potentially susceptible groups for O3-induced adverse health
12      effects. There was only one study with a limited number of subjects that evaluated the effects
13      of O3 exposure in hypertensive patients, and it did not find significant O3-induced changes in
14      clinical parameters, such as heart rate, blood pressure and ECG.
15           The observation of increased pathology in long-term animal exposure studies in the
16      absence of observable physiological changes also suggests that chronic exposure may increase
17      susceptibility to adverse health effects, but this needs to be validated via long-term
18      epidemiologic studies.
19
20      8.6.2   Potential Public Health Impacts
21           Exposure to ambient O3 is associated with various health outcomes, including increased
22      incidence of cough, reduction in lung function, increased inflammation, and increased hospital
23      admissions and mortality. In protecting public health, a distinction must be made between health
24      effects that are considered "adverse" and those that are not. What constitutes an adverse health
25      effect varies for different population groups, with some changes  in healthy individuals not being
26      viewed as adverse but those of similar type and magnitude in other susceptible individuals being
27      seen as adverse. Hence, the definition of adversity of health effects will be an important issue in
28      ultimately considering and describing the rationale for decisions concerning review of the O3
29      NAAQS.
30
        August 2005                               8-74       DRAFT -DO NOT QUOTE OR CITE

-------
 1      8.6.2.1  General Concepts Related to Defining of Adverse Health Effects
 2           The official statement of the American Thoracic Society (ATS) published on "What
 3      Constitutes an Adverse Health Effect of Air Pollution?" (ATS, 2000) updated guidance for
 4      defining adverse respiratory health effects published fifteen years earlier (ATS, 1985) to address
 5      new investigative approaches used to identify the effects of air pollution and to reflect the
 6      concern for the impacts of air pollution on specific susceptible groups.
 7           In the 2000 update, there is an increased focus on quality of life measures as indicators of
 8      adversity and also a more specific consideration of population risk. Exposure to air pollution
 9      that increases the risk of an adverse effect to the entire population is adverse, even though it may
10      not increase the risk of any identifiable individual to an unacceptable level. For example, a
11      population of asthmatics could have a distribution of lung function such that no identifiable
12      single individual has a level associated with significant impairment.  Exposure to air pollution
13      could shift the distribution to lower levels that still do not bring any identifiable individual to a
14      level that is associated with clinically relevant effects. However, this would be considered to be
15      adverse because individuals within the population would have diminished reserve function and,
16      therefore, would be at increased risk if affected  by another agent.
17           Reflecting new investigative approaches, the ATS statement also describes the potential
18      usefulness of research into the genetic basis for disease, including responses to environmental
19      agents, that will provide insights into the mechanistic basis for susceptibility, and provide
20      markers of risk status.  Likewise biomarkers, that are indicators of exposure, effect or
21      susceptibility, may someday be useful in defining the point at which a response should be
22      equated with an adverse effect.
23           The 1996 O3 AQCD provided information useful in helping to define adverse health effects
24      associated with  ambient O3 exposure by describing gradation of severity and adversity of
25      respiratory-related effects, and those definitions are reproduced and presented here as Tables 8-2
26      and 8-3. The severity of effects described in those tables and the approaches taken to define the
27      adversity still appear to be valid and reasonable even in the context of the new ATS statement
28      (ATS, 2000).
29
        August 2005                               8-75        DRAFT -DO NOT QUOTE OR CITE

-------
            Table 8-2. Gradation of Individual Responses to Short-Term Ozone Exposure in
                                           Healthy Personsa*
Functional Response
FEVj
Nonspecific
bronchial responsiveness'5
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 See text for discussion; see Appendix A for abbreviations and acronyms.
        b An increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PD20 or PD100.
        *This table is reproduced from the 1996 O3 AQCD (Table 9-1, page 9-24) (U.S. Environmental Protection
        Agency, 1996a).
1      8.6.2.2  Estimation of Potential Numbers of Persons in At-Risk Susceptible Population
2              Groups in the United States
3           Although O3-related increases in individual health risks may appear to be small, they are
4      likely significant from an overall public health perspective due to the large number of
5      individuals in potential risk groups. Numerous subpopulations may be identified as having
6      increased susceptibility or vulnerability to adverse health effects from O3, including older adults,
7      children, individuals with preexisting cardiopulmonary disease, those of lower socioeconomic
8      status, and those with higher exposure levels. Clearly, the impact of O3 on public health can be
9      very extensive.
       August 2005
8-76
DRAFT -DO NOT QUOTE OR CITE

-------
     Table 8-3.  Gradation of Individual Responses to Short-Term Ozone Exposure in
                       Persons with Impaired Respiratory Systemsa*
Functional
FEVj change
Nonspecific
bronchial
None
Decrements of
<3%
Within normal
range
Small
Decrements of
3 to < 10%
Increases of <100%
Moderate
Decrements of >10
but <20%
Increases of <300%
Large
Decrements of
>20%
Increases of >300%
 responsiveness

 Airway resistance
 (SRaw)
Within normal
range (±20%)
SRaw increased <100%
SRaw increased up to
200% or up to 15cm
H,O/s
SRaw increased
>200% or more than
15cmH,O/s
Duration of
response
Symptomatic
Wheeze
Cough
Chest pain
Duration of
response
Impact of
Responses
Interference with
normal activity
Medical treatment
None
Normal
None
Infrequent
cough
None
None
Normal
None
No change
<4 hours
Mild
With otherwise normal
breathing
Cough with deep breath
Discomfort just
noticeable on exercise
or deep breath
< 4 hours
Mild
Few individuals choose
to limit activity
Normal medication as
needed
>4 hours but
<24 hours
Moderate
With shortness of
breath
Frequent spontaneous
cough
Marked discomfort on
exercise or deep
breath
>4 hours, but
<24 hours
Moderate
Many individuals
choose to limit
activity
Increased frequency of
medication use or
additional medication
>24 hours
Severe
Persistent with
shortness of breath
Persistent
uncontrollable
cough
Severe discomfort
on exercise or deep
breath
>24 hours
Severe
Most individuals
choose to limit
activity
Physician or
emergency room
visit
 a See text for discussion; see Appendix A for abbreviations and acronyms.
 b An increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PD20
 *This table is reproduced from the 1996 O3 AQCD (Table 9-2, page 9-25) (U.S. Environmental Protection
  Agency, 1996a).
August 2005
                         8-77
                       DRAFT -DO NOT QUOTE OR CITE

-------
 1           One consideration in the assessment of potential public health impacts is the size of various
 2      population groups that may be at increased risk for health effects associated with O3-related air
 3      pollution exposure. Table 8-4 summarizes information on the prevalence of chronic respiratory
 4      and circulatory conditions in the U.S. population in 2002 and 2003 (Dey and Bloom, 2005;
 5      Lethbridge-Cejku et al., 2004).  Individuals with preexisting cardiopulmonary disease constitute
 6      a fairly large proportion of the population, with tens of millions of people included in each
 7      disease category. For respiratory conditions, approximately 11% of U.S. adults and 13% of
 8      children have been diagnosed with asthma, and 6% of adults have COPD (chronic bronchitis and
 9      emphysema). Table 8-5 provides further information on the number of various specific
10      respiratory conditions per 100 persons by age among the U.S. population during the mid-1990s.
11      Approximately 23 million people, or 11% of the U.S. adult population, have some type of heart
12      disease, with 6% reporting diagnoses of coronary heart disease. Approximately 21% of the U.S.
13      adult population has hypertension.  Cardiovascular conditions are more common among  older
14      age groups, while asthma prevalence is higher in children.
15           In addition, subpopulations based on age group or socioeconomic status also comprise
16      substantial segments of the population that may be potentially at risk for O3-related health
17      impacts. Based on U.S. census  data from 2003, about 26% of the U.S. population are under
18      18 years of age and 12% are 65  years of age or older.  Approximately 12% of the U.S.
19      population (including 18% of children) are below the poverty level and 16% do not have health
20      insurance coverage. Hence, large proportions of the U.S. population are included in groups that
21      are considered likely to have increased susceptibility and vulnerability for health effects  from
22      ambient O3 exposure.
23           The health statistics data illustrate what is known as the "pyramid" of effects. At the top of
24      the pyramid, there are approximately 2.5 millions deaths from all causes per year in the U.S.
25      population, with about 900,000  deaths due to circulatory diseases and 100,000 deaths from
26      chronic lower respiratory diseases (Kochanek et al., 2004). For circulatory disease morbidity,
27      there are approximately 6 million hospital discharges per year (DeFranees et al., 2005),
28      4.5 million emergency department visits (McCaig and Burt, 2005), and 80 million ambulatory
29      care visits (Woodwell and Cherry, 2004). For respiratory health diseases, there are nearly
30      4 million hospital discharges per year (DeFrances et al., 2005), 14 million emergency
31      department visits (McCaig and Burt, 2005), 112 million ambulatory care visits (Woodwell and

        August 2005                              8-78        DRAFT -DO NOT QUOTE OR CITE

-------
Table 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)
to
o
o






^
VO

o
H
6
O
0
H
O
O
H
W
O
O
H
W

Chronic Condition/Disease
Respiratory Conditions
Asthma
COPD
Chronic Bronchitis
Emphysema
Circulatory Conditions
All Heart Disease
Coronary Heart Disease
Hypertension
Stroke

Chronic Condition/Disease
Respiratory Conditions
Asthma
Source: Lethbridge-Cejku et al.


Age (Years) Region
Adults (18+ Years) 18-44 45-64 65-74 75+ Northeast Midwest South
Cases
(XlO)% % % % % % °/o °/o

21.9 10.6 11.5 10.6 8.4 7.6 11 10.9 9.8

9.1 4.4 3.5 5.5 5.5 5.3 3.8 4 5.4
3.1 1.5 0.3 2 4.9 4.7 1.5 1.8 1.7

22.7 11.2 4 12.7 26.3 36.6 10.5 11.7 11.6
12.5 6.2 0.9 7.1 18.7 24.5 5.7 6.2 6.8
43.3 21.2 7.4 29 49.6 51.8 19.7 21.1 23.3
4.8 2.4 0.4 2.5 6.4 11.1 2.4 2.3 2.4
Age (Years) Region
Children
(<18 years) 0-4 5-11 12-17 Northeast Midwest South
Cases
(y 1 (\6\ o/ o/ o/ o/ o/ o/ o/
* 1U ) /o /o /o /o /o /o /o

9.1 12.5 7.5 14 14.7 14 13.5 11.8
(2004) for data on adults (18+ years); Dey and Bloom (2005) for data on children (<18 years).



West

%

11.8

3.8
1.1

10.7
5.8
18.9
2.7


West
o/o

11.2




-------
           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
l.T
2.4*
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
        Tigure does not meet standard of reliability or precision.
        Source: Adams et al. (1999).
 1     Cherry, 2004), and an estimated 700 million restricted activity days per year due to respiratory
 2     conditions (Adams et al., 1999). Combining small risk estimates with relatively large baseline
 3     levels of health outcomes can result in quite large public health impacts.  Thus, even a small
 4     percentage reduction in O3 health impacts on cardiopulmonary diseases would reflect a large
 5     number of avoided cases.
 6          Another key input for public health impact assessment is the range of concentration-
 7     response functions for various health outcomes.  Epidemiologic studies have reported
 8     associations between short-term exposure to O3 with mortality, hospitalizations for
 9     cardiopulmonary diseases,  reduced lung function, incidence of respiratory symptoms, and
10     changes in heartbeat rhythm and rate.  Effect estimates for morbidity responses to short-term
11     changes in O3 tend to be larger in magnitude than those for mortality.
12          A limited number of studies assessed the impact of reductions in air pollution levels on
13     health outcomes. A study by Neidell (2004) examined the relationship between air pollutants
14     and asthma hospitalizations in California. The most recent EPA O3 report (U.S. Environmental
15     Protection Agency, 2004b) indicated that the fourth highest daily 8-h max O3 levels in the pacific
       August 2005
8-80
DRAFT -DO NOT QUOTE OR CITE

-------
 1      southwest region had decreased by 16% from 1990 to 2003. This downward trend in O3 levels
 2      was mostly influenced by the improvements in Los Angeles and other southern California
 3      metropolitan areas.  Results from this study noted declines in levels of air pollutants since 1992
 4      and decreased asthma admissions in  1998 for children aged 1 to 18 years ranging from 5 to 14%,
 5      depending on the age group. The greatest decline (>10%) in air pollution-related asthma
 6      admissions was observed among 3- to 12-year old children.  Although this benefit analysis was
 7      not specific to O3, it provides evidence of decreased morbidity resulting from reduced air
 8      pollutant concentrations, including O3.
 9           An intervention study in Atlanta, GA, during the 1996 Summer Olympic Games examined
10      the impact of a city wide decrease in automobile traffic on air quality and childhood asthma
11      (Friedman et al., 2001). Citywide acute care visits and hospitalizations for asthma during the
12      17 days of the Olympic Games were compared to a baseline period consisting of the four weeks
13      before and after the  Olympic Games.  During the Olympic Games, levels for all pollutants
14      generally declined, but the most dramatic change was observed for O3. The 1-h max O3
15      concentration in Atlanta decreased 27.9% from a mean of 81.3 ppb during the baseline period to
16      58.6 ppb during the  Olympic Games.  The number of asthma acute care events also decreased by
17      41.6% in the Georgia Medicaid claims file,  compared to a 3.1% decline in nonasthma acute care
18      events. Combining  data from the baseline and intervention periods, a 31% (95% CI: 0.8, 69.9)
19      excess risk of asthma events was observed per 40 ppb increase in  1-h max O3 at a cumulative lag
20      of 0- to 2-days. Although a 16.1% decrease in PM10 concentrations also occurred during the
21      Olympic Games, there was no association between PM10 and asthma acute care events.
22           Many studies have examined O3-related health effects, yet only a few have addressed the
23      question as to the extent to which reductions in ambient O3 actually lead to reductions in adverse
24      health outcomes attributable to O3. While the findings from these studies suggest that decreases
25      in ambient O3 levels will likely lead to a reduction in asthma-related hospital admissions and
26      emergency visits, more studies are needed to diminish uncertainty regarding this issue of
27      accountability.
28           In addition to attribution of risks for the various health outcomes to O3 and other
29      copollutants, important considerations in assessing the impact of O3 on public health include the
30      size of the population at risk as well as the concentration-response relationship  and the potential
        August 2005                              8-81       DRAFT -DO NOT QUOTE OR CITE

-------
 1      identification of threshold levels. Taken together, it can be concluded that exposure to ambient
 2      O3 likely has a significant impact on public health in the United States.
 3
 4
 5      8.7  SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS
 6           This section summarizes the main conclusions derived from this integrated synthesis of
 7      information regarding health effects associated with ambient O3 exposures. The conclusions
 8      derived are based on an integrated analysis of animal, human clinical toxicological and
 9      epidemiological studies that have evaluated health effects associated with short-term, repeated,
10      and long-term exposures to O3 alone or in combination with other ambient pollutants.
11      Experimental evidence from human and animal toxicological studies presented in Chapters 4,
12      5, and 6 was utilized to provide biological plausibility for the health effects observed in
13      epidemiologic studies.  These empirical efforts are also aimed at identifying susceptible
14      populations that are at potentially greater risk for effects of O3 exposure.
15
16      1. Health effects of acute (short-term) exposures to Ozone
17           Numerous field panel and time-series epidemiologic studies (using better weather models
18      and adjustments to confounding copollutants than compared to those assessed in the 1996 O3
19      AQCD ) have evaluated the effects  of short-term exposure to O3  on a wide range of health
20      endpoints, from lung function decrements to mortality.  Results from the majority of studies
21      continue to support the conclusions reported in the 1996 O3 AQCD.
22
23      »D Panel studies typically have evaluated the effects of short-term O3 exposure on both healthy
24         individuals  and people with cardiopulmonary diseases.  These evaluations included
25         measurement of lung function changes, respiratory symptoms and  use of asthma medication.
26
27      »D Clinical controlled exposure studies in humans indicate changes in lung function and
28         respiratory symptoms that vary  as a function of exposure concentration, duration and level
29         of exercise.
30
        August 2005                              8-82       DRAFT -DO NOT QUOTE OR CITE

-------
 1      •  Newer meta-analyses confirmed the interindividual differences in lung function decrements
 2         reported in the 1996 O3 AQCD.  Age-specific differences in the lung function responses were
 3         also observed. Spirometric responses (due to decrements in lung function) in healthy adults
 4         exposed to near ambient O3 levels typically resolve to near baseline values within 4-6 h.
 5
 6      •  Meta-analyses of four controlled human exposure studies (two new and two reported in the
 7         1996 O3 AQCD) reporting the effects of prolonged  (6.6 h) exposures to 0.08  ppm O3 during
 8         moderate exercise on pulmonary function in young  healthy adults (M = 90, F = 30; mean
 9         age, 23 yrs) indicate an absolute FEVj decrease of 6%, whereas FEVj increased by 1%
10         following free air (FA)  exposures.
11
12      •  Inflammatory responses (PMN, inflammation mediators such as cytokines and chemokines)
13         and permeability changes (proteins, albumin), typically measured in BALF, also exhibit
14         intersubject variability.  Recent meta-analyses on numerous clinical studies indicate
15         interindividual differences in response to short-term O3 exposures.
16
17      •  Inflammatory and permeability responses also resolve (in some instances complete recovery)
18         and exhibit differential  attenuation profiles between normal healthy subjects  and people with
19         preexisting respiratory diseases. Some lung inflammation markers may not resolve readily
20         and mild persistent inflammation has been reported.
21
22      •  Field/panel studies of healthy individuals and asthmatics have revealed a positive association
23         between short-term exposure to O3 and decrements  in lung function.
24
25      •  An association between on short-term O3 exposures and school absenteeism (due to
26         respiratory illness) has also been suggested.
27
28      •  With regard to cardiac impacts, a limited number of field studies that examined the
29         relationship between short-term O3 exposures and cardiovascular effects (heart rate
30         variability, myocardial infarction) suggest an association.
31

        August 2005                              8-83        DRAFT -DO  NOT QUOTE OR CITE

-------
 1      •  A large multicity and several single-city studies have indicated a positive association
 2         between increased O3 levels (especially during the warm season) and increased risk for
 3         hospital admissions. On the other hand epidemiologic data on emergency department visits
 4         do not suggest such an association with increase in ambient O3 levels.
 5
 6      •  Data from two large multicity studies from the U.S. and several single-city studies suggest a
 7         positive association between increase in O3 levels and all cause (non-accidental) daily
 8         mortality. Meta-analyses on the influence of season suggest a causal association. Additional
 9         meta-analyses on cause-specific mortality are suggestive of a likely positive association
10         between increases in ambient O3 levels and cardiovascular mortality.
11
12      •  Short-term O3-induced lung function decrements, respiratory symptoms, inflammation and
13         permeability changes observed in animal toxicology studies  are consistent with human
14         studies.
15
16      2. Health effects of repeated short-term exposures to Ozone
17           The results of new controlled human exposure studies of repeated short-term O3 exposures
18      continue to support the health effects findings/conclusions reported in the 1996 O3 AQCD.
19
20      •  Repeated exposure studies at higher concentrations typically show that FEVj response to O3
21         is enhanced on the second of several days of exposure. Such an enhanced response was not
22         observed at lower O3 concentrations. With repeated O3 exposures over several days,
23         spirometric and symptom responses become  attenuated, but this tolerance is lost after about a
24         week without exposure.
25
26      •  In humans repeatedly exposed to 0.4 ppm O3 for 5 consecutive days, several indicators of
27         inflammation (e.g., PMN influx, IL-6, PGE2, BAL protein, fibronectin) were attenuated after
28         5 days of exposure.  Lung injury and permeability markers (LDH, IL-8, total protein,
29         epithelial cells) did not show attenuation, indicating that tissue damage probably continues  to
30         occur during repeated exposure. The recovery of the inflammatory response occurred for
31         some markers after 10 days, but some responses were not normalized even after 20 days.

        August 2005                               8-84       DRAFT -DO NOT QUOTE OR CITE

-------
 1         The continued presence of cellular injury markers indicates a persistent effect that may not
 2         necessarily be recognized due to the attenuation of spirometric and symptom responses.
 3
 4      •  Repeated daily exposure to lower concentrations of O3 (0.125 ppm for 4 days) causes an
 5         increased response to bronchial allergen challenge in subjects with preexisting allergic
 6         airway disease, with or without asthma.  In these subjects, changes in airway responsiveness
 7         after O3 exposure appear to be resolved more slowly than changes in FEVj or respiratory
 8         symptoms.
 9
10      3. Health effects of long-term exposures to Ozone
11           Assessment of human health effects associated with long-term O3 exposures is hampered
12      by the lack of pertinent data from human clinical and epidemiologic studies. Chronic animal
13      toxicology studies continue to support structural alterations in several regions of the respiratory
14      tract and identify the CAR as the most affected region.
15
16      •  Animal toxicology studies that utilized exposure regimens to simulate seasonal exposure
17         pattern also report increased lung injury compared to conventional chronic stable exposures.
18         One long term study of infant rhesus monkeys exposed to simulated seasonal O3 patterns
19         (0.5 ppm 8h/day for 5 days, every 14 days for 11 episodes) demonstrated: (1) remodeling in
20         the distal airways; (2) abnormalities in tracheal basement membrane;  (3) eosinophil
21         accumulation in conducting airways; (4) decrements in airway innervation.  These findings
22         advance earlier information regarding possible injury-repair processes occurring with
23         seasonal O3 exposures.
24
25      •  Effects of O3 on the upper respiratory tract of F344 rats exposed to O3 (0.12, 0.5, or 1.0 ppm
26         for 20 months) included marked mucous cell metaplasia in the rats exposed to 0.5 and
27         1.0 pm O3, but not at 0.12 ppm O3.  The persistent nature of the O3-induced mucous cell
28         metaplasia suggests that O3 exposure may have the potential to induce similar long-lasting
29         alterations in the airways of humans. Hyperplasia in the nasal  epithelium of rats exposed to
30         0.25 and 0.5 ppm, 8h/day, 7 days/week, for 13 weeks has been reported.
31

        August 2005                              8-85       DRAFT -DO NOT QUOTE OR CITE

-------
 1      •  Pathophysiological changes associated with chronic O3 exposures observed in animal studies
 2         suggest possible similar alterations in humans. The pulmonary function changes observed in
 3         children in polluted metropolitan areas and lung structural alterations reported in autopsy
 4         study in Los Angeles suggest a role for long-term ambient O3 exposure and need to be
 5         critically evaluated with proper study design.
 6
 7      4.  Susceptibility factors associated with exposure to ozone
 8           Various factors such as age, gender, nutrition, socioeconomic, activity patterns, and disease
 9      status have been shown to influence the response to environmental air pollutants. Controlled
10      human exposure studies clearly established differential biological  response to O3 based on
11      physical activity (exertion) and age.  These studies also demonstrated a large variation in
12      sensitivity and responsiveness to O3. The specific factors that contribute to this intersubject
13      variability are yet to be identified.
14
15      •  Increased hospital admissions for asthma and COPD in summer (with increased levels of
16         ambient O3) suggest that people with these respiratory diseases as potential sub-population
17         for O3-induced health effects.
18
19      •  Similarly, based on O3-induced differential responses in lung inflammation and in airway
20         hyperresponsiveness, asthmatics (including children) appear to have potentially increased
21         susceptibility to O3. However, there is no supportive data from controlled human studies
22         suggesting individuals with COPD are more sensitive to O3-induced health effects.
23
24      •  Animal toxicology studies provided supportive evidence to the observations of varied
25         susceptibility. Various strains of mice and rats have demonstrated the importance of genetic
26         background in O3 susceptibility.  Moreover, genetic and molecular characterization studies in
27         laboratory animals identified genetic loci responsible for both  sensitivity and resistance.
28
29      •  Consistent with the 1996 O3 AQCD, the scarcity of data prevents determination of the role of
30         ethnic or racial background and nutrition status on O3-induced health effects. However, as
31         presented in this document, exercising (moderate to high physical exertion) healthy,

        August 2005                               8-86        DRAFT -DO NOT QUOTE OR CITE

-------
 1         adolescents, and asthmatics appear to demonstrate increased responsiveness to ambient
 2         concentrations of O3 and may be susceptible for O3-induced health effects.
 3
 4      5. Health effects of binary pollutant mixtures containing ozone
 5           A limited number of controlled human exposure studies and few animal toxicology studies
 6      with the binary mixtures containing O3 suggest potential interactions depending on the exposure
 7      regimens  and copollutant constituents.
 8
 9      •  Continuous exposure to SO2 and NO2 increased inhaled bolus O3 absorption, while
10         continuous exposure to O3 decreased O3 bolus absorption.  Asthmatics exhibited enhanced
11         airway reactivity to house dust mite following exposures to O3, NO2, and the combination of
12         the two gases.  Spirometric response, however, was impaired only by O3 and O3+NO2 at
13         higher concentrations.
14
15      •  Animal toxicology studies with O3 in mixture with NO2, formaldehyde, and PM
16         demonstrated additive, synergistic or antagonistic effects depending on the exposure regimen
17         and the endpoints evaluated.
18
19      •  One controlled exposure study of children, designed to approximate exposure conditions of
20         an epidemiologic study by matching the population and exposure atmosphere (0.1 ppm O3,
21         0.1 ppm SO2 and 101 |ig/m2 H2SO4), failed to support the findings of the epidemiologic study.
22         This study points out the difficulties in attempting to link the outcomes of epidemiologic and
23         controlled studies with binary pollutant mixtures.
24
        August 2005                              8-87       DRAFT -DO NOT QUOTE OR CITE

-------
  1       REFERENCES

  2       Adams, P. F.; Hendershot, G. E.; Marano, M. A. (1999) Current estimates from the National Health Interview
  3             Survey, 1996. Hyattsville, MD: U.S. Department of Health and Human Services, Public Health Service,
  4             National Center for Health Statistics; DHHS publication no. (PHS) 99-1528. (Vital and health statistics: v. 10,
  5             data from the National Health Survey, no. 200). Available:
  6             http://www.cdc.gov/nchs/products/pubs/pubd/series/srlO/pre-200/pre-200.htm [12 March, 2001].
  7       Alexis, N.; Urch, B.; Tarlo, S.; Corey, P.; Pengelly, D.; O'Byrne, P.; Silverman, F. (2000) Cyclooxygenase
  8             metabolites play a different role in ozone-induced pulmonary function decline in asthmatics compared to
  9             normals. Inhalation Toxicol. 12: 1205-1224.
10       American Thoracic Society. (1985) Guidelines as to what constitutes an adverse respiratory health effect, with
11             special reference to epidemiologic studies of air pollution. Am. Rev. Respir. Dis. 131: 666-668.
12       American Thoracic Society. (1991) Lung function testing: selection of reference values and interpretative strategies
13             Am. Rev. Respir. Dis. 144: 1202-1218.
14       American Thoracic Society. (2000) What constitutes an adverse health effect of air pollution? Am. J. Respir. Crit.
15             Care Med. 161:665-673.
16       Anderson, H. R.; Spix, C.; Medina, S.; Schouten, J. P.; Castellsague, J.; Rossi, G.; Zmirou, D.; Touloumi, G.;
17             Wojtyniak, B.; Ponka, A.; Bacharova, L.; Schwartz, J.; Katsouyanni, K. (1997) Air pollution and daily
18             admissions for chronic obstructive pulmonary disease in 6 European cities: results from the APHEA project.
19             Eur. Respir. J. 10: 1064-1071.
20       Avol, E. L.; Navidi,  W. C.; Rappaport, E. B.; Peters, J. M. (1998) Acute effects of ambient ozone on asthmatic,
21             wheezy, and healthy children. Cambridge, MA: Health Effects Institute; research report no. 82.
22       Basha, M. A.; Gross, K. B.; Gwizdala, C. J.;  Haidar, A. H.; Popovich, J., Jr. (1994) Bronchoalveolar lavage
23             neutrophilia in asthmatic and healthy volunteers after controlled exposure to ozone and filtered purified air.
24             Chest 106:  1757-1765.
25       Bayram, H.; Rusznak, C.; Khair, O. A.; Sapsford, R. J.; Abdelaziz, M. M. (2002) Effect of ozone and nitrogen
26             dioxide on the permeability of bronchial epithelial cell cultures of non-asthmatic and asthmatic subjects.
27             Clin. Exp. Allergy 32: 1285-1292.
28       Bell, M. L.; McDermott, A.; Zeger, S. L.; Samet, J. M.; Dominici, F. (2004) Ozone and short-term mortality in
29             95 US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292: 2372-2378.
30       Bell, M. L.; Dominici, F.; Samet, J. M. (2005) A meta-analysis of time-series studies of ozone and mortality with
31             comparison to the national morbidity,  mortality, and air pollution study. Epidemiology 16: 436-445.
32       Bosson, J.; Stenfors, N.; Bucht, A.; Helleday, R.; Pourazar, J.;  Holgate, S. T.; Kelly, F. J.; Sandstrom, T.; Wilson, S.;
33             Frew, A. J.; Blomberg, A. (2003) Ozone-induced bronchial epithelial cytokine expression differs between
34             healthy  and asthmatic subjects. Clin. Exp. Allergy 33: 777-782.
35       Broeckaert, F.; Clippe, A.; Wattiez, R.; Falmagne, P.; Bernard, A. (2003) Lung hyperpermeability, Clara-cell
36             secretory potein (CC16), and susceptibility to ozone of five inbred strains of mice. Inhalation Toxicol.
37             15: 1209-1230.
38       Brook, R. D.; Brook, J.  R.; Urch, B.; Vincent, R.; Rajagopalan, S.; Silverman, F. (2002) Inhalation of fine paniculate
39             air pollution and ozone causes acute arterial vasoconstriction in healthy adults. Circulation 105:  1534-1536.
40       Brook, R. D.; Franklin, B.; Cascio, W.; Hong, Y.; Howard, G.; Lipsett, M.; Luepker, R.; Mittleman, M.; Samet, J.;
41             Smith, S. C., Jr.; Tager, I. (2004) Air pollution and cardiovascular disease.  A statement for healthcare
42             professionals from the Expert Panel on Population and Prevention Science  of the American Heart
43             Association. Circulation 109: 2655-2671.
44       Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M.;
45             Dann, T.; Brook,  J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of respiratory
46             admissions to Ontario hospitals. Environ. Res. 65: 172-194.
47       Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997) Association between ozone and
48             hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72: 24-31.
49       Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997) The role of paniculate size and chemistry in the
50             association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
51             Environ. Health Perspect. 105: 614-620.
52       Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M.  E.; Brook, J. R.; Dales, R. E.; Leech, J. A.; Cakmak, S.;
53             Krewski, D. (2001) Association between ozone and hospitalization for acute respiratory diseases in children
54             less than 2 years of age. Am. J. Epidemiol. 153: 444-452.
         August 2005                                    8-88         DRAFT -DO NOT QUOTE OR CITE

-------
  1       Bush, M. L.; Zhang, W.; Ben-Jebria, A.; Ultman, J. S. (2001) Longitudinal distribution of ozone and chlorine in the
  2             human respiratory tract: simulation of nasal and oral breathing with the single-path diffusion model. Toxicol.
  3             Appl. Pharmacol. 173: 137-145.
  4       Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. (2000) Elementary school absenteeism and air pollution.
  5             Inhalation Toxicol. 12: 997-1016.
  6       Cho, H.-Y.; Zhang, L.-Y.; Kleeberger, S. R. (2001) Ozone-induced lung inflammation and hyperreactivity are
  7             mediated via tumor necrosis factor-a receptors. Am. J. Physiol. 280: L537-L546.
  8       Chock, D. P.; Winkler, S. L.; Chen, C. (2000) A study of the association between daily mortality and ambient air
  9             pollutant concentrations in Pittsburgh, Pennsylvania.  J. Air Waste Manage. Assoc. 50: 1481-1500.
10       Cody, R. P.; Weisel, C.  P.; Birnbaum, G.; Lioy, P. J. (1992) The effect of ozone associated with summertime
11             photochemical smog on the frequency of asthma visits to hospital emergency departments. Environ. Res.
12             58: 184-194.
13       Cohen, M. D.; Sisco, M.; Li, Y.; Zelikoff, J. T.; Schlesinger, R. B. (2001) Ozone-induced modulation of
14             cell-mediated immune responses in the lungs. Toxicol. Appl. Pharmacol. 171: 71-84.
15       Coleridge, J. C.  G.; Coleridge, H. M.; Schelegle, E. S.; Green, J. F. (1993) Acute inhalation of ozone stimulates
16             bronchial C-fibers and rapidly adapting receptors in dogs. J. Appl. Physiol. 74: 2345-2352.
17       David, G. L.; Romieu, I.; Sienra-Monge, J. J.; Collins, W. J.; Ramirez-Aguilar, M.; Del Rio-Navarro, B. E.;
18             Reyes-Ruiz, N. I.; Morris, R. W.; Marzec, J. M.; London, S. J. (2003) Nicotinamide adenine dinucleotide
19             (Phosphate) reduced:quinone oxidoreductase and glutathione s-transferase ml polymorphism and childhood
20             asthma. Am. J. Respir. Crit.  Care Med. 168: 1199-1204.
21       DeFrances, C. J.; Hall, M. J.; Podgornik, M. N. (2005) 2003 National Hospital Discharge Survey. Hyattsville, MD:
22             National  Center for Health Statistics; DHHS publication no. (PHS) 2004-1250. (Advance data from vital and
23             health statistics; no. 359). Available: http://www.cdc.gov/nchs/data/ad/ad359.pdf [3 August, 2005].
24       Delaunois, A.; Segura, P.; Montano, L. M.; Vargas, M. H.; Ansay, M.; Gustin, P.  (1998) Comparison of
25             ozone-induced effects on lung mechanics and hemodynamics in the rabbit. Toxicol. Appl. Pharmacol.
26             150:58-67.
27       De Leon, S. F.; Thurston, G. D.; Ito, K. (2003) Contribution of respiratory disease to nonrespiratory mortality
28             associations with air pollution. Am. J. Respir. Crit. Care Med. 167: 1117-1123.
29       Delfino, R. J.; Becklake, M.  R.; Hanley, J. A. (1994) The relationship of urgent hospital admissions for respiratory
30             illnesses to photochemical air pollution levels in Montreal. Environ. Res. 67: 1-19.
31       Delfino, R. J.; Zeiger, R. S.;  Seltzer, J. M.; Street, D. H.; Matteucci, R. M.; Anderson, P. R.; Koutrakis, P. (1997)
32             The effect of outdoor fungal spore concentrations on  daily asthma severity. Environ.  Health Perspect. 105:
33             622-635.
34       Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Squinazi, F.; Momas, I. (2002a) Short-term effects of low-level air
3 5             pollution on respiratory health of adults suffering from moderate to severe asthma. Environ. Res. A 89: 29-37.
36       Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Le Moullec, Y.;  Momas, I. (2002b) Effects of air pollution on adults
37             with chronic obstructive pulmonary disease. Arch. Environ. Health 57:  554-560.
38       Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S. (1991)
3 9             Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the
40             lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
41       Devlin, R. B.; Folinsbee, L. J.; Biscardi, F.; Hatch, G.; Becker, S.; Madden, M. C.; Robbins, M.; Koren, H. S. (1997)
42             Inflammation and cell damage induced by repeated exposure of humans to ozone. Inhalation Toxicol.
43             9:211-235.
44       Dey, A. N.; Bloom, B. (2005) Summary health statistics for  U.S. children: National Health Interview Survey, 2003.
45             Hyattsville, MD:  U.S. Department of Health & Human Services, National Center for Health Statistics. (Vital
46             and health statistics, series 10, no. 223). Available: http://www.cdc.gov/nchs/data/series/sr_10/srlO_223.pdf
47             [3 August, 2005].
48       Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with particulates
49             and acid aerosols. Environ. Res. 59: 362-373.
50       Dominici, F.; McDermott, A.; Daniels, M.; Zeger, S. L.;  Samet, J. M. (2003) Mortality among residents of 90 cities.
51             In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA: Health
52             Effects Institute;  pp. 9-24. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
53       Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
54             Perspect. 107: 637-641.
         August 2005                                     8-89         DRAFT -DO NOT QUOTE OR CITE

-------
  1       Fairley, D. (2003) Mortality and air pollution for Santa Clara County, California, 1989-1996. In: Revised analyses of
  2             time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute;
  3             pp. 97-106. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October, 2004].
  4       Fiore, A. M.; Jacob, D. I; Mathur, R.; Martin, R. V. (2003) Application of empirical orthogonal functions to
  5             evaluate ozone simulations with regional and global models. J. Geophys. Res. (Atmos.)
  6             108(014): 10.1029/2002JD003151.
  7       Foster, W. M.; Stetkiewicz, P. T. (1996) Regional clearance of solute from the respiratory epithelia: 18-20 h
  8             postexposure to ozone. J. Appl. Physiol. 81: 1143-1149.
  9       Foster, W. M.; Costa, D. L.; Langenback, E. G.  (1987) Ozone exposure alters tracheobronchial mucociliary function
10             in humans. J. Appl. Physiol. 63: 996-1002.
11       Foster, W. M.; Silver, J. A.; Groth, M. L. (1993) Exposure to ozone alters regional function and particle dosimetry in
12             the human lung. J. Appl. Physiol. 75: 1938-1945.
13       Foster, W. M.; Weinmann,  G. G.; Menkes,  E.; Macri, K. (1997) Acute exposure of humans to ozone impairs small
14             airway function. Ann. Occup. Hyg. 41(suppl. 1): 659-666.
15       Friedman, M. S.; Powell, K. E.; Hutwagner, L.;  Graham, L. M.; Teague, W. G. (2001) Impact of changes in
16             transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality and
17             childhood asthma. JAMA J. Am. Med. Assoc. 285: 897-905.
18       Fung, K. Y.; Luginaah, L; Gorey, K. M.;  Webster, G. (2005) Air pollution and daily hospital admissions for
19             cardiovascular diseases in Windsor,  Ontario. Can. J. Public Health 96: 29-33.
20       Gamble, J. L.  (1998) Effects of ambient air pollution on daily mortality: a time series analysis of Dallas, Texas,
21             1990-1994. Presented at: 91st annual meeting and exhibition of the Air & Waste Management Association;
22             June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 98-MP26.03.
23       Gent, J. F.; Triche, E. W.; Holford, T. R.; Belanger, K.; Bracken, M. B.; Beckett, W. S.; Leaderer, B. P. (2003)
24             Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA
25             J. Am.  Med.  Assoc. 290: 1859-1867.
26       Gilliland, F. D.; Berhane, K.; Rappaport,  E. B.; Thomas, D. C.; Avol, E.; Gauderman, W. J.; London, S. J.;
27             Margolis, H. G.; McConnell, R.; Islam, K. T.; Peters, J. M. (2001) The effects of ambient air pollution on
28             school  absenteeism due to respiratory illnesses. Epidemiology  12: 43-54.
29       Goldberg, M.  S.; Burnett, R. T.; Brook, J.; Bailar, J. C., Ill; Valois, M.-F.; Vincent, R. (2001) Associations between
30             daily cause-specific mortality and  concentrations of ground-level ozone in Montreal, Quebec. Am. J.
31             Epidemiol. 154: 817-826.
32       Gong, H., Jr.; Shamoo, D. A.; Anderson,  K. R.; Linn, W. S. (1997a) Responses of older men with and without
33             chronic obstructive pulmonary disease to prolonged ozone exposure. Arch. Environ. Health 52: 18-25.
34       Gong, H., Jr.; McManus, M. S.; Linn, W. S. (1997b) Attenuated response to repeated daily ozone exposures in
35             asthmatic subjects. Arch. Environ. Health 52: 34-41.
36       Gong, H., Jr.; Wong, R.; Sarma, R. J.; Linn, W.  S.; Sullivan, E. D.; Shamoo, D. A.; Anderson, K. R.; Prasad, S. B.
37             (1998)  Cardiovascular effects of ozone exposure in human volunteers. Am. J. Respir. Crit. Care Med.
38             158: 538-546.
39       Gryparis, A.; Forsberg, B.; Katsouyanni,  K.; Analitis, A.; Touloumi, G.; Schwartz, J.; Samoli, E.; Medina,  S.;
40             Anderson, H. R.; Niciu, E. M.; Wichmann, H.-E.; Kriz, B.; Kosnik, M.; Skorkovsky, J.; Vonk, J. M.;
41             Dortbudak, Z. (2004) Acute effects of ozone on mortality from the "air pollution and health: a European
42             approach" project. Am. J. Respir. Crit. Care Med.  170: 1080-1087.
43       Gwynn, R. C.; Burnett, R. T.; Thurston, G.  D. (2000) A time-series analysis of acidic paniculate matter and daily
44             mortality and morbidity in the Buffalo, New York, region. Environ. Health Perspect. 108: 125-133.
45       Hatch, G. E.; Slade, R.; Harris, L. P.; McDonnell, W. F.; Devlin, R. B.; Koren, H. S.; Costa, D. L.; McKee, J. (1994)
46             Ozone  dose and effect in humans and rats: a comparison using oxygen-18 labeling and bronchoalveolar
47             lavage. Am.  J. Respir. Crit. Care Med. 150: 676-683.
48       Hazucha, M. J.; Sant'Ambrogio, G. (1993) Effects of ozone on the activity of slowly (SAR) and rapidly adapting
49             (RAR) receptors in cats. FASEB J. 7: 407A.
50       Hazucha, M. J.; Folinsbee, L. J.; Bromberg, P. A. (2003) Distribution and reproducibility of spirometric response to
51             ozone by gender and age. J. Appl.  Physiol. 95: 1917-1925.
52       Hiltermann, T. J. N.; Stolk,  J.; Van der Zee, S. C.; Brunekreef, B.; De Bruijne, C. R.; Fischer, P. H.; Ameling, C. B.;
53             Sterk, P. J.; Hiemstra, P. S.; Van Bree, L. (1998) Asthma severity and susceptibility to air pollution. Eur.
54             Respir. J. 11:686-693.
         August 2005                                    8-90         DRAFT -DO NOT QUOTE OR CITE

-------
  1       Hiltermann, J. T. N.; Lapperre, T. S.; Van Bree, L.; Steerenberg, P. A.; Brahim, J. I; Sont, J. K.; Sterk, P. I;
  2             Hiemstra, P. S.; Stolk, J. (1999) Ozone-induced inflammation assessed in sputum and bronchial lavage fluid
  3             from asthmatics: a new noninvasive tool in epidemiologic studies on air pollution and asthma. Free Radical
  4             Biol. Med. 27: 1448-1454.
  5       Holz, O.; Mticke, M; Paasch, K.; Bohme, S.; Timm, P.; Richter, K.; Magnussen, H.; Torres, R. A. (2002) Repeated
  6             ozone exposures enhance bronchial allergen responses in subjects with rhinitis or asthma. Clin. Exp. Allergy.
  7             32:681-689.
  8       Holz, O.; Tal-Singer, R.; Kanniess, F.; Simpson, K. J.; Gibson, A.; Vessey, R. S. J.; Janicki, S.; Magnussen, H.;
  9             Torres, R. A.; Richter, K. (2005) Validation of the human ozone challenge model as a tool for assessing
10             anti-inflammatory drugs in early development. J. Clin. Pharmacol. 45: 498-503.
11       Hoppe, P.; Peters, A.; Rabe, G.; Praml, G.; Lindner, J.; Jakobi, G.; Fruhmann, G.; Nowak, D. (2003) Environmental
12             ozone effects in different population subgroups. Int. J. Hyg. Environ. Health 206: 505-516.
13       Horstman, D. H.; Ball, B. A.; Brown, J.;  Gerrity, T.; Folinsbee, L. J. (1995) Comparison of pulmonary responses of
14             asthmatic and nonasthmatic subjects performing light exercise while exposed to a low level of ozone.
15             Toxicol. Ind. Health 11: 369-385.
16       Huang, Y.; Dominici, F.; Bell, M. L. (2005) Bayesian hierarchical distributed lag models for summer ozone
17             exposure and cardio-respiratory mortality. Environmetrics  16: 547-562.
18       Inoue, H.; Aizawa, H.; Matsumoto, K.; Shigyo, M.; Takata, S.; Kara, M.; Kara, N. (1997) Effect of p-agonists on
19             histamine-induced airway microvascular leakage in ozone-exposed guinea pigs. Am. J. Respir. Crit. Care
20             Med. 156: 723-727.
21       Ito, K.; Thurston, G. D. (1996) Daily PMj/mortality associations: an investigation of at-risk subpopulations.
22             J. Exposure Anal. Environ. Epidemiol. 6: 79-95.
23       Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and
24             meta-analysis. Epidemiology 16: 446-457.
25       Jaffe, D. H.; Singer, M. E.; Rimm, A. A. (2003) Air pollution and emergency department visits for asthma among
26             Ohio Medicaid recipients, 1991-1996. Environ. Res.  91: 21-28.
27       Torres, R.; Nowak, D.; Magnussen, H.; Speckin, P.; Koschyk, S. (1996) The effect of ozone exposure on allergen
28             responsiveness in subjects with asthma or rhinitis. Am. J. Respir. Crit. Care Med. 153: 56-64.
29       Torres, R.  A.;  Holz, O.; Zachgo, W.; Timm, P.; Koschyk, S.; Mtiller, B.; Grimminger, F.; Seeger, W.; Kelly, F. J.;
30             Dunster,  C.; Frischer, T.; Lubec, G.; Waschewski, M.; Niendorf, A.; Magnussen, H. (2000) The effect of
31             repeated ozone exposures on inflammatory markers in bronchoalveolar lavage fluid and mucosal biopsies.
32             Am. J. Respir. Crit. Care Med. 161: 1855-1861.
33       Kehrl, H. R.; Vincent, L. M.; Kowalsky, R. J.; Horstman, D. H.; O'Neil, J. J.; McCartney, W. H.; Bromberg, P. A.
34             (1987) Ozone exposure increases  respiratory epithelial permeability in humans. Am. Rev. Respir.  Dis.
35             135: 1124-1128.
36       Kim, S.-Y.; Lee, J.-T.; Hong, Y.-C.; Ahn, K.-J.; Kim, H. (2004) Determining the threshold effect of ozone on daily
37             mortality: an analysis of ozone and mortality in Seoul, Korea, 1995-1999. Environ. Res. 94: 113-119.
3 8       King, T. K. C.; Briscoe, W. A. (1968) The distribution of ventilation, perfusion, lung volume and transfer factor
39             (diffusing capacity) in patients with obstructive lung disease. Clin. Sci. 35: 153-170.
40       Kinney, P. L.; Ozkaynak, H. (1991) Associations of daily mortality and air pollution in Los Angeles County.
41             Environ.  Res. 54: 99-120.
42       Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
43             In:  Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
44             mortality and morbidity; January  1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
45       Kinney, P. L.; Thurston, G. D.; Raizenne, M. (1996) The effects of ambient ozone on lung function in children:
46             a reanalysis of six summer camp studies. Environ. Health Perspect. 104:  170-174.
47       Kleeberger, S. R.; Levitt, R. C.; Zhang, L.-Y.; Longphre, M.; Harkema, J.; Jedlicka, A.; Eleff, S. M.; DiSilvestre, D.;
48             Holroyd, K. J. (1997) Linkage analysis of susceptibility to  ozone-induced lung inflammation in inbred mice.
49             Nat. Genet. 17: 475-478.
50       Klemm, R. J.; Mason, R. M., Jr. (2000) Aerosol Research and Inhalation Epidemiological Study (ARIES): air
51             quality and daily mortality statistical modeling-interim results. J. Air. Waste Manage. Assoc. 50: 1433-1439.
         August 2005                                    8-91         DRAFT -DO NOT QUOTE OR CITE

-------
  1       Kochanek, K. D.; Murphy, S. L.; Anderson, R. N.; Scott, C. (2004) Deaths: final data for 2002. Hyattsville, MD:
  2             U.S. Department of Health & Human Services, National Center for Health Statistics; DHHS publication no.
  3             (PHS) 2005-1120. (National vital statistics reports: v. 53, no. 5). Available:
  4             http://www.cdc.gov/nchs/data/nvsr/nvsr53/nvsr53_05.pdf [3 August, 2005].
  5       Kopp, M. V.; Ulmer, C.; Ihorst, G.; Seydewitz, H. H.; Frischer, T.; Forster, I; Kuehr, J. (1999) Upper airway
  6             inflammation in children exposed to ambient ozone and potential signs of adaptation. Eur. Respir. J.
  7             14:854-861.
  8       Koren, H. S.; Devlin, R. B.; Graham, D. E.; Mann, R.; McGee, M. P.; Horstman, D. H.; Kozumbo, W. J.; Becker, S.;
  9             House, D. E.; McDonnell, W. F.; Bromberg, P. A. (1989) Ozone-induced inflammation in the lower airways
10             of human subjects.  Am. Rev. Respir. Dis. 139:  407-415.
11       Korrick, S.  A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K. D.; Rosner, B. A.;
12             Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
13             Environ. Health Perspect. 106: 93-99.
14       Krishna, M. T.; Madden, J.; Teran, L. M.; Biscione, G. L.; Lau, L. C. K.; Withers, N. J.; Sandstrom, T.; Mudway, L;
15             Kelly, F. J.; Walls,  A.; Frew, A.  J.; Holgate, S.  T. (1998) Effects of 0.2 ppm ozone onbiomarkers of
16             inflammation in bronchoalveolar lavage fluid and bronchial mucosa of healthy subjects. Eur. Respir. J.
17             11: 1294-1300.
18       Kronenberg, R. S.; Drage, C. W.; Ponto, R. A.; Williams, L. E. (1973) The effect of age on the distribution of
19             ventilation and perfusion in the lung. Am. Rev. Respir. Dis. 108:  576-586.
20       Leikauf, G. D.; Simpson, L. G.; Santrock, J.; Zhao, Q.; Abbinante-Nissen, J.; Zhou, S.; Driscoll, K. E. (1995)
21             Airway  epithelial cell responses  to ozone injury. Environ. Health Perspect. 103(suppl. 2): 91-95.
22       Lethbridge-Cejku, M.; Schiller, J. S.; Bernadel, L. (2004) Summary health statistics for U.S. adults: National Health
23             Interview Survey, 2002. Hyattsville, MD: Centers for Disease Control and Prevention; DHHS publication no.
24             (PHS) 2004-1550. (Vital and health statistics, series 10, number 222). Available:
25             http://www.cdc.gov/nchs/data/series/sr_10/srlO_222.pdf [3, August, 2005].
26       Levy, J. L; Carrothers, T. J.; Tuomisto,  J. T.; Hammitt, J. K.; Evans, J. S. (2001) Assessing the public health benefits
27             of reduced ozone concentrations. Environ. Health Perspect.  109: 1215-1226.
28       Levy, J. I.; Chemerynski, S. M.; Sarnat, J. A. (2005) Ozone exposure and mortality, an empiric Bayes
29             metaregression analysis. Epidemiology 16: 458-468.
30       Liao, D.; Duan, Y.; Whitsel, E. A.; Zheng, Z.-J.; Heiss, G.; Chinchilli, V. M.; Lin, H.-M. (2004) Association of
31             higher levels of ambient criteria pollutants with impaired cardiac autonomic  control: a population-based
32             study. Am. J. Epidemiol. 159: 768-777.
33       Linn, W. S.; Szlachcic, Y.; Gong, H., Jr.; Kinney, P. L.; Berhane, K. T. (2000) Air pollution and daily hospital
34             admissions in metropolitan Los Angeles. Environ. Health Perspect. 108: 427-434.
35       Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000) Daily mortality  in the Philadelphia metropolitan area and
36             size-classified paniculate matter. J. Air Waste Manage. Assoc.  50: 1501-1513.
37       Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R.  T. (2000) Association of paniculate matter components with daily
3 8             mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report
39             no. 95.
40       Mannino, D. M.; Ford, E.  S.; Redd, S. C. (2003) Obstructive and restrictive lung disease and markers of
41             inflammation: data from the Third National Health and Nutrition Examination. Am. J. Med. 114: 758-762.
42       McBride, D. E.; Koenig, J. Q.; Luchtel, D. L.; Williams, P. V.; Henderson, W. R., Jr. (1994) Inflammatory effects of
43             ozone in the upper airways of subjects with asthma. Am. J. Respir. Crit. Care Med. 149: 1192-1197.
44       McCaig, L. F.; Burt, C. W. (2005) National Hospital Ambulatory Medical Care Survey: 2003 Emergency
45             Department Summary. Hyattsville, MD: National Center for Health Statistics; DHHS publication no. (PHS)
46             2005-1250. (Advance data from vital and health statistics; no. 358). Available:
47             http://www.cdc.gov/nchs/data/ad/ad358.pdf [3 August, 2005].
48       McDonnell, W. F. (1996)  Individual variability in human lung function responses to ozone exposure. Environ.
49             Toxicol. Pharmacol. 2: 171-175.
50       Moolgavkar, S. H. (2003) Air pollution and daily mortality in two  U.S. counties: season-specific analyses and
51             exposure-response relationships. Inhalation Toxicol. 15: 877-907.
52       Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A.; Anderson, E. L. (1995) Air pollution and daily mortality in
53             Philadelphia. Epidemiology 6: 476-484.
54       Mortimer, K. M.; Neas, L. M.; Dockery, D. W.; Redline, S.; Tager, I. B. (2002) The effect of air pollution on
55             inner-city children with asthma. Eur. Respir. J.  19: 699-705.
56       Mudway, I. S.; Kelly, F. J. (2000) Ozone and the lung: a sensitive  issue. Mol. Aspects. Med. 21:  1-48.
         August 2005                                    8-92        DRAFT -DO NOT QUOTE OR CITE

-------
  1       Mudway, I. S.; Kelly, F. J. (2004) An investigation of inhaled ozone dose and the magnitude of airway inflammation
  2             in healthy adults. Am. J. Respir. Crit. Care Med. 169: 1089-1095.
  3       Neidell, M. J. (2004) Air pollution, health, and socio-economic status: the effect of outdoor air quality on childhood
  4             asthma. J. Health Econ. 23: 1209-1236.
  5       Ostro, B. (1995) Fine paniculate air pollution and mortality in two Southern California counties. Environ. Res.
  6             70:98-104.
  7       Park, H.; Lee, B.; Ha, E.-H.; Lee, J.-T.; Kim, H.; Hong, Y.-C. (2002) Association of air pollution with school
  8             absenteeism due to illness. Arch. Pediatr. Adolesc. Med. 156: 1235-1239.
  9       Park, S. K.; O'Neill, M. S.; Vokonas, P. S.; Sparrow, D.; Schwartz, J. (2005) Effects of air pollution on heart rate
10             variability: the VA normative aging study. Environ. Health Perspect. 113: 304-309.
11       Passannante, A. N.; Hazucha, M. J.; Bromberg, P. A.; Seal, E.; Folinsbee, L.; Koch, G. (1998) Nociceptive
12             mechanisms modulate ozone-induced human lung function decrements. J. Appl. Physiol. 85: 1863-1870.
13       Peel, J. L.; Tolbert, P. E.; Klein, M.; Metzger,  K. B.; Flanders, W. D.; Knox, T.; Mulholland, J. A.; Ryan, P. B.;
14             Frumkin, H. (2005) Ambient air pollution and respiratory emergency department visits. Epidemiology
15             16: 164-174.
16       Peters, A.; Dockery, D. W.; Muller, J. E.; Mittleman, M. A. (2001) Increased paniculate air pollution and the
17             triggering of myocardial infarction. Circulation 103: 2810-2815.
18       Pope, C. A., Ill; Burnett, R. T.; Thun, M.  J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
19             cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA J.  Am. Med.
20             Assoc. 287: 1132-1141.
21       Pulfer, M. K.; Murphy, R. C. (2004) Formation of biologically active oxysterols during ozonolysis of cholesterol
22             present in lung surfactant. J. Biol. Chem. 279: 26,331-26,338.
23       Pulfer, M. K.; Taube, C.; Gelfand, E.; Murphy, R. C. (2005) Ozone exposure in vivo and formation of biologically
24             active oxysterols in the lung. J. Pharmacol. Exp. Ther. 312: 256-264.
25       Rigas, M. L.; Ben-Jebria, A.; Ultman, J. S. (1997) Longitudinal distribution of ozone absorption in the lung: effects
26             of nitrogen dioxide, sulfur dioxide, and ozone exposures. Arch. Environ. Health 52: 173-178.
27       Romieu, I.; Meneses, F.; Ramirez, M.; Ruiz, S.; Padilla, R. P.; Sienra, J. J.; Gerber, M.; Grievink, L.; Dekker, R.;
28             Walda, I.; Brunekreef, B. (1998) Antioxidant supplementation and respiratory functions among workers
29             exposed to high levels of ozone. Am. J. Respir. Crit. Care Med. 158: 226-232.
30       Romieu, I.; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Tellez-Rojo, M. M.; Moreno-Macias, H.; Reyes-Ruiz, N. I.;
31             Del Rio-Navarro, B. E.; Ruiz-Navarro, M. X.; Hatch, G.; Slade, R.; Hernandez-Avila, M. (2002) Antioxidant
32             supplementation and lung functions among children with asthma exposed to high levels of air pollutants. Am.
33             J. Respir. Crit. Care Med. 166: 703-709.
34       Romieu, I.; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. I.; Estela del
3 5             Rio-Navarro, B.; Hernandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and
3 6             antioxidant supplementation influence lung function in relation to ozone exposure in asthmatic children in
37             Mexico City. Thorax 59: 8-10.
38       Ruidavets, J.-B.; Cournot, M.; Cassadou,  S.; Giroux, M.; Meybeck, M.; Ferrieres, J. (2005) Ozone air pollution is
39             associated with acute myocardial infarction. Circulation 111: 563-569.
40       Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, L; Dockery, D. W.; Schwartz, J.; Zanobetti, A.
41             (2000) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
42             pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94, part II.
43       Samet, J. M.; Hatch, G. E.; Horstman, D.; Steck-Scott, S.; Arab, L.; Bromberg, P. A.; Levine, M.; McDonnell,
44             W. F.; Devlin, R. B. (2001) Effect of antioxidant supplementation on ozone-induced lung injury in human
45             subjects. Am. J. Respir. Crit. Care Med. 164: 819-825.
46       Scannell, C.; Chen, L.; Aris, R. M.; Tager, I.; Christian, D.; Ferrando, R.; Welch, B.; Kelly, T.; Balmes, J. R. (1996)
47             Greater ozone-induced inflammatory responses in subjects with asthma. Am. J. Respir. Crit. Care Med.
48             154:24-29.
49       Schierhorn, K.; Hanf, G.; Fischer, A.; Umland, B.; Olze, H.; Kunkel, G.  (2002) Ozone-induced release of
50             neuropeptides from human nasal mucosa cells. Int. Arch. Allergy Immunol. 129: 145-151.
51       Schierhorn, K.; Zhang, M.; Matthias, C.; Kunkel, G. (1999) Influence of ozone and nitrogen dioxide on histamine
52             and interleukin formation in a human nasal mucosa culture system. Am. J. Respir. Cell Mol. Biol.
53             20: 1013-1019.
54       Schwartz, J. (2004) The effects of paniculate air pollution on daily deaths: a multi-city case crossover analysis.
55             Occup. Environ. Med. 61:956-961.
         August 2005                                     8-93         DRAFT -DO NOT QUOTE OR CITE

-------
  1       Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature?
  2             Am. J. Respir. Crit. Care Med. 171: 627-631.
  3       Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.; Ponce de
  4             Leon, A.; Ponka, A.; Rossi, G.; Saez, M; Schouten, J. P. (1996) Methodological issues in studies of air
  5             pollution and daily counts of deaths or hospital admissions. In: St Leger, S., ed. The APHEA project. Short
  6             term effects of air pollution on health: a European approach using epidemiological time  series data.
  7             J. Epidemiol. Commun. Health 50(suppl. 1): S3-S11.
  8       Selwyn, B.  J.; Stock, T. H.; Hardy, R. J.; Chan, F. A.; Jenkins, D. E.; Kotchmar, D. J.; Chapman, R. S. (1985)
  9             Health effects of ambient ozone exposure in vigorously exercising adults. In: Lee, S. D., ed. Evaluation of the
10             scientific basis for ozone/oxidants standards: proceedings of an APCA international specialty conference;
11             November 1984; Houston, TX. Pittsburgh,  PA: Air Pollution Control Association; pp. 281-296. (APCA
12             international specialty conference transactions: TR-4).
13       Sherwin, R. P.; Richters, V.; Kraft, P.; Richters, A. (2000) Centriacinar region inflammatory disease in young
14             individuals: a comparative study of Miami  and Los Angeles residents. Virchows Arch. 437: 422-428.
15       Sin, D. D.; Man, S. F. P. (2003) Why are patients  with chronic obstructive pulmonary disease at increased risk of
16             cardiovascular diseases? Circulation 107: 1514-1519.
17       Sin, D. D.; Wu, L.-L.; Man, S. F. P. (2005) The relationship between reduced lung function and cardiovascular
18             mortality: a population-based study and a systematic review of the literature. Chest 127: 1952-1959.
19       Solway, J.;  Leff, A. R.  (1991) Sensory neuropeptides and airway function. J. Appl. Physiol. 71: 2077-2087.
20       Spektor, D. M.; Lippmann, M.; Thurston, G. D.; Lioy, P.  J.; Stecko, J.; O'Connor, G.; Garshick, E.; Speizer, F. E.;
21             Hayes, C. (1988) Effects of ambient ozone on respiratory function in healthy adults exercising outdoors.
22             Am. Rev. Respir. Dis. 138: 821-828.
23       Sprenger, M.; Maspoli, M. C.; Wernli, H. (2003) Tropopause folds and cross-tropopause exchange: a global
24             investigation based upon ECMWF analyses for the time period March 2000 to February 2001. J. Geophys.
25             Res. (Atmos.) 108(D12): 10.1029/2002JD002587.
26       Stenfors, N.; Pourazar, J.; Blomberg, A.; Krishna, M. T.; Mudway, I.; Helleday, R.; Kelly, F. J.; Frew, A. J.;
27             Sandstrom, T. (2002) Effect of ozone on bronchial mucosal inflammation in asthmatic and healthy subjects.
28             Respir. Med.  96: 352-358.
29       Stieb, D. M.; Burnett, R. T.; Beveridge, R. C.; Brook, J. R. (1996) Association between ozone and asthma
30             emergency department visits in Saint John, New Brunswick, Canada. Environ. Health Perspect.
31             104:1354-1360.
32       Stieb, D. M.; Judek, S.; Burnett, R. T. (2003) Meta-analysis of time-series studies of air pollution and mortality:
3 3             update in relation to the use of generalized additive models. J. Air Waste Manage. 53: 258-261.
34       Stohl, A. (2001) A 1-year Lagrangian "climatology" of airstreams in the Northern Hemisphere  troposphere and
35             lowermost stratosphere. J. Geophys. Res. (Atmos.) 106: 7263-7279.
36       Tepper, J. S.; Wiester, M. J.; Weber, M. F.; Menache, M.  G. (1990) Measurements of cardiopulmonary response in
37             awake rats during acute exposure to near-ambient concentrations of ozone. J. Appl. Toxicol. 10: 7-15.
38       Thurston, G. D.; Ito, K. (2001) Epidemiological studies of acute ozone exposures and mortality. J. Exposure Anal.
39             Environ. Epidemiol. 11: 286-294.
40       Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
41             hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers.
42             J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
43       Thurston, G. D.; Ito, K.; Hayes, C.  G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital admissions and
44             summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ. Res.
45             65:271-290.
46       Tolbert, P. E.; Mulholland,  J. A.; Macintosh, D. L.; Xu, F.; Daniels, D.; Devine, O. J.; Carlin, B. P.; Klein, M.;
47             Dorley, J.; Butler, A. J.; Nordenberg, D. F.; Frumkin, H.; Ryan, P. B.; White, M. C. (2000) Air quality and
48             pediatric emergency room visits for asthma in Atlanta, Georgia. Am. J. Epidemiol. 151:  798-810.
49       Trenga, C. A.; Koenig, J. Q.; Williams, P. V. (2001) Dietary antioxidants and ozone-induced bronchial
50             hyperresponsiveness in adults with asthma. Arch. Environ. Health 56: 242-249.
51       Tuazon, E.  C.; Winer, A. M.; Pitts, J. N., Jr. (1981) Trace pollutant concentrations in a multiday smog episode in the
52             California South Coast Air Basin by long path length Fourier transform infrared spectroscopy. Environ. Sci.
53             Technol. 15: 1232-1237.
54       U.S. Environmental Protection Agency. (1990) National air quality and emissions trends report, 1988.  Research
5 5             Triangle Park, NC: Office of Air Quality Planning  and Standards; report no. EPA/450/4-90/002.
         August 2005                                    8-94         DRAFT -DO NOT QUOTE OR CITE

-------
  1       U.S. Environmental Protection Agency. (1996a) Air quality criteria for ozone and related photochemical oxidants.
  2             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
  3             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
  4             http://cfpub2.epa.gov/ncea/.
  5       U.S. Environmental Protection Agency. (1996b) Review of national ambient air quality standards for ozone:
  6             assessment of scientific and technical information. OAQPS staff paper. Research Triangle Park, NC:
  7             Office of Air Quality Planning and Standards; report no. EPA/452/R-96/007. Available from: NTIS,
  8             Springfield, VA; PB96-203435. Available: http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_pr_sp.html
  9             (9 September 2003).
10       U.S. Environmental Protection Agency. (2003) National air quality and emissions trends report. 2003 special studies
11             edition. Research Triangle Park, NC: Office of Air Quality Standards; Emissions Monitoring and Analysis
12             Division; report no. EPA 454/R-03-005. Available: http://www.epa.gov/air/airtrends/aqtrnd03/toc.html
13             (27 August, 2004).
14       U.S. Environmental Protection Agency. (2004a) Air quality criteria for paniculate matter. Research Triangle Park,
15             NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
16             http://cfpub.epa.gov/ncea/ [9 November, 2004].
17       U.S. Environmental Protection Agency. (2004b) The ozone report: measuring progress through 2003. Research
18             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/K04-001. Available:
19             http://www.epa.gov/air/airtrends/pdfs/2003ozonereport.pdf [12 May, 2005].
20       Ultman, J. S.; Ben-Jebria, A.; Arnold, S. F. (2004) Uptake distribution of ozone in human lungs: intersubject
21             variability in physiologic response. Boston, MA: Health Effects Institute.
22       Vagaggini, B.; Taccola, M.; Clanchetti, S.; Carnevali, S.; Bartoli, M. L.; Bacci, E.; Dente, F. L.; Di Franco, A.;
23             Giannini, D.; Paggiaro, P. L. (2002) Ozone exposure increases eosinophilic airway response induced by
24             previous allergen challenge. Am. J. Respir. Crit.  Care Med. 166: 1073-1077.
25       Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
26             pollution. Environ. HealthPerspect. Ill: 45-51.
27       Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of submicron
28             particle exposure. Environ. Health Perspect. 108: 1139-1145.
29       Wattiez, R.; Noel-Georis, L; Cruyt, C.; Broeckaert, F.; Bernard, A.; Falmagne, P. (2003) Susceptibility to oxidative
30             stress: proteomic analysis of bronchoalveolar lavage from ozone-sensitive and ozone-resistant strains of mice.
31             Proteomics 3: 658-665.
32       Wayne, R. P. (1991) Chemistry of Atmospheres: an introduction to the chemistry of the atmospheres of Earth, the
33             planets, and their satellites. 2nd ed. New York, NY: Oxford University Press,  Inc.
34       Weschler, C. J. (2004) Chemical reactions among indoor pollutants: what we've learned in the new millennium.
35             Indoor Air 14(suppl. 7): 184-194.
36       Wilson, A. M.; Wake, C. P.; Kelly, T.; Salloway, J. C. (2005) Air pollution, weather, and respiratory emergency
37             room visits in two northern New England cities: an ecological time-series study. Environ. Res. 97: 312-321.
38       Woo,  K.-S.; Chen, D.-R.; Pui, D. Y. H.; McMurry, P. H. (2001) Measurement of Atlanta aerosol size distributions:
39             observations of ultrafine particle events. Aerosol Sci. Technol. 34: 75-87.
40       Woodwell, D. A.; Cherry, D.  K. (2004) National Ambulatory Medical Care Survey: 2002 summary. Hyattsville,
41             MD: National Center for Health Statistics; DHHS publication no. (PHS) 2004-1250. (Advance data from vital
42             and health statistics; no. 346). Available: http://www.cdc.gov/nchs/data/ad/ad346.pdf [3 August, 2005].
43       World Health Organization. (2004) Meta-analysis of time-series studies and panel studies of paniculate matter (PM)
44             and ozone (O3): report of a WHO task group. Copenhagen, Denmark: WHO Regional Office for Europe;
45             document no. EUR/04/5042688. Available: http://www.euro.who.int/document/E82792.pdf
46             [18 November, 2004].
47
48
         August 2005                                     8-95         DRAFT -DO NOT QUOTE OR CITE

-------
 i        9.  ENVIRONMENTAL EFFECTS:  OZONE EFFECTS
 2                  ON VEGETATION AND ECOSYSTEMS
 3
 4
 5     9.1   INTRODUCTION
 6          A number of ozone (O3) effects studies were published between 1996 and 2004, and they
 7     are reviewed in this document in the context of the previous O3 air quality criteria documents
 8     (AQCDs) (U.S. Environmental Protection Agency, 1978, 1986,  1992, 1996). Data published
 9     since 1996 continue to support the conclusions of previous O3 AQCDs that there is strong
10     evidence that ambient O3 concentrations cause foliar injury along with growth and yield damage
11     to numerous common and economically valuable plant and tree  species. Research to date has
12     continued to be focused at the species level with very few studies at the ecosystem level. The
13     lack of quantification of biotic and abiotic factors impinging on  the individual to population
14     organizational levels results in a limited ability to scale O3 responses to the ecosystem level.
15     Therefore, a high degree of uncertainty remains in our ability to assess ozone risk to ecological
16     resources and the services they provide.
17          In general, there has been a shift away from chamber studies in favor of more field-based
18     approaches, although chamber exposures still dominate the effects literature. Field-based
19     approaches include surveys of visible injury, as well as physiological and growth studies using
20     the non-chambered free-air CO2 exposure (FACE) systems. The FACE systems have
21     substantiated earlier growth and yield results for crop and tree species obtained in open-top
22     chamber  (OTC) systems.  Increased emphasis has also been placed on quantifying aspects of
23     ozone uptake to better link ambient exposure monitoring with plant/tree response. Much of the
24     progress in quantifying uptake has occurred in Europe in the development of their ozone air
25     quality management tool, the "critical level". The research has developed exposure-response
26     functions for several crops and tree seedlings using OTC studies, as well as in developing and
27     testing models that simulate uptake. Evaluation of this new information has added to our
28     knowledge and provides new research directions, but has not fundamentally altered the
29     conclusions of 1996 O3 AQCD (U.S.  Environmental Protection Agency, 1996).
30          It is well known that O3 is phytotoxic and that toxicity occurs only if O3 or its reaction
31     products  reach the target tissues in the plant cell. Recent studies have provided an increased

       August 2005                              9-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1      understanding of how ozone interacts with the plant at the cellular level. This increased
 2      understanding of cellular-level O3 effects have translated into better models, more detailed
 3      schema of how O3 alters much of the basic metabolism of plants and how to construct an index
 4      that more aptly captures the species, climate, and site factors that alter uptake.  These results
 5      have and will continue to lead to a better quantification of exposure and effect. However, the
 6      translation of these mechanisms into how O3 is involved in altered cell metabolism and
 7      subsequent reductions in whole-plant productivity and ecosystem-level responses remain to be
 8      more been fully resolved.
 9           The ensuing sections of this chapter (Section 9.2 to 9.8) are not intended to provide a
10      complete review of the environmental effects of O3, but rather an assessment of key information
11      published since the 1996 O3 AQCD.  More detailed discussion of the research since 1996 is
12      provided in Chapter 9 Annex Sections AX9.1 to AX9.7 (in Volume 3 of this document).  The
13      framework for Chapter 9 follows the environmental effects chapter of the 1996 O3 AQCD. First,
14      an overview of various methodologies that have been, and continue to be, central to the
15      quantification of O3 effects on vegetation is provided in Section 9.2 below (see Section AX9.1
16      for more detailed discussion). The adequacy of each methodology is discussed in the context of
17      developing statistically robust data appropriate for assessing the risk of O3 to vegetation
18      resources.  In Section 9.3, research is then reviewed from the molecular to the biochemical and
19      physiological levels in plants that are impacted, which offers insight into the mode of action of
20      O3 (see also AX9.2). The manner in which plants respond to O3, as influenced by the numerous
21      biotic and abiotic factors present in the environment, is next discussed (see also AX9.3).
22      Quantifying these various modifiers is critical to being able to scale the response of individual
23      plants to the community level and across varied landscapes and climates and is needed for
24      regional to national assessments of risk. The development of indices of exposure or O3 uptake is
25      discussed in the context of their adequacy to realistically describe the ambient concentration-
26      response relationships (see also AX9.4). The exposure-response relationships for a large number
27      of crop species and cultivars, native vegetation, and tree species are reviewed, tabulated,  and
28      compared to form the basis for an assessment of the potential risk of current levels of O3 on
29      vegetation resources (see also AX9.5).  Available research by which to assess the impact of O3
30      on ecosystems is also reviewed, along with the potential data available for estimating the loss of
        August 2005                               9-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1     various ecosystem services (see also AX9.6). Finally, available research on the economic
 2     impact of ozone effects on vegetation resources is briefly discussed (see also AX9.7).
 3
 4
 5     9.2  METHODOLOGIES USED IN VEGETATION RESEARCH
 6          New methodological advancements since 1996 have not fundamentally altered our
 7     understanding of O3 effects on plants or ecosystems. Most of the new information confirms
 8     earlier conclusions and provides additional support for OTC use in assessing sensitive species
 9     and developing exposure-response relationships. A more in-depth discussion of this topic can be
10     found in Annex Section AX9.1.
11          The majority of ozone effects studies are fumigation studies conducted in controlled
12     chambers, as noted in the 1996 O3 AQCD (U.S. Environmental Protection Agency,  1996). That
13     document noted that OTCs represented the best technology for determining statistically robust
14     exposure response models of O3 and crop yield and plant biomass at that time. While OTCs are
15     still the best method for conducting controlled exposures of varying length and frequency for
16     developing exposure-response relationships, several new approaches have been applied toO3
17     effects research — most notably free-air exposure or "plume" systems. Free-air exposure
18     systems (FACE) eliminate many of the concerns raised about closed or open-top chamber
19     experiments including small plot size,  altered microclimate within OTCs, and the effect of
20     charcoal filtering on overall air quality within OTCs. FACE systems have increased our
21     understanding in some areas; and results from FACE studies have, on the whole, confirmed what
22     was already understood or hypothesized about how plants and plant assemblages respond to O3.
23     Some shortcomings of using plume systems in O3 research have also been identified, namely the
24     relatively poor control of exposure levels, the presence of "hotspots" and the inability to
25     decrease O3 concentrations to below ambient levels when ambient concentrations are phytotoxic.
26     Nonetheless, the application of FACE  systems and other open-air systems to ozone  exposure
27     research have greatly helped our scaling efforts and are, perhaps, the best approach  for studying
28     the response of plant species mixtures  to O3 (Nussbaum and Fuhrer, 2000).
29          One of the advantages of the application of plume systems to O3 research is the ability to
30     compare response of plants in open-field systems with results from OTCs. In particular, studies
31     with quaking aspen (Populus tremuloides L.) performed in OTCs, FACE, and also sites along an

       August 2005                              9-3        DRAFT-DO NOT QUOTE OR CITE

-------
 1      ambient ozone gradient showed that ozone symptom expression was generally similar across
 2      these methodologies, supporting the previously observed level of variation among aspen clones
 3      in OTC studies (Isebrands et al., 2000, 2001; Karnosky et al., 1999). While this perhaps
 4      represents the first time direct comparisons were made, it supports the use of OTC data in the
 5      development of O3 response functions for individual species. Concerns raised earlier about
 6      microclimate differences in chambers and the role of "chamber effects" (Fuhrer, 1994; Manning
 7      and Krupa, 1992) still persist; however, most evidence suggests that chamber effects result in
 8      altered O3 uptake without altering the fundamental response of plants to ozone, thus reducing
 9      uncertainty in the use of data from OTCs. Extrapolation of the results from chamber studies
10      depends on fully characterizing temperature, light, turbulence, and other chamber characteristics
11      during exposures (Nussbaum and Fuhrer, 2000), but study design is equally important.
12      Conducting studies with a large number of plant species across regions of the country where
13      those species are indigenous is important in capturing  regional climatic differences to reduce the
14      uncertainty associated with extrapolating composited response functions across regions and to
15      identify relative risk to vegetation in relation to given  O3 exposure values (U.S. Environmental
16      Protection Agency, 1996).
17           The lack of rural monitors continues to be a major problem in the characterization of O3
18      exposures in remote areas, as well as in linking effects to exposure in natural ecosystems.  Since
19      the 1996 O3 AQCD, the use of passive samplers has expanded monitoring efforts to include
20      remote areas that were previously uncharacterized.  This has greatly enhanced our ability to link
21      ozone symptomology with elevated O3 exposure in such remote areas.  However, passive
22      samplers do not capture the temporal dynamics of exposure. Therefore, passive samplers  cannot
23      substitute for active monitors when attempting to link  exposure dynamics to plant response or
24      when developing exposure/dose-response relationships of much value as inputs for the standard
25      setting process. To overcome this problem, Krupa et al. (2001, 2003) used models  and data from
26      a collocated O3 monitor to estimate the underlying frequency distribution of hourly O3
27      concentrations from passive samplers. Future development of passive monitor technology and
28      data synthesis techniques holds promise, particularly since it is unlikely that extensive O3
29      monitoring networks will be established in rural areas  in the near future.
30           Exclusion methods that employ protective chemicals such as ethylenediurea (EDU) are the
31      least disruptive of ambient culture conditions in the  field, as noted in the 1996 O3 AQCD.

        August 2005                               9-4        DRAFT-DO NOT QUOTE OR CITE

-------
 1     However, the level of protection afforded by EDU is site- and species-specific and is subject to
 2     local meteorologic conditions. In addition, new evidence suggests that EDU does not always
 3     have greater effects at higher O3 exposures and that the degree of protection by EDU largely
 4     depends on environmental conditions. Because of the variability observed in the level of
 5     protection provided, and the fact that mechanisms of protection afforded by EDU and other
 6     exclusion methods are unknown, caution is needed in applying this approach to the study of O3
 7     effects in the field.
 8          Advancements in biomonitoring have been made since the 1996 O3 AQCD, primarily in
 9     the area of identification and symptom verification of sensitive species (Flagler, 1998; Krupa
10     et al., 1998; Innes et al., 2001; Smith et al., 2003). The U.S. Department of Agriculture (USDA)
11     Forest Service continues its program to monitor ozone effects in forested ecosystems throughout
12     the United States. Currently, 33 states participate in the program, which uses a grid system to
13     identify the location of plants showing foliar injury. Although results cannot be used for
14     developing exposure-response relationships or for quantifying responses to O3, they can provide
15     an annual assessment and correlative information regarding the extent of O3 injury occurring
16     across many regions of the United States.
17
18
19     9.3  SPECIES RESPONSE/MODE-OF-ACTION
20          There are several steps in the process of O3 uptake and toxicity that are now better
21     understood than in 1996.  These advancements are important in refining hypotheses on O3 uptake
22     and mode of action on plants and in developing a flux-based index for use in quantifying
23     response and, ultimately, for potential use in developing a secondary national ambient air quality
24     standard (SNAAQS). The new information available on the mode of action of O3 is, in part, a
25     result of improved molecular tools for following rapid changes that occur within the leaf (Pell
26     et al., 1997; Sandermann, 2000; Ward et al.,  1991). This new information is discussed in greater
27     detail in Annex Section AX9.2.
28          Clearly, many changes occur within hours or possibly days following O3 exposure
29     (Sandermann,  1998). However, other O3 effects take longer to occur and tend to be most
30     obvious only under exposure to low O3 concentrations for long periods (Andersen et al., 1997;
31     Hogsett et al.,  1989; Langebartels et al., 1997). These low-exposure chronic effects have been

       August 2005                              9-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      linked to the senescence process or some physiological response very closely linked to
 2      senescence (e.g., translocation, reabsorption, allocation of nutrients and carbon).
 3           Langebartels et al. (1997) discussed "memory" or "carry-over effects" within the plant to
 4      explain sensitivity to frost in the winter following summertime O3 exposure.  Others have argued
 5      that this sensitivity is due to the nutrient status of the tree during the over-wintering phase of its
 6      life and to chronic (on-going, less severe levels with fewer peaks at very high levels) exposure to
 7      ambient O3 inducing (1) mineral nutrient deficiency; (2)  alterations of normal metabolism,
 8      including translocation and allocation of carbohydrates and probably nitrogen; and/or (3)
 9      disturbance of normal transpiration and diurnal cycling, leading to water stress (Schmieden and
10      Wild, 1995). While general nutrient concentrations within the foliage may not occur, localized
11      deficiencies might.  This is  difficult to observe or prove without  a great deal of work on all
12      portions of a tree and without a general hypothesis of what is occurring.
13           It is important to note that the dramatic strides made over the last few years in
14      understanding the genetic make-up of plants, gene control, and signal transduction/control will
15      accelerate in the future and  translate into better models of the hypotheses listed above as well as
16      more detailed schemes of how  O3 alters basic plant metabolism.  Thus, while our understanding
17      of how O3 interacts with the plant at the cellular level has dramatically improved (Assmann,
18      2003; Assmann and Wang,  2001; Rao and Davis, 2001),  the translation of those mechanisms
19      into how O3 is involved with altered cell metabolism and the subsequent reductions in whole
20      plant productivity and other physiological facts has not yet been  fully achieved.  As the
21      understanding of wounding responses in plants and more information on genome details and
22      varied plant mutants becomes available, the cellular and  physiological responses of plants to O3
23      exposures are slowly becoming clearer. However, more  studies  on a larger variety of species are
24      needed before this type of information can be incorporated into indices of response and for
25      consideration in developing SNAAQS.
26
27
28      9.4  MODIFICATION OF FUNCTIONAL AND GROWTH RESPONSES
29           It has been known for decades that several factors,  both biotic and abiotic, alter plant
30      response to ozone. However, only a few studies  reported since the 1996 O3 AQCD have
31      improved our understanding of the role of these interactions in modifying plant O3 response.

        August 2005                               9-6        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Quantifying how these interactions alter plant O3 response is a critical first step to reducing the
 2      uncertainty in extrapolating individual plant responses to higher levels of biological
 3      organization, e.g., ecosystems. None of the recent studies have significantly improved our
 4      ability to quantify the degree to which these factors modify plant O3 response; however, they
 5      have reinforced the conclusions of the 1996 O3 AQCD with regard to factors known to alter
 6      plant response to O3. This new information is discussed in greater detail in Annex Section
 7      AX9.3.
 8           In the area of biotic interactions, new evidence with regard to insect pests and diseases (see
 9      Docherty et al. (1997) and Fliickiger et al. (2002) for recent reviews) has not reduced the
10      uncertainties noted in the 1996 O3 AQCD.  Most interactions thought to affect crops, forest trees
11      and other natural vegetation have yet to be studied. Recent studies have supported the earlier
12      conclusion that O3 often increases the likelihood and success of insect attacks, but only with
13      respect to chewing insects (e.g., Percy et al., 2002; Kopper and Lindroth, 2003).  With the
14      economically important group of sucking insects (such as aphids), no clear trends have been
15      revealed in the latest studies (see reviews by Docherty et al., 1997; Fliickiger et al., 2002).
16      Hence, although it seems likely that some insect problems could increase as a result of greater O3
17      levels, we are still far from being able to predict the nature of any particular O3-plant-insect
18      interaction, its likelihood, or its severity.
19           The situation is somewhat clearer with respect to interactions involving facultative
20      necrotrophic plant pathogens, with O3 exposure generally contributing to increased disease
21      (Fliickiger et al., 2002).  With obligate biotrophic fungal, bacterial, and nematode diseases,
22      however, twice as many reports indicate O3-induced inhibitions than enhancements. This pattern
23      is supported by the concept put forth by Dowding (1988) that pathogens that benefit from
24      damage to cells are enhanced by pollution stress of their hosts, whereas pathogens and pests that
25      require healthy hosts are depressed by pollution  stress. The frequent reports that infection by
26      obligate biotrophs reduces the severity of O3-induced foliar injury (e.g., Schraudner et al., 1996)
27      does not result in true "protection", as the disease per se causes negative effects on the host
28      plant. With obligate biotrophs, the nature of any interaction with  O3 is probably dictated by the
29      unique, highly specific biochemical relationships between the pathogen and the host plant. At
30      this time, therefore, although some diseases may become more widespread or severe as a result
        August 2005                                9-7         DRAFT-DO NOT QUOTE OR CITE

-------
 1      of exposure to O3, it is still not possible to predict exactly which diseases are likely to present the
 2      greatest risks to crops and forests.
 3           Recent studies have not greatly added to our understanding of the nature of interactions
 4      between O3 and root symbionts, but have served to support conclusions put forth in the 1996 O3
 5      AQCD. Several studies have indicated that the functioning of tree root symbioses with
 6      mycorrhizae may be adversely affected by O3 (e.g., Kytoviita et al., 2001), but there is also
 7      evidence that the presence of mycorrhizae may overcome O3-enhanced root diseases (Bonello
 8      et al., 1993). There is also evidence that O3 may encourage the spread of mycorrhizae to the
 9      roots of uninfected trees. The role of O3 in altering root symbionts, its interactions with soil
10      organisms, and the subsequent feedback effects on plant growth represent one of the greatest
11      areas of uncertainty in assessing the influence of O3 on ecosystems (Andersen, 2003).
12           The few recent studies of the impact of O3 on intraspecific plant competition confirmed
13      that grasses frequently show greater resilience than other types of plants. In grass-legume
14      pastures, the leguminous species tend to suffer greater growth inhibition (Johnson et al., 1996;
15      Nussbaum et al., 2000).  The suppression of Ponderosa pine (Pinusponderosa Laws.) seedling
16      growth by blue wild-rye grass (Elymus glaucus Buckl.) was markedly increased by O3 (Andersen
17      et al., 2001). However, we are far from being able to predict the outcome of the impact of O3 on
18      specific competitive situations, such as success!onal plant communities or crop-weed
19      interactions.
20           Physical or abiotic factors play a large role in modifying plant response to O3,  and new
21      information is available that supports the conclusions of the 1996 O3 AQCD. Although some
22      recent field studies have indicated that O3 impact significantly increases with increased ambient
23      temperature (Ball et al., 2000; Mills et al., 2000), other studies have indicated that temperature
24      has little effect (Balls et al., 1996; Fredericksen et al., 1996).  Temperature affects the rates of all
25      physiological processes based on enzyme-catalysis and diffusion; each process and overall
26      growth (the integral of all processes) has a distinct optimal temperature range. It is important to
27      note that a plant's response to changes in temperature will depend on whether it is growing near
28      its optimum temperature for growth or near its maximum temperature (Rowland-Bamford,
29      2000).  But temperature is unquestionably an important variable affecting plant O3 response in
30      the presence of the elevated CO2 levels contributing to global climate change.  In contrast,
31      evidence continues to accumulate to indicate  that O3 exposure sensitizes plants to low

        August 2005                               9-8        DRAFT-DO NOT QUOTE OR CITE

-------
 1      temperature stress (Colls and Unsworth, 1992) and, also, that O3 decreases below-ground
 2      carbohydrate reserves, which may lead to responses in perennial species ranging from rapid
 3      demise to impaired growth in subsequent seasons (i.e., carry-over effects) (Andersen et al.,
 4      1997).
 5           Light, a component of the plant's physical environment, is an essential "resource" whose
 6      energy content drives photosynthesis and CO2 assimilation.  It has been suggested that increased
 7      light intensity may increase O3 sensitivity of light-tolerant species while decreasing that of
 8      shade-tolerant species, but this appears to be an oversimplification with many exceptions.  As
 9      pointed out by Chappelka and Samuelson (1998) and Topa et al. (2001), the interaction between
10      O3 sensitivity and light environment is complicated by developmental stage as well as the light
11      environment of individual leaves in the canopy.
12           Although the relative humidity of the ambient air has generally been found to increase the
13      adverse effects of O3 by increasing stomatal conductance, and thereby increasing O3 flux,
14      abundant evidence also indicates that the ready availability of soil moisture results in greater O3
15      sensitivity (Mills, 2002). The partial "protection" against the adverse effects of O3 afforded by
16      drought (as noted in previous O3 AQCDs) has been observed in field experiments and modeled
17      in computer simulations (Broadmeadow and Jackson, 2000). There is also compelling evidence
18      that O3 can predispose plants to drought stress (Maier-Maercker, 1998).  Hence, the response
19      will depend to some extent upon the sequence in which the stresses occur, but, even though the
20      nature of the response is largely species-specific, successful  applications of model simulations
21      have led to larger-scale predictions of the consequences of O3 * drought interactions.  However,
22      regardless of the  interaction, the net result on short-term growth is negative; although in tree
23      species, other responses such as increased water use  efficiency could benefit long-term survival.
24           Somewhat analogous to temperature, it appears that any shift away from the nutritional
25      optimum may lead to greater O3 sensitivity; but the shift would  have to be substantial before a
26      significant effect on O3 response was observed.  Mineral nutrients in the soil, other gaseous air
27      pollutants, and agricultural chemicals constitute chemical factors in the environment.  The
28      evidence regarding interactions with specific nutrients is still contradictory: some experimental
29      evidence indicates that low general fertility increases sensitivity to O3 (Whitfield et al., 1998;
30      Landolt et al., 1997),  although others have found less sensitivity with decreased fertility
31      (Cardoso-Vilhena and Barnes, 2001).  Simulation modeling  of trees suggests that nutrient

        August 2005                               9-9        DRAFT-DO NOT QUOTE OR CITE

-------
 1      deficiency and O3 may act less than additively, but too many examples of contrary trends exist to
 2      permit any sweeping conclusions at this time.
 3           Interactions of O3 with other air pollutants have received relatively little recent attention
 4      since 1996 (see Barnes and Wellburn [1998] and Fangmeier et al. [2002] for recent reviews).
 5      The situation with SO2 remains inconsistent, but SO2 seems unlikely to pose any additional risk
 6      to those related to other individual pollutants.  With NO and NO2, the situation is complicated by
 7      their nutritional value as a N source. Much more investigation is needed before we will be able
 8      to predict the outcomes of different O3-NO-NO2 scenarios.  The latest research into O3 * acid
 9      rain interactions has confirmed that, at realistic acidities, significant interactions are unlikely
10      (Momen et al., 1997; 1999; Laurence et al., 1997; Sayre and Fahey, 1999).  A continuing lack of
11      information precludes our offering  any generalizations about interactive effects of O3 with NH3,
12      FTP, or heavy metals. More evidence has been reported for protective effects against O3 afforded
13      by the application of fungicides (Wu and Tiedemann, 2002).
14           Considerable emphasis during the last decade has been placed on research evaluating
15      potential O3 interactions with the components of global climate change: increased atmospheric
16      CO2, increased mean global temperatures, and increased surface level UV-B radiation.
17      However, it must be noted that most of these studies have tended to regard increased CO2 levels
18      and increased mean temperatures as unrelated  phenomena.  Experiments into the effects of
19      doubled CO2 levels at today's mean ambient temperatures are of questionable value in trying to
20      assess the impact of climate change on responses to O3.  To date, the limited experimental
21      evidence and that obtained by computer simulation suggest that even though an enriched CO2
22      atmosphere (-600 ppm) would more than offset the impact of O3 on responses as varied as wheat
23      (Triticum aestivum L.) yield or young Ponderosa pine growth, the concurrent increase in
24      temperature would reduce, but probably not eliminate, the net gain (Batts et al., 1997; Van Oijen
25      and Ewart, 1999; Constable et al., 1996). There is also some recent evidence that O3 and UV-B
26      interact in their effects on plant injury and photosynthesis (Schnitzler et al., 1999), but additional
27      research is needed to fully understand how O3  interacts with multiple climate change factors.
28
        August 2005                              9-10        DRAFT-DO NOT QUOTE OR CITE

-------
 1      9.5  EFFECTS-BASED AIR QUALITY EXPOSURE INDICES
 2           Since the 1996 O3 AQCD , there has been no direct experimental testing of the adequacy of
 3      exposure indices proposed in 1996; therefore, there is no new information to alter the basic
 4      conclusions put forth in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996) with
 5      regard to exposure indices.  A more detailed discussion of effects-based air quality indices can
 6      be found in Annex Section AX9.4.
 7           Exposure indices are metrics that relate measured plant damage (i.e., reduced growth) to
 8      monitored ambient O3 concentrations over time to provide a consistent metric for reviewing and
 9      comparing exposure-response effects obtained from various studies. The 1996 O3 AQCD (U.S.
10      Environmental Protection Agency, 1996) focused on the research used to develop various
11      exposure indices to quantify growth and yield effects in crops, perennials, and trees (primarily
12      seedlings), and not foliar injury. The proposed indices included various functional and statistical
13      summaries of monitored hourly O3 concentrations over designated time periods.  The indices
14      were developed through regression analyses of earlier exposure studies and was  accomplished by
15      ordering the measured responses of growth and/or yield of crops and tree (seedling) species in
16      response to O3. Their development focused on consideration and inclusion of some, but not all,
17      the factors that affect O3 uptake and expression of effects (e.g., Lee et al. [1988]).
18           The few studies that have been published since the 1996  O3 AQCD continue to support the
19      earlier conclusions, including the importance of peak concentrations, and the duration and
20      occurrence of O3 exposures in altering plant growth and yield.  In addition, a large body of new
21      research, mostly out of Europe, addresses the need for an index related to the actual uptake of O3
22      by the plant and the flux of O3 from the atmosphere to the O3 affected plant tissues.  Despite
23      additional research linking estimates of flux with plant response since 1996, information is still
24      insufficient to identify a flux-based model that incorporates the necessary complexity across
25      space and time to be non-site or non-species specific. Based on the current state of knowledge,
26      exposure indices that cumulate and differentially weight the higher hourly average O3
27      concentrations, but include the mid-level values, still represent the best approach for relating
28      vegetation effects to O3 exposure in the United States.
29           The new studies have also substantiated earlier conclusions on the role of exposure
30      components including concentration,  duration, and exposure patterns in determining plant
31      growth response to O3 (Oksanen and Holopainen, 2001; Yun and Laurence,  1999).  Recent

        August 2005                              9-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      studies using different exposure patterns have confirmed earlier studies on the role of higher
 2      concentrations and exposure duration (Nussbaum et al., 1995). A role for higher concentrations
 3      is inferred based on improved air quality in regions in the Western United States (Lefohn and
 4      Shadwick, 2000). For example, the O3 reductions in the San Bernardino Mountain area since the
 5      late 1970s are associated with reductions in the higher hourly average O3 concentrations, the
 6      number of hours of concentrations >0.95 ppm, and the cumulative concentration-weighted
 7      exposure index (Lee et al., 2003).  The mid-range concentrations appeared to be relatively
 8      unchanged or even slightly increasing over the period of 1980 to 2000. General forest
 9      improvement has been reported following a decrease of O3 along a decreasing gradient of
10      exposure (Miller and Rechel, 1999; Arbaugh et al., 2003;  Tingey et al., 2004). These studies
11      suggest the focus should be on the higher O3 concentrations, while including the lower levels,
12      when estimating the  effects of O3 precursor emission reduction strategies on vegetation.
13          New studies have demonstrated the potential disconnection of peak events and maximal
14      stomatal conductance at xeric to mesic sites in California  (Panek et al., 2002; Grulke et al., 2002;
15      Panek, 2004). In addition, a few studies have indicated that O3 uptake during nighttime hours is
16      greater than previously thought (Grulke et al., 2004; Massman, 2004); and a review of the
17      literature suggests a large number of species exhibit some degree of conductance at night
18      (Musselman and Minnick, 2000). These studies  suggest a reconsideration of cumulating
19      exposure 24 h/day and not just during daylight hours in exposure index determinations. This
20      lack of coincidence in temporal patterns of conductance and peak ambient concentrations
21      introduces uncertainty in assessing the impact of O3.  The use of an exposure index that does not
22      consider regionally unique climate and site factors that modify stomatal conductance may, as a
23      result, under- or over- estimate growth effects. The shortcomings of an ambient exposure-based
24      index is especially apparent when assessing the potential impact of O3 across broad climatic
25      regions of the United States or Europe.  Various  means to overcome this potential problem were
26      addressed with several new studies; one solution would be to add other components to the
27      present statistical summaries of exposure indices (e.g., meteorological) to develop flux-based
28      indices.  However, the increased biological and meteorological information in these indices may
29      make them more regional in their applicability.
30          A number of studies have taken a flux-based approach to improve upon the
31      concentration-based  (i.e., exposure indices) approach as a means to address the issue of

        August 2005                              9-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1      assessing risk of O3 across different climatic regions.  The European acceptance and use of the
 2      flux-based critical values is, in part, a recognition of the landscape scaling problems associated
 3      with ambient exposure-based indices. A great deal of progress has occurred in developing and
 4      testing stomatal models that may be generally applicable across certain vegetation types
 5      (Danielsson et al., 2003; Emberson et al., 2000; Griinhage and Jager, 2003; Matyssek et al.,
 6      2004; Pleijel et al., 2000).  While a flux-based approach is preferred, a cautionary argument was
 7      advanced in a few publications based on the nonlinear relationship between O3 uptake and foliar
 8      injury (growth was not assessed).  The concern is that not all O3 stomatal uptake results in a
 9      yield reduction, which depends to  some degree on the amount of internal detoxification
10      occurring with each particular species.  Those species having high amounts of detoxification
11      potential may, in fact, show little relationship  between O3 stomatal uptake and plant response
12      (Musselman and Massman, 1999).
13          Given the current state of knowledge and the best available data, exposure indices that
14      cumulate and differentially weight the higher hourly average concentrations, and also include the
15      mid-level values, continue to offer the most defensible approach for use in developing response
16      functions and comparing studies as well as future indices for vegetation protection.  A large
17      database exists that has been used for establishing exposure-response relationships; however, at
18      this time, such a database does not exist for relating O3 flux to growth response.
19          It is anticipated that as the  overlapping relationships  of conductance, concentration, and
20      defense mechanisms are better defined, the flux-based indices may be able to predict vegetation
21      injury and/or damage across varied landscapes and climates with more accuracy than the
22      exposure-response models. However, it is unclear that such is the case at this time.  The
23      translation of these indices from research and  assessment tools to air quality standards has the
24      additional need to be simple, understandable, and adaptive to a manageable monitoring program.
25
26
27      9.6  OZONE EXPOSURE-PLANT RESPONSE RELATIONSHIPS
28          Data published since 1996 continue to support the  conclusions of previous O3 AQCDs that
29      there is strong evidence that ambient concentrations of O3  cause foliar injury and growth and
30      yield damage to numerous common and economically valuable plant and tree species. For
31      annual vegetation, the data summarized in Table AX9-16 (see Annex Section AX9.5) show a

        August 2005                              9-13       DRAFT-DO NOT QUOTE OR CITE

-------
 1     range of growth and yield responses both within species and among species. Nearly all of these
 2     data were derived from studies in OTCs, with only two studies using open-air systems in the
 3     United Kingdom (Ollerenshaw et al., 1999; Ollerenshaw and Lyons, 1999). It continues to be
 4     difficult to compare studies that report O3 exposure using different indices, such as AOT40,
 5     SUM06, W126, or 7-h or 12-h mean values.
 6
 7          The AOT40, SUM06, and W126 indices are defined as follows:
 8     AOT60:  the seasonal sum of the difference between an hourly concentration above the
 9     threshold  value of 60 ppb, minus the threshold value of 60 ppb;
10
11     SUM06:  the seasonal sum of hourly concentrations at or above the threshold value of 60 ppb;
12     and
13
14     W126:  a sigmoid functional weighting of all hourly concentrations for the season.
15
16          When such index comparisons can be made, the results of recent research confirm earlier
17     results summarized in the 1996 O3 AQCD  (U.S. Environmental Protection Agency, 1996).
18     A summary of earlier literature concluded that a 7-h, 3-month mean of 49 ppb O3 corresponding
19     to a SUM06 exposure of 26 ppm-h, would  cause 10%  loss in 50% of 49 experimental cases
20     (Tingey et al., 1991). Recent data summarized in Table 9-16 support this conclusion and more
21     generally  indicate that ambient O3 exposures can reduce the growth and yield of annual species.
22     Some annual species such as soybean [Glycine  max (L.) Merr.] are more sensitive, and greater
23     losses in such species may be expected (Table 9-16).  Thus, the recent scientific literature
24     supports the conclusions of the 1996 O3 AQCD that ambient O3 concentrations are reducing the
25     yield of major crops in the United States.
26          Much research in Europe has used the cutoff-concentration weighted,  cumulative-exposure
27     statistic AOT40; and substantial effort has gone into developing "Level-1" critical levels for
28     vegetation using this index.  Based on regression analysis of 15 OTC studies of spring wheat
29     including  one U.S. study and 14 studies from locations ranging from southern Sweden to
30     Switzerland, an AOT40 value of 5.7 ppm-h was found to correspond to a 10% yield loss, and a
31     value of 2.8 ppm-h corresponded to a 5% yield  loss (Fuhrer et al., 1997). Because a 4 to 5%
32     decrease could be detected with a confidence level of 99%, 3 ppm-h was selected by the
33     European Union as the AOT40 critical level in  1996 (Karenlampi and Skarby, 1996).
       August 2005                             9-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1           In addition to reductions in crop yield, O3 may also reduce the quality or nutritive value of
 2      annual species. Many studies have found O3 effects on various measures of plant organs that
 3      affect quality, with most of those studies focusing on characteristics important for food or
 4      fodder. These studies indicate that ambient O3 may have economically important effects on the
 5      quality of crop and forage species. Previous O3 AQCDs have concluded that visible symptoms
 6      on marketable portions of crops and ornamental plants can occur with seasonal 7-h mean O3
 7      exposures of 40 to 100 ppb (U.S. Environmental Protection Agency, 1978; 1986; 1992; 1996).
 8      The recent scientific literature does not refute this conclusion.
 9           The use of OTCs may reverse the usual vertical gradient in O3 that occurs within a few
10      meters above the ground surface (see Annex Section AX9.1).  This reversal suggests that OTC
11      studies may,  to some degree, overestimate the effects of an O3 concentration as measured several
12      meters above the ground. However, such considerations do not invalidate the conclusion of the
13      1996 O3 AQCD that ambient O3 exposures are sufficient to reduce the yield of major crops in the
14      United States.
15           As found for single-season agricultural crops, yields of multiple-year forage crops are
16      reduced at ozone exposures that occur over large  areas of the United States.  This result is
17      similar to that reported in the 1996 O3 AQCD (U.S. Environmental Protection Agency, 1996).
18      When species are grown in mixtures, O3 exposure can lead to the increased growth of O3-tolerant
19      species while exacerbating the growth decrease of O3-sensitive species. Because of this
20      competitive interaction, the total growth of the mixed-species community may not be affected by
21      O3 exposure.  However, in some cases, mixtures of grasses and clover species have shown
22      significant decreases in total biomass growth in response to O3 exposure in studies in the United
23      States and in  Sweden. In Europe, a provisional AOT40 critical level of 7 ppm-h over 6 months
24      has been proposed by the European Union as a value to protect sensitive herbaceous perennial
25      plant species  from the adverse effects of O3.
26           For deciduous tree species, recent evidence from FACE and OTC studies supports results
27      observed in previous OTC studies. For example, a series of O3-FACE studies was undertaken in
28      Rhinelander,  WI (Isebrands et al., 2000, 2001). These studies showed that O3 symptom
29      expression was generally similar in OTCs, FACE, and also at sites along an ambient O3 gradient,
30      supporting the previously observed variation among aspen clones obtained using OTCs
        August 2005                              9-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (Karnosky et al., 1999). As has been observed in previous O3 AQCDs, root growth is often
 2      found to be the most sensitive biomass response to O3.
 3           Results since 1996 support the conclusion of the 1996 O3 AQCD (U.S. Environmental
 4      Protection Agency, 1996) that deciduous trees are generally less sensitive to O3 than are most
 5      annual plants, with the exception of a few very sensitive genera such as Populus and sensitive
 6      species such as black cherry (Prunus serotina Ehrh.). However, the data presented in
 7      Table AX9-18 (see Annex Section AX9.5) suggest that ambient exposures that occur in the
 8      United States  can sometimes reduce the growth of seedlings of deciduous species. Results from
 9      multi-year studies sometimes show a pattern of increased effects in subsequent years. In some
10      cases, however, growth decreases due to O3 may become less significant or even disappear over
11      time. While some mature trees show greater O3 sensitivity than do seedlings in physiological
12      parameters such as net photosynthetic rate, these effects may not translate into measurable
13      reductions in biomass growth. However, because even multi-year experiments do not expose
14      trees to O3 for more than a small fraction of their life span and because competition may, in
15      some cases, exacerbate the effects of O3 on individual species, determining O3 effects on mature
16      trees remains  a significant challenge.
17           In Europe, a Level I critical level has been set for forest trees based on OTC studies of
18      European beech (Fagus sylvatica L.) seedlings: defined as an AOT40 value of 10 ppm-h for
19      daylight hours for a 6-month growing season (Karenlampi and Skarby, 1996).  However, other
20      studies show that other species,  such as silver birch(Betulapendula Roth.), may be more
21      sensitive to O3 than beech (Paakkonen et al., 1996).
22           As found for other tree species, various evergreen tree species and genotypes have widely
23      varying O3 sensitivities. Based on OTC studies with seedlings, major evergreen species in the
24      United States  are generally less  sensitive than are most deciduous trees, and  slower-growing
25      evergreen  species are less sensitive than are faster-growing species. There is evidence that
26      interacting stress, such as competition, may increase the O3 sensitivity of trees. As in studies of
27      deciduous species, most experiments with evergreen species have only covered a very small
28      portion of the  life span of a tree  and have been conducted with seedlings, so  estimating effects
29      on mature evergreens is difficult.
30           For all types of perennial vegetation, cumulative effects over more than one growing
31      season may be important; and studies for only a single season may underestimate effects.

        August 2005                               9-16       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Mature trees may be more or less sensitive to O3 than are seedlings, depending on the species,
 2      but information on physiological traits can be used to predict such differences in specific cases.
 3      In some cases, mature trees may be more sensitive to O3 than seedlings due to differences in gas
 4      exchange rates, differences in growth rates, greater cumulative exposure, or the interaction of O3
 5      with other stressors.
 6
 7
 8      9.7  EFFECTS OF OZONE EXPOSURE  ON NATURAL ECOSYSTEMS
 9           There is evidence that tropospheric O3 is an important stressor of ecosystems, with
10      documented impacts on the biotic condition, ecological processes, and chemical/physical nature
11      of natural ecosystems  (See Table AX9-22; Annex Section AX9.6). In turn, the effects of O3 on
12      individual plants and processes are  scaled up through the ecosystem affecting processes such as
13      energy and material flow, inter- and intraspecies competition, and net primary productivity
14      (NPP).  Thus, effects on individual  keystone species and their associated microflora and fauna,
15      which have been shown experimentally, may cascade through the ecosystem  to the landscape
16      level, although this has not yet been demonstrated. By affecting water balance, cold hardiness,
17      tolerance to wind and  by predisposing plants to  insect and disease pests, O3 may even impact the
18      occurrence and impact of natural disturbance (e.g., fire, erosion). Despite the probable
19      occurrence of such effects, there are essentially  no instances where ecosystem level, highly
20      integrated studies have conclusively shown that ozone is indeed altering ecosystem structure
21      and/or function.
22           Systematic injury surveys demonstrate that foliar injury occurs to O3 sensitive species in
23      many regions of the United States (Smith et al.,  2003; Coulston et al., 2003; Chappelka et al.,
24      1997) (Campbell et al., 2000) and Europe (Braun et al., 1999). However, the frequent lack of
25      correspondence between foliar symptoms and growth effects means that other methods must be
26      used to estimate the regional effects of O3 on tree growth rates (e.g., Rebbeck, 1996; Kouterick
27      et al., 2000). Investigations of the radial  growth of mature trees in combination with data from
28      many controlled studies with seedlings and a few studies with mature trees suggest that ambient
29      O3 is reducing the growth of mature trees in some  locations (Somers et al., 1998).  Studies using
30      models based on tree physiology and forest stand dynamics suggest that modest effects of O3 on
31      growth may accumulate over time and may interact with other stressors (Laurence et al., 2001)

        August 2005                              9-17        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (Laurence et al., 2003).  For mixed-species stands, such models predict that overall stand growth
 2      rate is generally not likely to be affected.  However, competitive interactions among species may
 3      change as a result of growth reductions of O3-sensitive species (Weinstein et al., 2001).  These
 4      results suggest that O3 exposure over decades may be altering the species composition of forests
 5      in some regions.
 6           Despite increased understanding of possible ecosystem effects of ozone, the data base
 7      demonstrating and quantifying the degree to which O3 is altering natural ecosystems is sparse.
 8      Much of the speculation of ozone impact on ecosystems must be inferred from a number of case
 9      studies of forest plot field-based data reporting on  a number of different species. One means to
10      discuss our  current knowledge is by listing the areas in which more information is needed.
11      These include:
12           Ecosystem processes. Very little is known about the effects of O3 on water, carbon, and
13      nutrient cycling, particularly at the stand and community levels. Effects on below-ground
14      ecosystem processes in response to O3 exposure alone, and in combination with other stressors,
15      are critical to projections at the watershed and landscape levels. Little is yet known about the
16      effects of O3 on  structural or functional components of soil food webs or how these impacts
17      could affect plant species diversity (Andersen, 2003).
18           Biodiversity and genetic diversity. The study of genetic aspects of O3 impacts on natural
19      ecosystems  has been largely based on correlations, and it remains to be shown conclusively
20      whether O3  affects biodiversity or genetic diversity (Pitelka, 1988; Winner et al., 1991; Davison
21      and Barnes, 1998).  Studies of competitive interactions under elevated O3 levels are needed
22      (Laurence and Andersen, 2003). Reexaminations via new sampling of population studies to
23      bring in a time component to previous studies showing spatial variability in population responses
24      to O3 are  also needed. These studies could be strengthened by modern molecular methodologies
25      to quantify impacts on diversity.
26           Natural ecosystem interactions with the atmosphere. Little is known about feedbacks
27      between O3  and  climate change on production of volatile organic compounds, which, in turn,
28      could affect O3 production (Fuentes et al., 2001).  At moderate to high O3 exposure sites,
29      aberrations  in stomatal behavior could significantly affect  individual tree water balance of O3-
30      sensitive trees; and if the sensitive tree species is dominant, the hydrologic balance at the
31      watershed and landscape levels could be affected.  This has not been addressed in any model,

        August 2005                              9-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1      because O3-exposure effects, if included at all in the modeling effort, have assumed a linear
 2      relationship between assimilation and stomatal conductance.  Interaction studies with other
 3      components of global change (i.e., warming, increasing atmospheric CO2, N deposition, etc.) or
 4      with various biotic stressors are needed to better predict complex interactions likely in the future
 5      (Laurence and Andersen, 2003). Whether O3 will negate the positive effects of an elevated CO2
 6      environment on plant carbon and water balances is not yet known, nor is it known if these effects
 7      will scale up through the ecosystem.  How O3 affects the progress of pest epidemics and insect
 8      outbreaks as concentrations increase is unclear (Ball et al., 1998). Information concerning the
 9      impact of O3 on plant pest and insect reproductive processes and reproductive development
10      under realistic field or forest conditions are needed as well as examination of reproductive
11      effects under interacting pollutants (Black et al., 2000).
12           Scaling. The vast majority of O3 studies of trees have been conducted with young,
13      immature trees and in trees that have not yet formed a closed canopy. Questions remain as to the
14      comparability of O3 effects on juvenile and mature trees and on trees grown in the open versus
15      those in a closed forest canopy in a competitive environment (Chappelka and Samuelson, 1998;
16      Kolb and Matyssek, 2001; Samuelson and Kelly, 2001). Merging the effects of O3 across spatial
17      scales is also difficult. Scaling responses of a single or a few plants to effects on communities
18      and ecosystems are complicated matters that will require a combination of manipulative
19      experiments with model ecosystems; community and ecosystem studies along natural O3
20      gradients; and extensive modeling efforts to project landscape level, regional, national and
21      international impacts of O3. Linking these various studies via impacts on common research
22      quantification across various scales using measures of such factors as leaf area index or spectral
23      reflective data, which could eventually be remotely sensed (Kraft et al.,  1996; Panek et al.,
24      2003), would provide powerful new tools for ecologists.
25           Identifying endpoints. In general, methodologies to determine the important values  of
26      services and benefits derived from natural ecosystems are lacking.  Identifying and quantifying
27      factors that could be used in comprehensive risk assessment for O3 effects on natural ecosystems
28      would increase societal awareness of the importance of protecting ecosystems (Heck et al.,
29      1998).
30
31

        August 2005                              9-19         DRAFT-DO NOT QUOTE OR CITE

-------
 1      9.8  ECONOMICS
 2           Substantial progress has been made over the past two decades in our understanding of the
 3      effects of ozone and other oxidants on vegetation, particularly for agriculturally important plant
 4      species.  See Annex Section AX9.7 for a more detailed discussion.  The physical and economic
 5      effects on agriculture are well documented and provide useful information for the consideration
 6      of establishing air quality standards for crops (e.g., Spash, 1997).
 7           Since the completion of the National Crop Loss Assessment Newwork (NCLAN) program
 8      in the late 1980s, the number of economic assessments of air pollution studies focusing on
 9      terrestrial ecosystems in general, and agriculture in particular, has declined.  For example, for the
10      period of 1980 to 1990, 33 economic studies of O3 and other air pollutant effects on U.S. crops
11      were published in peer-reviewed journal outlets (Spash, 1997). However, in preparing this
12      section of the current O3 AQCD, only four peer-reviewed economic assessments were found for
13      the decade of 1991 to 2000 that addressed vegetation in the United States. In addition, one
14      peer-reviewed article (Kuik et al., 2000) was found dealing with agriculture in the Netherlands.
15      Recent interest in global climate change and the potential effects of global warming on O3 and
16      other photochemical oxidants, has renewed interest in the effects of air pollution on both
17      managed and unmanaged terrestrial ecosystems (Adams et al., 1998).  In addition, concern is
18      growing for regarding the effects of air pollutants on natural ecosystems and on the services they
19      provide (Daily, 1997).  Unfortunately, this interest has not yet translated into additional
20      peer-reviewed publications addressing O3 or other air pollutants effects on ecosystems.
21           A study by Murphy et al. (1999) of the economic effects of tropospheric O3 on U.S.
22      agriculture is of note here, because it confirms the general magnitude of economic effects
23      reported by the two key studies performed a decade earlier (Adams, 1986; 1985). Specifically,
24      Murphy  et al. (1999) evaluated benefits to eight major crops associated with several scenarios
25      concerning the reduction or elimination of O3 precursor emissions from motor vehicles in the
26      United States. Their analysis reported a $2.8 to 5.8 billion (1990 dollars) benefit from complete
27      elimination of O3 exposures from all sources, i.e., ambient O3 reduced to a background level
28      assumed to be 0.025 to 0.027 ppm. While the analytical framework is similar to Adams et al.
29      (1986) in the use of NCLAN-based yield response functions and a mathematical
30      programming-based economic optimization model, the study is novel in its focus on the role of
31      motor vehicle emissions of VOCs/NOx in total anthropogenic O3 levels.  The study is also

        August 2005                              9-20        DRAFT-DO NOT QUOTE OR CITE

-------
 1      notable in its careful attention to federal farm program effects, particularly the deficiency
 2      payment component.
 3           There have been a number of recent studies of air pollutant effects on tree species in the
 4      literature.  Some have reported changes in total biomass and focused on European species
 5      (Kurczynska et al., 1997). Other studies have assessed changes in composition of forest species
 6      (biodiversity) or forest health due to exposure to air pollutants (Bringmark and Bringmark, 1995;
 7      McLaughlin and Percy,  1999; Vacek et al., 1999).  As noted previously, changes in forest
 8      biomass and composition are more difficult to value than marketable products. However,
 9      measures of forest composition  or health have implications for an area of increasing policy
10      concern, that being the effect of air pollutants and other environmental stressors on unmanaged
11      (natural) ecosystems and the services they provide (Goulder and Kennedy, 1997; Pimentel et al.,
12      1997).  Considerable discussion has  occurred among ecologists and economists as to the
13      appropriate means for valuing these  services (Anderson, 1990; Carpenter and Dixon, 1985;
14      Common and Perrings, 1992).  A number  of conceptual articles have been published on this
15      issue in both economic and ecological journals (Bergstrom, 1990; Castle, 1993; Pearce, 1993;
16      Suter, II, 1990).
17           Effects on forests and natural ecosystems remain problematic, due to limitations in
18      biological response data and economic methods. The problem is even more acute for valuing
19      natural  ecosystem service flows. The current limitations surrounding forests and natural
20      ecosystems present a rich research agenda. Areas of greatest potential value in terms of regional
21      policymaking need to be prioritized. Such priority setting can be assisted by sensitivity analyses
22      with existing economic models. By  measuring the changes in economic effects arising from
23      changes in key parameters, it is  possible to identify those research  data gaps most likely to affect
24      economic values.
25
26
        August 2005                              9-21        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES

  2       Adams, R. M. (1986) Agriculture, forestry, and related benefits of air pollution control: a review and some
  3             observations. Am. J. Agric. Econ. 68: 464-472.
  4       Adams, R. M.; Hamilton, S. A.; McCarl, B. A. (1985) An assessment of the economic effects of ozone on U.S.
  5             agriculture. J. AirPollut. Control Assoc. 35: 938-943.
  6       Adams, R. M.; Hamilton, S. A.; McCarl, B. A. (1986) The benefits of pollution control: the case of ozone and U.S.
  7             agriculture. Am. J. Agric. Econ. 68: 886-893.
  8       Adams, R. M.; Kurd, B. H.; Lenhart, S.; Leary, N. (1998) Effects of global climate change on agriculture: an
  9             interpretative review. Clim. Res. 11: 19-30.
10       Anderson, E. (1990) The ethical limitations of the market. Econ. Philos. 6: 179-205.
11       Andersen, C. P. (2003) Source-sink balance and carbon allocation below ground in plants exposed to ozone.
12             New Phytol. 157: 213-228.
13       Andersen, C. P.; Wilson, R.; Plocher, M.; Hogsett, W. E. (1997) Carry-over effects of ozone on root growth and
14             carbohydrate concentrations of ponderosa pine seedlings. Tree Physiol. 17: 805-811.
15       Andersen, C. P.; Hogsett, W. E.; Plocher, M.; Rodecap, K.; Lee, E. H. (2001) Blue wild-rye grass competition
16             increases the effect of ozone  on ponderosa pine seedlings. Tree Physiol. 21:319-327.
17       Arbaugh, M.; Bytnerowicz, A.; Grulke, N.; Fenn, M.; Poth, M.; Temple, P.; Miller, P. (2003) Photochemical smog
18             effects in mixed conifer forests along a natural gradient of ozone and nitrogen deposition in the San
19             Bernardino Mountains. Environ. Int. 29: 401-406.
20       Assmann, S. M. (2003) Open stomata 1 opens the door to ABA signaling inArabidopsis guard cells. Trends Plant
21             Sci. 8: 151-153.
22       Assmann, S. M.; Wang, X. Q. (2001) From milliseconds to millions of years: guard cells  and environmental
23             responses. Curr. Opin. Plant Biol. 4: 421-428.
24       Ball, G. R.; Benton, J.; Palmer-Brown, D.; Fuhrer, J.; Skarby, L.; Gimeno, B. S.; Mills, G. (1998) Identifying factors
25             which modify the effects of ambient ozone on white clover (Trifolium repens) in Europe. Environ. Pollut.
26             103:7-16.
27       Ball, G. R.; Palmer-Brown, D.; Fuhrer, J.; Skarby, L.; Gimeno, B. S.; Mills, G. (2000) Identification of non-linear
28             influences on the seasonal ozone dose-response of sensitive and resistant clover clones using artificial neural
29             networks. Ecol. Model. 129:  153-168.
30       Balls, G. R.; Palmer-Brown, D.; Sanders, G. E. (1996) Investigating microclimatic influences on ozone injury in
31             clover (Trifolium subterraneum) using artificial neural networks. New Phytol. 132: 271-280.
32       Barnes, J. D.; Wellburn, A. R. (1998) Air pollutant combinations. In: De Kok, L.  J.; Stolen, L, eds. Responses of
33             plant metabolism to air pollution and global change.  Leiden, The Netherlands: Backhuys Publishers;
34             pp. 147-164.
35       Batts, G. R.; Morison, J. I. L.; Ellis, R. H.; Hadley, P.; Wheeler, T. R. (1997) Effects of CO2 and temperature on
36             growth and yield of crops of winter wheat over four seasons. Eur. J. Agron. 7: 43-52.
37       Bergstrom, J. C. (1990) Concepts and measures of the economic value of environmental quality: a review.
38             J. Environ. Manage. 31:215-228.
39       Black, V. J.; Black, C. R.; Roberts, J. A.;  Stewart, C. A.  (2000) Impact of ozone on the reproductive development of
40             plants. New Phytol. 147: 421-447.
41       Bonello, P.; Heller, W.; Sandermann, H.,  Jr. (1993) Ozone effects on root-disease susceptibility and defence
42             responses in mycorrhizal and non-mycorrhizal seedlings of Scots pine (Pinus sylvestris L). New Phytol.
43             124:653-663.
44       Braun, S.; Rihm, B.; Schindler, C.; Fluckiger, W. (1999)  Growth of mature beech in relation to ozone and nitrogen
45             deposition: an epidemiological approach. Water Air  Soil Pollut.  116: 357-364.
46       Bringmark, E.; Bringmark, L.  (1995) Disappearance of spatial variability and structure in forest floors - a distinct
47             effect of air pollution? Water Air SoilPollut. 85: 761-766.
48       Broadmeadow, M. S. J.; Jackson, S. B. (2000) Growth responses of Quercuspetraea, Fraxinus excelsior and Pinus
49             sylvestris to elevated carbon dioxide, ozone and water supply. New Phytol. 146: 437-451.
50       Campbell, S.; Temple, P.; Pronos, J.; Rochefort, R.; Andersen, C. (2000) Monitoring for ozone injury in west coast
51             (Oregon, Washington, California) forests in 1998.  Portland, OR: U.S. Department  of Agriculture, Forest
52             Service, Pacific Northwest Research Station; general technical report no. PNW-GTR-495. Available:
53             http://www.fs.fed.us/pnw/gtrs.htm [11 April, 2003].
54       Cardoso-Vilhena, J.; Barnes, J. (2001) Does nitrogen supply affect the response of wheat (Triticum aestivum cv
55             Hanno) to the combination of elevated CO2 and O3? J. Exp. Bot. 52: 1901-1911.
         August 2005                                    9-22         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Carpenter, R. A.; Dixon, J. A. (1985) Ecology meets economics: a guide to sustainable development. Environment
  2             27:6-32.
  3       Castle, E. N. (1993) A pluralistic, pragmatic and evolutionary approach to natural resource management. For. Ecol.
  4             Manage. 56: 279-295.
  5       Chappelka, A. H.; Samuelson, L. J. (1998) Ambient ozone effects on forest trees of the eastern United States:
  6             a review. New Phytol.  139: 91-108.
  7       Chappelka, A.; Renfro, J.; Somers, G.; Nash, B. (1997) Evaluation of ozone injury on foliage of black cherry
  8             (Prunus serotina) and tall milkweed (Asclepias exaltata) in Great Smoky Mountains National Park. Environ.
  9             Pollut. 95: 13-18.
10       Colls, J. J.; Unsworth, M. H. (1992) Air pollution interactions with natural stressors. In: Barker, J. R.; Tingey, D. T.,
11             eds. Air pollution effects on biodiversity. New York, NY: Van Nostrand Reinhold; pp. 93-108.
12       Common, M.; Perrings, C. (1992) Towards an ecological economics of sustainability. Ecol. Econ. 6: 7-34.
13       Constable, J. V. H.; Taylor, G. E., Jr.; Laurence, J. A.; Weber, J. A. (1996) Climatic change effects on the
14             physiology and growth of Pinusponderosa: expectations from simulation modeling. Can. J. For. Res.
15             26: 1315-1325.
16       Coulston, J. W.; Smith, G. C.; Smith, W. D. (2003) Regional assessment of ozone sensitive tree species using
17             bioindicator plants. Environ. Monit. Assess. 83:  113-127.
18       Daily, G. C. (1997) Introduction: what are ecosystem services? In: Daily, G. C., ed. Nature's services: societal
19             dependence on natural ecosystems. Washington, DC: Island Press; pp. 1-10.
20       Danielsson, H.; Karlsson, G. P.; Karlsson, P. E.; Pleijel, J. (2003) Ozone uptake modelling and flux-response
21             relationships—an assessment of ozone-induced yield loss in spring wheat. Atmos.  Environ. 37: 475-485.
22       Davison, A. W.; Barnes, J. D. (1998) Effects of ozone on wild plants. New Phytol. 139: 135-151.
23       Docherty, M.; Salt, D. T.; Holopainen, J. K. (1997) The impacts of climate change and pollution on forest pests.
24             In: Watt, A. D.; Stork,  N. E.; Hunter, M. D., eds. Forests and insects. Chapman and Hall; pp. 229-247.
25       Dowding, P. (1988) Air pollutant effects on plant pathogens. In:  Schulte-Hostede, S.; Darrall, N. M.; Blank, L. W.;
26             Wellburn, A. R., eds. Air pollution and plant metabolism. London, United Kingdom: Elsevier Applied
27             Science Publishers; pp. 329-355.
28       Emberson, L.; Ashmore, M. R.; Cambridge, H. M.; Simpson, D.; Tuovinen, J. P. (2000) Modelling stomatal ozone
29             flux across Europe. Environ. Pollut. 109: 403-413.
30       Fangmeier, A.; Bender, J.; Weigel, H. J.; Jager, H. J. (2002) Effects of pollutant mixtures. In: Bell, J. N. B.;
31             Treshow, M., eds. Air Pollution and Plant Life. 2nd ed. Chichester, United Kingdom: John Wiley & Sons
32             Ltd.; pp. 251-272.
33       Flagler, R. B. (1998) Recognition of air pollution injury to vegetation: a pictorial atlas. 2nd ed.  Pittsburgh, PA: Air &
34             Waste  Management Association.
35       Fluckiger, W.; Braun, S.; Hiltbrunner, E. (2002) Effects of air pollutants onbiotic stress. In: Bell, J. N. B.; Treshow,
36             M., eds. Air pollution and plant life. 2nd ed. Chichester, United Kingdom: John Wiley & Sons Ltd.;
37             pp. 379-406.
38       Fredericksen, T. S.; Skelly, J. M.; Snyder, K. R.; Steiner, K. C. (1996) Predicting ozone uptake from meteorological
39             and environmental variables. J. Air Waste Manage. Assoc. 46: 464-469.
40       Fuentes, J. D.; Hayden, B. P.; Garstang, M.; Lerdau, M.; Fitzjarrald, D.; Baldocchi, D. D.; Monson, R.; Lamb, B.;
41             Geron, C. (2001) New directions: VOCs and biosphere-atmosphere feedbacks.  Atmos. Environ. 35:  189-191.
42       Fuhrer, J. H. (1994) Effects of ozone on managed pasture. 1. Effects of open-top chambers on microclimate, ozone
43             flux and plant growth. Environ. Pollut. 86: 297-305.
44       Fuhrer, J.; Skarby, L.; Ashmore, M. R. (1997) Critical levels for ozone effects on vegetation in Europe. Environ.
45             Pollution 97: 91-106.
46       Goulder, L. H.; Kennedy, D. (1997) Valuing ecosystem services: philosophical bases and empirical methods. In:
47             Daily,  G. C. Nature's Services: Societal Dependence on Natural Ecosystems. Washington,  DC: Island Press;
48             pp. 23-47.
49       Grulke, N. E.; Preisler, H. K.; Rose, C.; Kirsch, J.; Balduman, L. (2002) O3 uptake and drought stress effects on
50             carbon acquisition of ponderosa pine in natural stands. New Phytol. 154: 621-631.
51       Grulke, N. E.; Alonso, R.; Nguyen, T.; Cascio, C.; Dobrowolski, W. (2004) Stomata open at night in pole-sized and
52             mature ponderosa pine: implications for O3 exposure metrics. Tree Physiol. 24: 1001-1010.
53       Grunhage, L.; Jager, H. J. (2003) From critical levels to critical loads for ozone: a discussion of a new experimental
54             and modelling approach for establishing flux-response relationships for agricultural crops and native plant
55             species. Environ. Pollut. 125: 99-110.
         August 2005                                     9-23          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Heck, W. W.; Furiness, C. S.; Cowling, E. B.; Sims, C. K. (1998) Effects of ozone on crop, forest, and natural
  2             ecosystems: assessment of research needs. EM (October): 11-22.
  3       Hogsett, W. E.; Tingey, D. T.; Hendricks, C.; Rossi, D. (1989) Sensitivity of western conifers to SO2 and seasonal
  4             interaction of acid fog and ozone. In: Olson, R. K.; Lefohn, A. S., eds. Effects of air pollution on western
  5             forests [an A&WMA symposium; June; Anaheim, CA].  Air Pollution Control Association; pp. 469-491
  6             (APCA transactions series: no. 16).; pp. Anaheim,CA-Anaheim491.
  7       Innes, J. L.; Skelly, J. M; Schaub, M. (2001) Ozone and broadleaved species A guide to the identification of ozone-
  8             induced foliar injury.  Bern, Switzerland: Paul Haupt Publishers.
  9       Isebrands, J. G.; Dickson, R. E.; Rebbeck, J.; Karnosky, D. F. (2000) Interacting effects of multiple stresses on
10             growth and physiological processes in northern forest trees. In: Mickler, R. A.; Birsdey, R. A.; Horn, J., eds.
11             Responses of northern U.S. forests to environmental change. New York, NY: Springer-Verlag; pp. 149-180.
12             (Ecological studies: v. 139).
13       Isebrands, J. G.; McDonald, E. P.; Kruger, E.; Hendrey, G.;  Percy, K.; Pregitzer, K.;  Sober, J.; Karnosky, D. F.
14             (2001) Growth responses ofPopulus tremuloides clones to interacting carbon dioxide and tropospheric ozone.
15             Environ. Pollut. 115:  359-371.
16       Johnson, E.G.; Hale, B. A.; Ormrod, D. P. (1996) Carbon dioxide and ozone effects on growth of a legume-grass
17             mixture. J. Environ. Qual. 25: 908-916.
18       Karenlampi, L.; Ska'rby, L. (1996) Critical levels for ozone in Europe: testing and finalizing the concepts UN-ECE
19             workshop report. In: Proceedings of UN-ECE convention on long-range transboundary air pollution
20             workshop; April; Kuopio, Finland; Kupio, Finland: University of Kuopio, Department of Ecology and
21             Environmental Science.
22       Karnosky, D. F.; Mankovska, B.; Percy, K.; Dickson, R. E.; Podila, G. K.; Sober, J.; Noormets, A.; Hendrey, G.;
23             Coleman, M. D.; Kubiske, M.; Pregitzer, K. S.; Isebrands, J. G. (1999) Effects of tropospheric ozone on
24             trembling aspen and interaction with CO2: results from an O3-gradient and a FACE experiment. Water Air
25             Soil Pollut. 116:311-322.
26       Kolb, T. E.; Matyssek, R. (2001) Limitations and perspectives about scaling ozone impacts in trees. Environ. Pollut.
27             115:373-393.
28       Kopper, B. J.; Lindroth, R. L. (2003) Effects of elevated carbon dioxide and ozone on the phytochemistry of aspen
29             and performance of an herbivore. Oecologia 134: 95-103.
30       Kouterick, K. B.; Skelly, J. M.; Fredericksen, T. S.; Steiner, K. C.; Kolb,  T. E.; Ferdinand, J. A. (2000) Foliar injury,
31             leaf gas exchange and biomass responses of black cherry (Primus semtina Ehrh) half-sibling families to
32             ozone exposure. Environ. Pollut. 107:  117-126.
33       Kraft, M.; Weigel, H. J.; Mejer, G. J.; Brandes, F. (1996) Reflectance measurements of leaves for detecting visible
34             and non-visible ozone damage to crops. J. Plant Physiol. 148: 148-154.
35       Krupa, S. V.; Tonneijck, A.  E. G.; Manning, W. J. (1998) Ozone. In: Flagler, R. B., ed. Recognition of air pollution
36             injury to vegetation -  a pictorial atlas.  Air& Waste Management Association; pp.  2-11.
37       Krupa, S. V.; Nosal, M.; Peterson, D. L. (2001) Use of passive ozone O3 samplers in vegetation effects assessment.
38             Environ. Pollut. 112:  303-309.
39       Krupa, S. V.; Nosal, M.; Ferdinand, J. A.; Stevenson, R. E.;  Skelly, J. M.  (2003) A multi-variate statistical model
40             integrating passive sampler and meteorology data to predict the frequency distributions of hourly ambient
41             ozone (O3) concentrations. Environ. Pollut. 124:  173-178.
42       Kuik, O. J.; Helming, J. F. M.; Dorland, C.; Spaninks, F. A.  (2000) The economic benefits to agriculture of a
43             reduction of low-level ozone pollution in The Netherlands. Eur. Rev. Agric Econ. 27: 75-90.
44       Kurczyiiska, E. U.; Dmuchowski, W.; Wloch, W.; Bytnerowicz. A. (1997) The influence of air pollutants on needles
45             and stems of scots pine (Firms Sylvestris L.) trees. Environ. Pollut. 98: 325-334.
46       Kytoviita, M. M.; Le Thiec,  D.; Dizengremel, P. (2001) Elevated CO2 and ozone reduce nitrogen acquisition by
47             Pinus halepensis from its mycorrhizal symbiont. Physiol.  Plant. 111:305-312.
48       Landolt, W.; Gunthardt-Goerg, M.  S.; Pfenninger, L; Einig,  W.; Hampp, R.; Maurer, S.; Matyssek, R. (1997) Effect
49             of fertilization on ozone-induced changes in the metabolism of birch (Betula pendula) leaves. New Phytol.
50             137: 389-397.
         August 2005                                     9-24          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Langebartels, C.; Ernst, D.; Heller, W.; Lutz, C.; Payer, H. D.; Sandermann, H., Jr. (1997) Ozone responses of trees:
  2             results from controlled chamber exposures at the GSF phytotron. In: Sandermann, H.; Wellburn, A. R.;
  3             Heath, R. L., eds. Forest decline and ozone. New York, NY: Springer-Verlag; pp. 163-200 (Ecological
  4             studies: v. 127).
  5       Laurence, J. A.; Andersen, C. P. (2003) Ozone and natural systems: understanding exposure, response, and risk.
  6             Environ. Int. 29: 155-160.
  7       Laurence, J. A.; Amundson, R. G.; Kohut, R. J.; Weinstein, D. A. (1997) Growth and water use of red spruce
  8             (Picea rubens Sarg) exposed to ozone and simulated acidic precipitation for four growing seasons. For. Sci.
  9             (Bethesda, Md.) 43: 355-361.
10       Laurence, J. A.; Retzlaff, W. A.; Kern, J. S.; Lee, E. H.; Hogsett, W. E.; Weinstein, D.  A. (2001) Predicting the
11             regional impact of ozone and precipitation on the growth of loblolly pine and yellow poplar using linked
12             TREGRO and ZELG models. For. Ecol. Manage. 146: 247-263.
13       Laurence, J. A.; Retzlaff, W. A.; Kern, J. S.; Lee, E. H.; Hogsett, W. E.; Weinstein, D.  A. (2003) Corrigendum to
14             "Predicting the regional impact of ozone and precipitation on the growth of loblolly pine and yellow-poplar
15             using linked TREGRO and ZELIG models" [Forest Ecol Manage 146 (2001) 247-263]. For. Ecol. Manage.
16             174:607.
17       Lee, E. H.; Tingey, D. T.; Hogsett, W. E. (1988) Evaluation of ozone exposure indices  in exposure-response
18             modeling. Environ. Pollut. 53: 43-62.
19       Lee, E. H.; Tingey, D. T.; Hogsett, W. E.; Laurence, J. A. (2003) History of tropospheric ozone for the San
20             Bernardino Mountains of southern California, 1963-1999. Atmos. Environ. 37: 2705-2717.
21       Lefohn, A. S.; Shadwick, D. S. (2000) Differences in trending estimates in the United States using several ozone
22             metrics. In: Proceedings of the 93rd annual meeting of the Air & Waste Management Association; June;
23             Salt Lake City, UT; Pittsburgh, PA: Air & Waste Management Association; paper no. AS ld-645.
24       Maier-Maercker, U. (1998) Predisposition of trees to drought stress by  ozone. Tree Physiol. 19: 71-78.
25       Manning, W. J.; Krupa, S. V. (1992) Experimental methodology for studying the effects of ozone on crops and trees.
26             In: Lefohn, A. S., ed. Surface level ozone exposures and their effects on vegetation. Chelsea, MI: Lewis
27             Publishers, Inc.; pp. 93-156.
28       Massman, W. J. (2004) Toward an ozone standard to protect vegetation based on effective dose: a review of
29             deposition resistance and a possible metric. Atmos. Environ. 38: 2323-2337.
30       Matyssek, R.; Wieser, G.; Nunn, A. J.; Kozovits,  A. R.; Reiter, I. M.; Heerdt, C.; Winkler, J. B.; Baumgarten, M.;
31             Haberle, K. H.; Grams, T. E. E.; Werner, H.; Fabian, P.; Havranek, W. M. (2004) Comparison between
32             AOT40 and ozone uptake in forest trees of different species, age and site conditions. Atmos. Environ.
33             38:2271-2281.
34       McLaughlin, S.; Percy, K. (1999) Forest health in North America: some perspectives on actual and potential roles of
3 5             climate and air pollution. Water Air Soil Pollut. 116: 151-197.
36       Miller, P. R.; Rechel, J. (1999) Temporal changes in crown condition indices, needle litterfall, and collateral needle
37             injuries of Ponderosa and Jeffrey pines. In: Miller, P. R.; McBride, J. R., eds. Oxidant air pollution impacts in
3 8             the Montane forests of southern California: a case study of the San Bernardino Mountains. New York, NY:
39             Springer; pp. 164-178.
40       Mills, G. (2002) Modification of plant response by environmental conditions. In: Bell, J. N. B.; Treshow, M., eds.
41             Air pollution and plant life. 2nd ed. Chichester, United Kingdom: John Wiley & Sons Ltd.; pp. 343-358.
42       Mills, G.; Ball, G.; Hayes, F.; Fuhrer, J.; Skarby, L.; Gimeno, B.; De Temmerman, L.; Heagle, A.; Members of the
43             ICP Vegetation programme. (2000) Development of a multi-factor model for predicting the effects of ambient
44             ozone on the biomass of white clover. Environ. Pollut. 109: 533-542.
45       Momen, B.; Anderson, P. D.; Helms, J. A.; Houpis, J. L. J. (1997) Acid rain and ozone effects on gas exchange of
46             Pinusponderosa: a comparison between trees and seedlings. Int. J. Plant Sci. 158: 617-621.
47       Momen, B.; Anderson, P. D.; Helms, J. A. (1999) Temperature dependency of acid-rain effect on photosynthesis of
48             Pinus ponderosa. For. Ecol. Manage. 113: 223-230.
49       Murphy, J. J.; Deluki, M. A.; McCubbin, D. R.; Kim, H. J. (1999) The  cost of crop damage caused by ozone air
50             pollution from mo tor vehicles. J. Environ. Manage.  55: 273-289.
51       Musselman, R. C.; Massman, W. J. (1999) Ozone flux to vegetation and its relationship to plant response and
52             ambient air quality standards. Atmos. Environ.  33: 65-73.
53       Musselman, R. C.; Minnick, T. J. (2000) Nocturnal stomatal conductance and ambient air quality standards for
54             ozone. Atmos. Environ. 34: 719-733.
55       Nussbaum, S.; Fuhrer, J. (2000) Difference in ozone uptake in grassland species between open-top chambers and
56             ambient air. Environ. Pollut.  109: 463-471.
         August 2005                                    9-25         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Nussbaum, S.; Geissmann, M.; Fuhrer, J. (1995) Ozone exposure-response relationships for mixtures of perennial
  2             ryegrass and white clover depend on ozone exposure patterns. Atmos. Environ. 29: 989-995.
  3       Nussbaum, S.; Bungener, P.; Geissmann, M.; Fuhrer, J. (2000) Plant-plant interactions and soil moisture might be
  4             important in determining ozone impacts on grasslands. New Phytol. 147: 327-335.
  5       Oksanen, E.; Holopainen, T. (2001) Responses of two birch (Betula pendula Roth) clones to different ozone profiles
  6             with similar AOT40 exposure. Atmos. Environ. 35: 5245-5254.
  7       Ollerenshaw, J. H.; Lyons, T. (1999) Impacts of ozone on the growth and yield of field-grown winter wheat.
  8             Environ. Pollut. 106: 67-72.
  9       Ollerenshaw, J. H.; Lyons, T.; Barnes, J. D. (1999) Impacts of ozone on the growth and yield of field-grown winter
10             oilseed rape. Environ. Pollut. 104: 53-59.
11       Paakkonen, E.; Holopainen, T.; Karenlampi, L. (1996) Relationships between open-field ozone exposures and
12             growth and senescence of birch (Betula pendula and Betula pubescens). In: Karenlampi, L.; Skarby, L. eds.
13             Critical levels for ozone in Europe: testing and finalizing the concepts: UN-ECE workshop report. UN-ECE
14             Convention on Long-Range Transboundary Air Pollution; April; Kuopio, Finland; Kuopio, Finland:
15             University of Kuopio, Department of Ecology and Environmental Science; pp. 298-302.
16       Panek, J. A. (2004) Ozone uptake, water loss and carbon exchange dynamics in annually drought-stressed Pinus
17             ponderosa forests: measured trends and parameters for uptake modeling. Tree Physiol. 24: 277-290.
18       Panek, J.; Kurpius, M. R.; Goldstein, A. H. (2002) An evaluation of ozone exposure metrics for a seasonally
19             drought-stressed ponderosa pine ecosystem. Environ. Pollut. 117: 93-100.
20       Panek, J. A.; Baldocchi, D. D.; Goldstein, A.  H. (2003) The need for spatially and functionally integrated models of
21             ozone deposition to Sierra Nevada forests. In: Bytnerowicz, A.; Arbaugh, M. J.; Alonso, R., eds. Ozone air
22             pollution in the Sierra Nevada: distribution and effects on forests. New York, NY: Elsevier Science Ltd;
23             pp. 325-357. (Developments in environmental science: v.  2).
24       Pearce, D. W. (1993) Economic values and the natural world. London, United Kingdom: Earthscan Publications,
25             Ltd.
26       Pell, E. J.; Schlagnhaufer, C. D.; Arteca, R. N. (1997) Ozone-induced oxidative stress: mechanisms of action and
27             reaction. Physiol. Plant. 100: 264-273.
28       Percy, K. E.; Awmack, C. S.; Lindroth, R. L.; Kubiske, M. E.; Kopper, B. J.; Isebrands, J. G.; Pregitzer, K.  S.;
29             Hendry, G. R.; Dickson, R. E.; Zak, D. R.; Oksanen, E.; Sober,  J.; Harrington, R.; Karnosky, D. F. (2002)
3 0             Altered performance of forest pests under atmospheres enriched with CO2 and O3. Nature (London)
31             420:403-407.
32       Pimentel, D.; Wilson, C.; McCullum, C.; Huang, R.; Dwen, P.; Flack,  J.; Tran, Q.; Saltman, T.; Cliff, B. (1997)
33             Economic and environmental benefits  of biodiversity. BioScience 47: 747-757'.
34       Pitelka, L. F. (1988) Evolutionary responses of plants to anthropogenic pollutants. Trends Ecol. Evol. 3: 233-236.
35       Pleijel, H.; Danielsson, H.; Karlsson, G. P.; Gelang, J.; Karlsson, P. E.; Sellden, G. (2000) An ozone flux-response
36             relationship for wheat. Environ. Pollut. 109: 453-462.
37       Rao, M. V.; Davis, K. R. (2001) The physiology of ozone induced cell death. Planta 213: 682-690.
3 8       Rebbeck, J. (1996) Chronic ozone effects on three northeastern hardwood species: growth and biomass. Can. J. For.
39             Res. 26: 1788-1798.
40       Rowland-Bamford, A. J. (2000) Plant responses to changing carbon dioxide and temperature. In: Singh, S. N., ed.
41             Trace gas emissions and plants. Dordrecht,  The Netherlands: Kluwer Academic Publishers; pp. 63-74.
42       Samuelson, L.; Kelly, J. M. (2001) Scaling ozone effects from seedlings to forest trees. Tansley review no 21. New
43             Phytol. 149: 21-41.
44       Sandermann, H., Jr. (1998) Ozone: an air pollutant acting as a plant-signaling molecule. Naturwissenschaften
45             85: 369-375.
46       Sandermann, H., Jr. (2000) Ozone/biotic disease interactions: molecular biomarkers as a new experimental tool.
47             Environ. Pollut. 108: 327-332.
48       Sayre, R. G.; Fahey, T. J. (1999) Effects of rainfall acidity and ozone on foliar leaching in red spruce Picea  rubens.
49             Can. J. For. Res. 29: 486-496.
50       Schmieden, U.; Wild, A. (1995) The contribution of ozone to forest decline. Physiol. Plant. 94: 371-378.
51       Schnitzler, J. P.; Langebartels, C.; Heller, W.; Liu, J.; Lippert, M.; Dohring, T.; Bahnweg, G.;  Sandermann, H.
52             (1999) Ameliorating effect of UV-B radiation on the response of Norway spruce and Scots pine to ambient
53             ozone concentrations. Global Change Biol.  5: 83-94.
54       Schraudner, M.; Langebartels, C.; Sandermann, H., Jr. (1996) Plant defence systems and ozone. Biochem. Soc.
55             Trans. 24: 456-461.
         August 2005                                    9-26          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Smith, G.; Coulston, J.; Jepsen, E.; Prichard, T. (2003) A national ozone biomonitoring program - results from field
  2             surveys of ozone sensitive plants in northeastern forests (1994-2000). Environ. Monit. Assess. 87: 271-291.
  3       Somers, G. L.; Chappelka, A. H.; Rosseau, P.; Renfro, J. R. (1998) Empirical evidence of growth decline related to
  4             visible ozone injury. For. Ecol. Manage. 104: 129-137.
  5       Spash, C. L.  (1997) Assessing the economic benefits to agriculture from air pollution control. J. Econ. Surv.
  6             11:47-70.
  7       Suter, G. W., II. (1990) Endpoints for regional ecological risk assessments. Environ. Manage. (N. Y.) 14: 9-23.
  8       Tingey, D. T.; Hogsett, W. E.; Lee, E. H.; Herstrom, A. A.; Azevedo, S. H. (1991) An evaluation of various
  9             alternative ambient ozone standards based on crop yield loss data. In: Berglund, R. L.; Lawson, D. R.;
10             McKee, D. J. eds. Tropospheric ozone and the environment: papers from an international conference; March
11             1990; Los Angeles, CA; Pittsburgh, PA: Air & Waste Management Association (A&WMA transactions series
12             no. TR-19); pp. 272-288.
13       Tingey, D. T.; Hogsett, W. E.; Lee, E. H.; Laurence, J. A. (2004) Stricter ozone ambient air quality standard has
14             beneficial effect on ponderosa pine in California. Environ. Manage. 34: 397-405.
15       Topa, M. A.; Vanderklein, D. W.; Corbin, A. (2001) Effects of elevated ozone and low light on diurnal and seasonal
16             carbon gain in sugar maple. Plant Cell Environ.  24: 663-677.
17       U.S. Environmental Protection Agency (1978) Air quality criteria for ozone and other photochemical oxidants.
18             Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
19             Assessment Office; report no. EPA-600/8-78-004. Available from: NTIS, Springfield, VA; PB80-124753.
20       U.S. Environmental Protection Agency (1986) Air quality criteria for ozone and other photochemical oxidants.
21             Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
22             Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA; PB87-
23             142949.
24       U.S. Environmental Protection Agency (1992) Summary of selected new information on effects of ozone on health
25             and vegetation: supplement to 1986 air quality criteria for ozone and other photochemical oxidants. Research
26             Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
27             Office; report no. EPA/600/8-88/105F. Available from: NTIS, Springfield, VA; PB92-235670.
28       U.S. Environmental Protection Agency (1996) Air quality criteria for ozone and related photochemical oxidants.
29             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
30             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
31             http://cfpub2.epa.gov/ncea/.
32       Vacek, S.; Bastl, M.; Leps, J. (1999) Vegetation changes in forests of the Krkonose Mountains over a period of air
33             pollution stress (1980-1995). Plant Ecol. 143: 1-11.
34       Van Oijen, M.; Ewart, F. (1999) The effects of climatic variation in Europe on the yield response of spring wheat cv
3 5             Minaret to elevated CO2 and O3: an analysis  of open-top chamber experiments by means of two crop growth
36             simulation models. Eur. J. Agron.  10: 249-264.
37       Ward, E. R.; Uknes, S. J.; Williams,  S. C.; Dincher, S.  S.; Wiederhold, D. L.; Alexander, D. C.; Ahi-Goy, P.;
38             Metraux, J. P.; Ryals, J. A. (1991) Coordinate gene activity in response to agents that induce systemic
39             acquired resistance. Plant Cell 3: 1084-1094.
40       Weinstein, D. A.; Gollands, B.; Retzlaff,  W. A. (2001) The effects of ozone on a lower slope forest of the Great
41             Smoky Mountain National Park: simulations linking an individual tree model to a stand model. For. Sci.
42             47:29-42.
43       Whitfield, C. P.; Davison, A. W.; Ashenden, T. W.  (1998) The effects of nutrient limitation on the response of
44             Plantago major to ozone. New Phytol. 140: 219-230.
45       Winner, W. E.; Coleman, J. S.; Gillespie, C.; Mooney, H. A.; Pell, E. J. (1991) Consequences of evolving resistance
46             to air pollutants. In: Taylor, G. E.; Pitelka, L. F.; Clegg, M. T., eds. Ecological genetics and air pollution.
47             Springer-Verlag; pp. 177-202.
48       Wu, Y. X.; Tiedemann, A. V. (2002) Impact of fungicides on active oxygen species and antioxidant enzymes in
49             spring barley (Hordeum vulgare L) exposed  to ozone. Environ. Pollut. 116: 37-47.
50       Yun, S. C.; Laurence, J. A. (1999) The response of sensitive and tolerant clones ofPopulus tremuloides to dynamic
51             ozone exposure under controlled environmental conditions. New Phytol. 143: 305-313.

52
         August 2005                                    9-27         DRAFT-DO NOT QUOTE OR CITE

-------
 i               10.  TROPOSPHERIC OZONE EFFECTS ON
 2        UV-B FLUX, AND ITS ROLE IN CLIMATE CHANGE
 3
 4
 5     10.1   INTRODUCTION
 6          In addition to exerting direct effects on human health and vegetation/ecosystems, as
 7     discussed in earlier chapters, tropospheric ozone (O3) influences the ground-level flux of solar
 8     ultraviolet (UV) radiation, as well as other processes that alter the Earth's radiative balance and
 9     contribute to climate change. This chapter discusses tropospheric O3 and (1) its role in in
10     determining surface-level UV flux and, (2) its involvement in global climate change.
11
12
13     10.2   THE ROLE OF TROPOSPHERIC OZONE IN DETERMINING
14            GROUND-LEVEL UV-B FLUX
15          Atmospheric O3 plays  a crucial role in reducing the exposure of living organisms to solar
16     UV radiation. Approximately 90% of the total atmospheric O3 burden is located in the
17     stratosphere; therefore, photochemical processes that alter the concentration of stratospheric O3
18     are of particular concern within the global community.  The importance of stratospheric O3
19     depletion due to the release of long-lived anthropogenic chlorinated- and fluorinated
20     hydrocarbons was recognized over a period of several years during the 1970s and early 1980s
21     and led to the international treaty for the protection of stratospheric O3: the 1987 Montreal
22     Protocol on Substances that  Deplete the Ozone Layer.
23          While roughly representing only 10% of the total  atmospheric  O3 burden, tropospheric O3,
24     like other tropospheric pollutants, can influence the flux of biologically-damaging UV radiation
25     at the Earth's surface.  This  section summarizes the available information on the factors
26     governing UV flux at the Earth's surface, first, followed by  a discussion the status of scientific
27     understanding of the factors governing human UV exposure and, then, of the links between
28     UV exposure and human disease.
       August 2005                             10-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1      10.2.1  Factors Governing Ultraviolet Radiation Flux at the Earth's Surface
 2           The Montreal Protocol requires routine review of the latest scientific information available
 3      on the status of the O3 layer and of UV radiation levels at the Earth's surface. The World
 4      Meteorology Organization (WMO) and U.N. Environmental Program (UNEP) are responsible
 5      for assessing the state of the science regarding the O3 layer and for reporting on trends in surface
 6      UV radiation levels.  The latest WMO/UNEP assessment was published in 2002 (WMO/UNEP,
 7      2002).
 8           An outcome of the on-going atmospheric chemistry research effort devoted to tracking
 9      stratospheric O3 depletion and its effects is a body of literature, though limited, that describes the
10      effects of tropospheric pollutants, particulate matter (PM), and O3, on ground-level UV radiation
11      flux.  Drawing from the WMO/UNEP assessment and more recent literature, this section
12      describes the current level of scientific understanding of the factors influencing ground-level UV
13      radiation flux such as geophysical factors, tropospheric O3, PM, and cloud cover. Figure 10-1
14      visually summarizes  some of the factors that influence the flux of UV-B at the Earth's surface.
15
16      10.2.1.1  UV Radiation:: Wavelengths, Energies and Depth of Atmospheric Penetration
17           Designations for portions of the electromagnetic spectrum have evolved over time and are
18      usually associated with general functions or effects caused by photons within a given wavelength
19      range. The energy possessed by a photon is inversely proportional to its wavelength. For
20      example, gamma rays, having wavelengths <0.1 nm, are especially damaging high-energy
21      photons emitted during radioactive decay and by stellar activity.  Radiowaves, having
22      wavelengths > 108 nm, are very low in energy and function as carriers for broadcast
23      communications.
24           The wavelengths ranging between 50 and 400 nm in length are denoted "ultraviolet."
25      Solar radiation of wavelengths <280 nm, including UV-C  (200 to 280 nm), is almost entirely
26      blocked by the Earth's upper atmosphere due to photoionization and photodissociation
27      processes.  Figure 10-2 compares the solar flux above the atmosphere with ground-level flux.
28      Solar UV-B radiation (280 - 320 nm) is absorbed or scattered in part within the  atmosphere,
29      while UV-A radiation (320 - 400 nm) can be scattered but not absorbed to any meaningful
30      degree by atmospheric gases.  Both UV-B and UV-A photons contain sufficient energy to break
31      (photolyze) chemical bonds and are associated with human health- and ecosystem-damaging

        August 2005                              10-2       DRAFT-DO NOT QUOTE OR CITE

-------
                 ^^^1   Backscattered ^^^1
                   ^B    Radiation   ^B
                                       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.
 1     effects. However, because UV-B is more energetic, it is potentially capable of producing
 2     substantially more biological damage than UV-A.
 3
 4     10.2.1.2 Temporal Variations in Solar Flux
 5           The magnitude of the solar radiation flux entering the atmosphere depends upon long-term
 6     solar activity, sunspot cycle (11 years), solar rotation (27 days) and the position of the Earth in
 7     its orbit around the sun. A variety of changes in solar irradiance can be found in historical data,
 8     from 1700 to the present. Solanki and Fligge (2000) concluded that solar irradiation changes on
 9     time-scales of days to  centuries can be attributed to variations in solar magnetic features. Since
10     the last Maunder minimum in 1700, solar irradiance has increased slightly, at -3.0% for
11     wavelengths in the UV-C range and at -1.3% for wavelengths in the UV-B and UV-A ranges.
       August 2005
10-2
DRAFT-DO NOT QUOTE OR CITE

-------
                  0)
                  o
                     1.0000
                  I  0.1000
                  es
                  E
                  x  0.0100
                     0.0010
                     0.0001
                                              Solar Flux
                         290
                                      300           310           320
                                              Wavelength (nm)
                                                                                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 SOdeg,
                    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: Krotkovetal. (1998).
1     Including visible wavelengths, Solanki and Fligge (2000) estimated that the overall increase in

2     solar irradiance was -0.3%. Rozema et al. (2001) pointed out that any increase in wavelengths

3     <300 nm (UV-C) would initiate additional O3 formation in the stratosphere. This suggests that

4     any increase in UV-B and/or UV-A solar flux would be offset by a more absorptive stratosphere.
5          Solar rotation and sunspot activity have the greatest effects on radiation flux originating in

6     the highest levels of the solar atmosphere. The amplitude  of the associated cyclical changes in
7     solar shortwave radiation flux follows an inverse relationship between photon wavelength and

8     the solar altitude at which it was emitted. The maximum level of radiation (solar-max) differs
      August 2005
10-4
DRAFT-DO NOT QUOTE OR CITE

-------
 1      from the minimum (solar-min) by as much as 10% for wavelengths near 160 nm. This peak-to-
 2      trough difference declines to around 1% for 300 nm (UV-B range) (Salby, 1996).
 3           The combined effects of the Earth's obliquity (the angle of the Earth's axis of rotation with
 4      respect to the plane of its orbit around the sun) and its precession (the rotation of the Earth's axis
 5      with respect to a perpendicular line through the plane of its solar orbit) yield variations of up to
 6      30% in total summertime solar flux, depending on latitude (Hartmann,  1994).
 7
 8      Zenith Angle: Latitude, Season,  and Time of Day
 9           The sun's relative elevation is measured with respect to the vertical and is known as its
10      "zenith angle."  This angle varies hourly, seasonally, and with latitude. Daily and seasonal
11      changes in solar zenith angle result in the largest changes  in the magnitude of solar radiation
12      flux, with higher zenith angles corresponding to lower solar flux. The largest natural fluxes
13      occur in the tropical regions, where solar noon occurs at a zenith angle at or near 0°.  Seasonal
14      variation in solar flux ranges from small changes at the equator to very large changes at high
15      latitudes.  Daily variations in solar flux, from sunrise to sunset, show added wavelength
16      dependence as a function of zenith angle, because transmission of some wavelengths are
17      sensitive to atmospheric pathlength due to scattering and absorption processes.  These processes
18      will be discussed further below.
19
20      10.2.1.3  Atmospheric Radiative Interactions with Solar Ultraviolet Radiation
21      Radiative Interactions in the Stratosphere
22           As noted, earlier, the stratosphere contains 90% or more of the total column density of O3,
23      the principle gas phase absorber  of UV-B. Ozone interacts with UV radiation by scattering the
24      photon, or absorbing and transforming  its energy. Upon absorbing a UV photon, O3 may
25      photodissociate, or become electronically and vibrationally excited.
26           Photoabsorption by O3 occurs with very high efficiency. After electronically excited O3
27      (O3*) is formed, it will either lose its excess electronic energy via a collision with another gas
28      molecule (M) or dissociate into ground-state oxygen,  O2, and an electronically excited oxygen
29      radical, O(JD) (See Reactions 1 and 2). Intermolecular collisions degrade the excess electronic
30      energy of the O3* molecule by transferring it to other molecules as vibrational, rotational, and/or
31      translational energies, that warm the atmosphere.  An O(JD) radical can react with H2O to form

        August 2005                               10-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1     two hydroxyl (OH) radicals (Reaction 3). See Chapter 2 for further discussion of odd oxygen
 2     and HOX photochemistry.
                             O3 + hv ->• O3* ->• O^D) + O2                               (10-1)
                                            ->O3*+M-»O3 + M*                     (10-2)
 3
                                     O(1D) + H2O -» 2OH                              (10-3)
 5
 6
 7     Either of these photochemical processes transforms the energy of the UV photon into heat, a
 8     form of energy that, in this context, lacks the potential for human health or ecosystems damage.
 9          The WMO/UNEP (2002) scientific assessment reported that global average total
10     column O3 had declined by 3% from pre-1980 levels, due to the presence of anthropogenic O3-
1 1     depleting substances in the atmosphere.  Ozone depletion has a strong latitude and seasonal
12     dependence.  The seasonality of total O3 changes differ between the Northern and Southern
13     Hemispheres. In the northern midlatitudes, total column O3 declined by -4% during the
14     winter/spring seasons and by approximately half that amount in the summer/fall of the
15     1997-2001 time period, relative to pre-1980 total column O3 levels. In southern midlatitudes,
16     total column  O3 declined -6% during all seasons.
17          The concentration of O3 in a vertical column extending from the Earth's surface is
1 8     expressed in Dob son Units (DU) corresponding to the column height in hundredths of a
19     millimeter  of O3 at standard temperature and pressure (273 K and 1  atmosphere) (Wayne, 2000).
20     One DU =  2.587 x 1015 molecules of O3/cm2. The total O3 column effectively prevents any
21     UV-C from reaching the surface and reduces the penetration of UV-B to the surface, but it does
22     little to attenuate the intensity of UV-A except at the shorter wavelengths close to the cutoff for
23     UV-B.  Cutchis (1974) calculated that with overhead sun, a 10% decrease in the O3 column
24     would lead to 20,  250, and 500% increases in flux at 305, 290, and 287 nm, respectively, values
25     that have been supported by ground observations in Toronto, ON (49° N; Kerr and McElroy,
26     1993). Rapid changes of this magnitude appear to happen naturally. As seen in data collected
        August 2005                              10-6       DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
by the Total Ozone Mapping Satellite (TOMS) (Figure 10-3), the total O3 column undergoes
wide natural variation on short timescales (Cockell, 2001).
                                      Ozone Column (1990-1992)
               600
          c
          o
          (/)
          £1
          O
          a
                / / / / / /


                                                    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).
1          Nacreous and polar stratospheric clouds, and aerosols, such as those injected into the
2      stratosphere by explosive volcanic eruptions, both absorb and scatter radiation.  Relative to the
3      troposphere, the stratosphere is low in atmospheric pressure. Stratospheric clouds and aerosols
4      are also more dispersed than those in the troposphere. Consequently, UV radiation can traverse
5      the stratosphere with a substantially lower probability of encountering a gas molecule or cloud or
6      an aerosol particle than it would in the troposphere. In the radiative transfer literature, the
      August 2005
                                         10-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1      stratosphere is described as a "single scattering" regime for UV radiation, and UV that has
 2      penetrated the stratosphere is referred to as "direct beam UV."  The troposphere, due to its high
 3      gas and particle concentrations is referred to as a "multiple scattering" regime.
 4
 5      Radiative Interactions in the Troposphere:  Solar Irradiance Versus Actinic Flux
 6           The troposphere contains < 10% of the  total column O3 but -78% of the total atmospheric
 7      mass (including clouds, and gas- and particle-phase radiation scatterers and absorbers), making it
 8      a "multiple scattering" regime for UV radiation.  These scattering processes increase the mean-
 9      free path a photon must travel before reaching the Earth's surface, transforming the direct beam
10      UV solar irradiance that has penetrated the stratosphere into diffuse, or actinic, UV irradiance
11      (Briihl and Crutzen, 1988).
12           Rayleigh and Mie scattering of solar radiation are sensitive to molecular and particle size.
13      In Rayleigh scattering, gas molecules that are smaller than the wavelength of the incident photon
14      isotropically deflect incoming photons.  Conversely, aerosol and cloud droplets scatter incoming
15      radiation with distinctive forward- and backscattering tendencies, i.e., Mie scattering.  Actinic
16      flux, especially at the Earth's surface, is directly proportional to surface albedo (Wendisch and
17      Mayer, 2003). Surface albedo is very strongly wavelength dependent. For example, fresh and
18      wet snow reflect 60 to 90% of incident violet light, while soil and grass surfaces reflect <5%
19      (Xenopoulos and Schindler, 2001).  In their in situ measurement and modeling study of the
20      vertical distribution of solar irradiance, Wendisch and Mayer (2003) found that surface albedos
21      must be measured in order to accurately simulate solar flux, due to the large variations in albedo
22      that may occur within a given surface type.  Snow cover, even many kilometers from
23      measurement sites is known to increase detected UV irradiances. Complicated interactions
24      result when radiation is scattered by snow (or other bright surfaces) and backscattered or
25      absorbed by atmospheric particles and clouds in the same vicinity (WMO/UNEP, 2002).
26
27      Variation in Solar Flux with Altitude
28           Solar flux increases with altitude above sea level, due to the decreased presence  of clouds
29      and declining concentrations of scattering and absorbing atmospheric pollutants. Rayleigh
30      scattering, also lessens with decreasing atmospheric pressure.  A number of measurements of
31      UV radiation have been taken at various altitudes and are reviewed by Xenopoulos and Schindler

        August 2005                               10-8        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (2001). Increases in flux as a function of altitude are given as percent irradiance enhancement
 2      per 1000 m relative to sea level. The effect can range from 9 to 24% /1000 m as function of the
 3      altitude at which the measurement was taken (Xenopoulos and Schindler, 2001). The effect
 4      corresponds to the relative pathlength traveled by the solar photon:  flux is strongest when the
 5      photon is not impeded by atmospheric scattering or absorbing agents. Similarly, this effect is
 6      seen as a function of solar zenith angle, i.e., flux is at its maximum when the atmospheric depth
 7      through which the photon must pass is at its  shallowest.
 8
 9      Clouds
10           In principle, clouds have the largest influence on surface-level UV irradiance, but their
11      effects are difficult to quantify.  The depth and composition of a cloud determine, in part, the
12      amount and wavelengths of radiation that it will scatter or absorb.  Geometry is an especially
13      important factor, as the reduction in irradiance may be small with scattered or broken clouds - or
14      may be enhanced by scattering between clouds, increasing surface flux (WMO/UNEP, 2002).
15      Quantifying the effect of clouds on surface UV flux, therefore, requires detailed information on
16      cloud composition, geometry, altitude, and the position of the sun relative to the cloud and the
17      underlying surface as a function of time.  Provided that all of this information is available, a
18      three-dimensional model is then required to calculate surface-level reductions or enhancements
19      in UV flux.
20
21      Particulate Matter
22           On a zonally averaged basis, PM does not contribute significantly to lower tropospheric
23      absorption of UV radiation. However, in urban areas or other areas subject to high smog levels
24      (areas of significant biomass combustion), PM may be the most important determinant of
25      ground-level erythemal UV flux, second only to cloud cover (U.S. Environmental Protection
26      Agency, 2004; WMO/UNEP, 2002). Model-to-measurement comparisons of ground-level flux
27      for Greece and Toronto, Canada, have shown 20 and 5-10% reductions, respectively (McKenzie
28      et al., 2003).  Increases over the past 20 to 30 years in combustion-associated PM and black
29      carbon may account for the inability to detect a surface trend in UV-B radiation caused by a
30      known decrease in stratospheric O3 over the Northern Hemisphere (Barnard et al., 2003).
31

        August 2005                               10-9       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Gases
 2           In the upper troposphere, the UV-absorbing gases O3 and, of lesser importance,
 3      formaldehyde (CH2O) and SO2 are vented or diffuse from the surface. Stratospheric intrusions
 4      force O3-rich air into the troposphere where it mixes, increasing regional background O3 levels
 5      (see Chapter 3). Tropospheric O3 data are typically expressed on a concentration basis, e.g.,
 6      parts per billion by volume (ppbv), where 1 ppbv tropospheric O3 = 0.65 DU (IPCC, 200la).
 7      Ozone concentrations decrease with increasing altitude from the surface up to, roughly, the mid-
 8      troposphere, then increase up into the stratosphere.  Figure 10-4 shows a series of O3 vertical
 9      profiles for 4 sites within the continental U.S., i.e., plots of O3 concentrations as  a function of
10      atmospheric pressure (correlating to altitude). The mean values of O3 in the free troposphere
11      reported in the literature range from -50 to -80 ppbv, with higher values occurring at the
12      tropopause. For example, a series of ozonesonde soundings over France from 1976 to 1995
13      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
14      7.5 km layer, although the data revealed no statistically significant  increasing trend over time
15      (Ancellet and Beekmann, 1997).
16           Photochemistry produces a diurnal rise and fall in O3 and PM concentrations in polluted
17      urban settings. Temperature inversions that often occur in these settings prevent the upward
18      mixing and dilution of ground-level O3, also trapping primary and secondary PM within the
19      boundary layer. A recent study of the concurrence of O3 and  PM is provided by
20      Koloutsou-Vakakis et al. (2001).  No measurement technique is currently available that can
21      distinguish between absorption of incident UV radiation by O3 versus absorption by PM.
22           Ultraviolet absorption by gases becomes significant under aerosol- and cloud-free
23      conditions. Figure 10-5 shows a calculation by Krotkov et al. (1998) of the sensitivity, as a
24      function of wavelength, of ground-level UV flux to a 1-DU decrease  in total column O3 under
25      cloud- and aerosol-free conditions. A 1991 to 1992 study in Chicago in which ambient O3,
26      broadband UV irradiance, and total sunlight were monitored (Frederick et al., 1993) found a
27      significant negative correlation between the UV irradiance and ambient O3 when the atmosphere
28      was relatively free of clouds and haze. Although Frederick et al. (1993) estimated that a
29      10-ppbv reduction in O3 was associated with a 1.3% increase in erythemally-weighted UV-B,
30      they cautioned that this value had a comparatively large uncertainty (±1.2%, or nearly 100% of
31      the predicted increase).

        August 2005                               10-10       DRAFT-DO NOT QUOTE OR CITE

-------
                                    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).
August 2005
10-11       DRAFT-DO NOT QUOTE OR CITE

-------
                          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: Krotkov et al. (1998).
 1          Note that when attempting to apply the results of studies such as Krotkov et al. (1998) in
 2     an analysis of the importance of surface O3 to protecting humans from UVB-related diseases, is
 3     that flux is ordinarily determined as a function of total column O3 density, which includes the
 4     stratospheric and upper tropospheric O3, not simply surface-level pollutant O3.  Nearly all of the
 5     routine data on tropospheric O3 concentrations in the U.S. is from ground-level O3 monitors,
 6     such as those used to determine the attainment of the O3 air quality standards. Such
 7     measurements, alone, are not sufficient information for  making reliable estimates of ambient O3
 8     concentrations above the boundary layer.
 9
10     10.2.1.4   Data Requirements for a Surface UV-B Climatology
11          A means of establishing the range of variability in UV-B at ground level would be in the
12     development of a map of flux levels under typical seasonal conditions based on historical
       August 2005
10-12
DRAFT-DO NOT QUOTE OR CITE

-------
 1      records. In the atmospheric sciences community, a map of this type is referred to as a
 2      "climatology."
 3           The WMO/UNEP (2002) stated that, in principle, if the spatial distribution of all UV
 4      absorbers and scatterers were fully known, the wavelength and angular distribution of the UV
 5      irradiance at the Earth's surface could be determined with model calculations.  However, the
 6      very limited information available on the distribution of the primary components (i.e., clouds,
 7      particles, O3, and surface albedo) makes detailed predictions impossible.  In an earlier
 8      assessment of the environmental effects of stratospheric O3 depletion (WMO/UNEP, 1999), the
 9      UNEP concluded that, in view of the high spatial and temporal variability of surface UV
10      radiation and the difficulty in maintaining calibration within networks of UV monitoring
11      instruments, satellite-based observations are necessary to develop a satisfactory UV climatology.
12      Furthermore, satellite-derived estimates of surface UV are limited by the availability of
13      instruments in orbit, with available datasets comprising interpolations based upon a single
14      satellite overpass per day for a given region. Complete assessment of the uncertainties in
15      predictions of UV surface flux would require comparisons between sparse available ground-level
16      observations and satellite data over longer periods of time and for different geographical
17      locations. No such assessment has been reported in the scientific literature.
18
19      10.2.2   Factors Governing Human Exposure to Ultraviolet Radiation
20           An assessment of public health benefits due to the attenuation of UV-B radiation by
21      surface-level O3 requires appropriate consideration of: (1) the multiple factors that alter the flux
22      of UV-B radiation at ground-level, as described above; (2) the factors that influence the extent of
23      human exposure to UV-B radiation, particularly behavioral decisions; and (3) the effects of UV-
24      B radiation exposure on human health. Consideration must also be given to the public health
25      benefits from exposure to UV-B radiation.  The present section outlines the most recent
26      information on the determinants of exposure to UV-B radiation in human populations.
27      Quantitative evaluation of human exposure to UV-B radiation is scientifically necessary to
28      perform health risk assessment and to define subpopulations at risk for UV-B-related health
29      effects.
30
        August 2005                              10-13       DRAFT-DO NOT QUOTE OR CITE

-------
 1      10.2.2.1  Outdoor Activities
 2           Exposure to solar UV radiation is related to one predominating factor:  time spent outdoors
 3      during daylight hours. A large U.S.  study was conducted using the EPA National Human
 4      Activity Pattern Survey (NHAPS) to assess UV radiation dose in Americans (Godar, 2001;
 5      Godar et al., 2001, 2003).  The EPA NHAPS recorded the activity profiles of 9,386 Americans
 6      (age 0 to 60+ years) over a 24-month period to assess their exposure to various environmental
 7      pollutants, including UV radiation. Available UV radiation was assessed using the EPA UV-
 8      monitoring program. Solar radiation in the UV-A and UV-B waveband regions were measured
 9      daily at a monitoring site in each quadrant of the U.S.  There is considerable error associated
10      with quantifying UV radiation dose from exposure surveys and four UV-monitoring sites across
11      the country; however, the qualitative information regarding factors that increase human exposure
12      to UV radiation is still of relevance.  The EPA-UV monitoring network has since expanded to 21
13      sites, located in  14 U.S. national parks and 7 urban areas across the U.S.
14      (http://www.epa.gov/uvnet/).  A UV-B monitoring network by the U.S. Department of
15      Agriculture is also available for the quantitative assessment of UV radiation exposure
16      (http://uvb.nrel.colostate.edu/UVB/). This monitoring network has 30+ monitoring sites across
17      the U.S. and three additional sites in Canada and New Zealand.
18           Godar et al. (2001) observed a strong seasonal preference for outdoor activities, with
19      people spending the most time outdoors during the summer followed by spring, fall, and, lastly,
20      winter. Because the solar erythemal (i.e., skin reddening) UV radiation dose is also highest
21      during the summer, the estimated UV radiation dose of Americans was more than 10-fold greater
22      in the summer compared to the winter season (Godar et al., 2001).
23           Vacationing at the beach in the summer was associated with higher UV radiation exposures
24      (Godar et al., 2001; Thieden et al., 2001). Even after accounting for sunscreen use at the beach,
25      the erythemal UV radiation doses were more than 40% higher during a 3-week beach vacation
26      compared to a 3-week stay at home (Godar et al., 2001). Danish children and adolescents were
27      found to receive >50% of their annual UV radiation dose while vacationing at European beaches
28      (Thieden et al., 2004a).  Sunbathing also was associated with increased annual UV radiation
29      dose in the Canadian National Survey on Sun Exposure and Protective Behaviours (Shoveller
30      et al., 1998).  Among the 3,449 adults (age 25+ years) who completed the telephone household
31      survey, 21% stated that they spent time actively sunbathing.  In a Danish study with 164

        August 2005                             10-14       DRAFT-DO NOT QUOTE OR CITE

-------
 1      participants, all children (age 1 to 12 years) and teenagers (age 13 to 19 years) as well as 94% of
 2      adults (age 20 to 76 years) had days with risk behavior (Thieden et al., 2004b). Teenagers, who
 3      had the highest number of risk-behaviors days, were found to have the highest annual UV
 4      radiation doses.  Among teenagers, 76% (95% CI:  41, 98) of their UV radiation dose during the
 5      measurement period was received on risk-behavior days, as determined using personal electronic
 6      UV dosimeters and exposure diaries (Thieden et al., 2004b).
 7           An Australian study examining time profiles of daily UV radiation exposure among 8th
 8      grade students observed that up to 47% of the daily UV radiation dose fell within the time
 9      periods when students were outdoors during school hours, sitting under shaded structures during
10      lunch breaks and participating in routine outdoors or sports activities (Moise et al., 1999). Other
11      studies also have found that participation in outdoor sports (e.g., basketball, soccer, golfing,
12      swimming, cycling) significantly increased UV radiation exposure (Moehrle, 2001; Moehrle
13      et al., 2000; Thieden et al., 2004a,b).
14
15      10.2.2.2  Occupation
16           Of the various factors that affect human exposure to UV radiation, occupation is also
17      important. Approximately 5% of the American workforce work outdoors, as determined by the
18      EPA NHAPS (Godar et al., 2001).  On average, American indoor workers spend -10% of their
19      day outdoors.  During their time outdoors, they are exposed to -30% of the total ground-level
20      UV flux, as measured by the EPA UV-monitoring program  (Godar et al., 2001). Compared to
21      indoor or in-home workers, outdoor workers are exposed to much higher levels of UV radiation
22      (Kimlin et al., 1998a; Thieden et al., 2004a), frequently at levels that are above current exposure
23      limits set by the International Commission on Non-Ionizing Radiation Protection (ICNIRP,
24      2004).  For example, Thieden et al. (2004a) observed that the annual UV radiation dose,
25      estimated using personal electronic UV dosimeters and exposure diaries, was -70% higher for
26      gardeners than indoor workers.  The gardeners received the majority (55%) of their UV radiation
27      dose on working days (Thieden et al., 2004a). Another study found that outdoor workers
28      received three to four times the annual UV radiation exposure of indoor workers (Diffey, 1990).
29      At-risk working populations  include farmers (Airey et al., 1997; Schenker et al., 2002),
30      fishermen (Rosenthal et al., 1988), landscapers (Rosenthal  et al., 1988), building and
31      construction workers (Gies and Wright, 2003), physical education teachers (Vishvakarman et al.,

        August 2005                              10-15       DRAFT-DO NOT QUOTE OR CITE

-------
 1      2001), mail delivery personnel (Vishvakarman et al., 2001), and various other workers who
 2      spend the majority of their day outdoors during peak UV radiation hours.
 3
 4      10.2.2.3  Age
 5           Age may be a factor that influences human exposure to UV radiation. In a large U.S. study
 6      using the EPA NHAPS, the average UV radiation dose among American children (age
 7      <12 years) was estimated to be slightly higher (-20%) than that of adolescents (age 13 to
 8      19 years) (Godar, 2001). A large Canadian survey found that 89% of children  (age <12 years)
 9      had 30 minutes or more of daily UV exposure compared to 51% for both adults (age 25+ years)
10      and youth (age 15 to 24 years) (Lovato et al., 1998a, 1998b; Shoveller et al., 1998).  In an
11      English study (Diffey et al., 1996), UV radiation exposure was estimated in 180 children (age 9
12      to 10 years) and adolescents (age 14 to 15 years) using personal film badges and exposure
13      records.  Once again, children were found to have received higher UV radiation exposure
14      compared to adolescents (Diffey et al., 1996). However, as discussed earlier, a Danish study
15      found that the annual UV radiation dose in teenagers (age 13 to 19 years) was 14-24% higher
16      compared to children (age 1 to 12 years) and adults (Thieden et al., 2004b).  This increase in UV
17      radiation dose in the Danish teenagers was attributed to their increased risk-behavior days.
18      Therefore, age may affect human exposure to UV radiation by influencing other factors of
19      exposure, such as outdoor activity and risk behavior.
20           Two studies examined lifetime UV radiation exposure among persons in the U.S. (Godar
21      et al., 2003) and Denmark (Thieden et al., 2004b).  Both studies observed that while there are
22      slight differences in UV radiation dose by age, generally people receive fairly consistent UV
23      doses at different age intervals throughout their lives.
24
25      10.2.2.4    Gender
26           Studies have indicated that females generally spend less time outdoors and, consequently,
27      have lower UV radiation exposure compared to males (Gies et al., 1998; Godar et al., 2001;
28      Shoveller et al.,  1998). The U.S. study by Godar et al. (2001) observed that while both males
29      and females had relatively consistent erythemal UV radiation doses throughout their lives, males
30      consistently received higher overall UV doses compared to females at all age groups. Among all
31      Americans, the lowest exposure to UV radiation  was received in females during their child-

        August 2005                              10-16      DRAFT-DO NOT QUOTE OR CITE

-------
 1      raising years (age 22 to 40 years) (Godar et al., 2001).  The highest exposure was observed in
 2      males aged 41 to 59 years in the U.S. study (Godar et al., 2001).  A similar Canadian survey
 3      found that younger adult males had the greatest exposures to UV radiation (Shoveller et al.,
 4      1998).
 5
 6      10.2.2.5  Geography
 7           In the U.S. study by Godar et al. (2001), erythemal UV radiation doses were examined in
 8      persons living in northern and southern regions. Northerners and southerners were found to
 9      spend an equal amount of time outdoors; however, the higher solar flux at lower latitudes
10      significantly increased the annual UV radiation dose for southerners (Godar et al., 2001).  The
11      annual UV radiation doses in southerners were 25 and 40% higher in females and males,
12      respectively, compared to northerners (Godar et al., 2001).  Other studies also have shown that
13      altitude and latitude influence personal exposure to UV radiation (Kimlin et al., 1998b; Rigel
14      etal.,  1999).
15
16      10.2.2.6  Protective Behavior
17           Protective behaviors such as using sunscreen (e.g., Nole and Johnson, 2004), wearing
18      protective clothing (e.g., Rosenthal et al., 1988; Sarkar, 2004; Wong et al., 1996), and spending
19      time in shaded areas (Moise et al., 1999;  Parisi et al., 1999) have been shown to reduce exposure
20      to UV radiation. In one study, the use of sunscreen was associated with extended intentional UV
21      radiation exposure (Autier et al., 1999); however, a follow-up study indicated that sunscreen use
22      increased duration of exposures to doses of UV radiation that were below the threshold level  for
23      erythema (Autier et al., 2000).
24           In a national  study of U.S. youths aged 11 to 18 years, the most prevalent protective
25      behavior was sunscreen use (39.2%) followed by use of a baseball hat (4.5%) (Davis et al.,
26      2002). There were significant differences in the use  of sunscreen by age group and gender, with
27      the younger age group (age 11 to 13 years) and girls  having greater likelihood (47.4 and 48.4%,
28      respectively) of using sunscreen (Davis et al., 2002). The Canadian National Survey on Sun
29      Exposure and Protective Behaviours observed that less than half of the adults (age 25+ years,
30      n = 3,449) surveyed took adequate protective actions (Shoveller et al.,  1998). Once again,
31      children (age <12 years, n = 1,051) were most protected from exposure to UV radiation, with

        August 2005                              10-17       DRAFT-DO NOT QUOTE OR CITE

-------
 1      76% using sunscreen and 36% avoiding the sun, as reported by their parents (Lovato et al.,
 2      1998a). However, the protection level was still not adequate, as indicated by the high 45% rate
 3      of erythema in children. Among Canadian youth (age 15 to 24 years, n = 574), protective
 4      actions from UV radiation exposure included wearing a hat (38%) and seeking shade and
 5      avoiding the sun between the peak hours of 11:00 a.m. to 4:00 p.m.  (26%) (Lovato et al., 1998b).
 6      The lowest prevalence of protective behavior among the youth was  likely responsible for the
 7      highest proportion of erythema (68%) experienced in this age group. A Danish study observed
 8      that both children and teenagers applied sunscreen on more days than adults, but teenagers had
 9      the most days with erythema, due to their increased risk behavior (Thieden et al., 2004b).
10      A survey in Switzerland of 1,285 individuals, including children and parents, indicated that
11      sunscreen use was the protective action most commonly used, but only at the beach and not in
12      routine daily exposure (Berret et al., 2002). In general, protective clothing and avoiding the sun
13      were not highly  used among these individuals to protect against UV-related health effects.
14
15      10.2.2.7   Summary of Factors that Affect Human Exposures to  Ultraviolet Radiation
16          The factors that potentially influence UV radiation doses were discussed in the previous
17      sections and include outdoor activities, occupation, age,  gender, geography, and protective
18      behavior. Results from the various studies indicate that the following subpopulations may be at
19      risk for higher exposures to UV radiation:
20          •  Individuals who engage in high-risk behavior, viz., sunbathing;
21          •  Individuals who participate in outdoor  sports  and activities;
22          •  Individuals who work outdoors with inadequate shade, e.g., farmers, fishermen,
               gardeners, landscapers, building and construction workers; and
23          •  Individuals living in geographic areas with higher solar flux (i.e., lower latitudes
               [e.g., Honolulu, HI] and higher altitudes [e.g., Denver, CO]).
24
25      10.2.3  Factors Governing Human Health Effects  due to Ultraviolet Radiation
26          Ultraviolet radiation occupies a specific region of the electromagnetic spectrum of
27      wavelengths and can be further subdivided into three parts, UV-A (320 to 400 nm), UV-B
28      (280 to 320 nm), and UV-C (200 to 280 nm). Most of the health risks associated with UV
29      radiation exposure are wavelength dependent. Wavelengths <180 nm are of little practical

        August 2005                              10-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1      biological significance as they are almost completely absorbed by the stratosphere (ICNIRP,
 2      2004).
 3           "Action spectra" of a given biological response to UV radiation across its spectral range
 4      are used to estimate exposure by weighting individual wavelength intensities according to the
 5      associated response. The overall effectiveness of the incident flux at inducing the biological
 6      response of interest is computed by means of the relationship:
 7
 o                                                     (•
                                 effective irradiance = J /^ E^ d'k                         flO-4")
 9                                                     A,
10
11      where 1A and EA are, respectively, the irradiance and its relative effectiveness at wavelength A.
12           Until 1980, it was generally thought that wavelengths <315 nm were responsible for the
13      most significant adverse UV radiation health effects; however, recent studies have found that the
14      longer wavelengths in the UV-A range also may produce adverse responses at substantially
15      higher doses (ICNIRP, 2004).  As UV-A radiation is not absorbed by O3, health effects solely
16      induced by UV-A exposure are not relevant in an analysis of public health risks/benefits
17      associated with O3-related UV attenuation.  Therefore, this section focuses on the latest available
18      information on the various  adverse health effects associated with acute and chronic UV-B
19      radiation exposure.
20
21      10.2.3.1  Erythema
22      Association Between Ultraviolet Radiation Exposure and Erythema
23           The most conspicuous and well-recognized acute response to UV radiation is erythema, or
24      the reddening of the skin, which is likely caused by direct damage to DNA by UV-B and UV-A
25      radiation (Matsumura and Ananthaswamy, 2004). Indirect oxidative damage also may occur at
26      longer wavelengths (Matsumura and Ananthaswamy, 2004). Skin type appears to play a large
27      role in the sensitivity to UV radiation-induced erythema. The Fitzpatrick classifications for skin
28      types are: (1) skin type I - individuals with extremely sensitive skin that sunburns easily and
29      severely, and is not likely to tan (e.g., very fair skin, blue eyes, freckles); (2) skin type
30      II - individuals with very sensitive skin that usually sunburns easily and severely, and tans
31      minimally (e.g., fair skin, red or blond hair, blue, hazel or brown eyes); (3) skin type

        August 2005                               10-19       DRAFT-DO NOT QUOTE OR CITE

-------
 1      III - individuals with sensitive skin that sunburns moderately and tans slowly (e.g., white skin,
 2      dark hair); (4) skin type IV - individuals with moderately sensitive skin that sunburns minimally
 3      and usually tans well (e.g., white or light brown skin, dark hair, dark eyes); (5) skin type
 4      V - individuals with minimally sensitive skin that rarely sunburns and tans deeply (e.g., brown
 5      skin); and (6) skin type VI - individuals with nonsensitive skin that never sunburns and tans
 6      profusely (e.g., dark skin). Harrison and Young (2002) found that the perceptible minimal
 7      erythemal dose was approximately twofold greater for individuals with skin type IV compared to
 8      skin type I, although there was considerable overlap in the minimal erythemal dose among the
 9      four skin types.  Waterston et al. (2004) further observed that within an individual, erythemal
10      response differed by body site (e.g., abdomen, chest, front upper arm, back of thigh).  These
11      differences were likely attributable to body site-specific variations in melanin pigmentation.
12          Kollias et al. (2001) investigated the change in erythemal response following a previous
13      exposure to UV radiation. Body sites that received a second exposure to UV radiation always
14      showed a reduced erythemal response compared to body sites with a single exposure,  especially
15      when the first exposure was at levels greater than the minimal erythemal dose. The suppression
16      of erythema was more pronounced when the second exposure was given 48 hours after the first.
17      These findings support the well established notion that repeated exposures to UV radiation
18      results in adaptation (e.g., stimulation of melanogenesis). Kaidbey and Kligman (1981)
19      examined individuals with skin types I, II, and III, and  found that multiple exposures to
20      subthreshold doses of UV radiation at 24-hour intervals resulted in cumulative injury to the skin,
21      as indicated by a lowering of the minimal erythemal dose. These results suggest that a longer
22      time period than 24 hours may be  necessary to repair damage from a single exposure to UV
23      radiation. Henriksen et al. (2004) also observed a lowering of the minimal erythemal  dose with
24      repeated exposure at 24-hour intervals in 49 healthy volunteers with skin types II, III,  IV, and V.
25      However, adaptation was reached after the 4th consecutive exposure. Henriksen et al. further
26      found that the change in threshold depended on skin type.  After 4 days of repeated UV
27      radiation, there was little change (10 to 20%) in the erythemal threshold dose with repeated
28      exposure to UV radiation in the fair-skinned individuals.  Among the darker-skinned individuals,
29      the minimal erythemal dose was lowered by 40 to 50%. However, both the initial UV dose and
30      the dose to erythema after four days of exposure was still higher in the dark-skinned persons.
        August 2005                              10-20       DRAFT-DO NOT QUOTE OR CITE

-------
 1           A reference erythema action spectrum was adopted by the Commission Internationale de
 2      1'Eclairage (International Commission on Illumination, CIE) in 1987 (McKinlay and Diffey,
 3      1987). The CIE erythema action spectrum indicates that UV-B radiation is orders of magnitude
 4      more effective per unit dose than UV-A radiation.
 5
 6      Risk of Erythema from Changes in Tropospheric O3 Levels
 1           There is no literature examining the risk of erythema associated with changes specifically
 8     in tropospheric or ground-level O3 levels. The scientific studies, available to date, focus on the
 9     effects of a reduction in stratospheric ozone. One such study has assessed the effects of
10     stratospheric O3 depletion on the risk of erythema (Longstreth et al., 1998). The analysis by
11     Longstreth et al. (1998) concluded that the risk of erythema would not appreciably increase with
12     depletion of the  stratospheric O3 layer. This is due to the powerful adaptation of the skin to
13     different levels of UV radiation, as evidenced by its ability to cope with changes in UV radiation
14     by season (van der Leun and de Gruijl, 1993).  Gradual exposure to increasing UV radiation from
15     the winter to summer leads to decreased sensitivity of the skin. In midlatitudes, the UV-B
16     radiation in the summer is 10-fold greater than in the winter.  In contrast, the steady depletion of
17     the O3 layer has been estimated to result in an approximately 20% increase in UV-B over 10 years
18     (Longstreth et al., 1998). The comparatively small increase in UV radiation throughout the years,
19     therefore, would not significantly increase the risk of erythema.  Tropospheric O3 constitutes no
20     more than 10% of total atmospheric O3. Given that stratospheric O3 depletion was unlikely to
21     increase the risk of erythema, one could reasonably conclude that small changes in ground-level
22     O3 that take place with attainment of the O3 NAAQS would also not result in increased risk.
23
24      10.2.3.2   Skin Cancer
25           According to the American Academy of Dermatology, one in five Americans develop skin
26      cancer during their lifetime.  The three main forms of skin cancer include basal cell carcinoma
27      and squamous cell carcinoma, which are both nonmelanoma skin cancers, and malignant
28      melanoma.  Nonmelanoma skin cancers constitute more than one-third of all cancers in the U.S.
29      and -90% of all skin cancers, with basal cell carcinoma being approximately four times as
30      common as squamous cell carcinoma (Diepgen and Mahler, 2002; ICNIRP, 2004). The
31      incidence of malignant melanoma is much lower than nonmelanoma  skin cancers. In 2004 more

        August 2005                             10-21       DRAFT-DO NOT QUOTE OR CITE

-------
 1      than one million cases of basal and squamous cell skin cancer are expected to be newly
 2      diagnosed, compared to 40,780 cases of melanoma (Jemal et al., 2004).  However, melanoma
 3      has great metastatic potential and accounts for the majority of skin cancer deaths.
 4           Exposure to UV radiation is considered to be a major risk factor for all three forms of skin
 5      cancer (Gloster and Brodland,1996; Diepgen and Mahler, 2002; IARC, 1992). Ultraviolet
 6      radiation is especially effective in inducing genetic mutations and acts as both a tumor initiator
 7      and promoter. Keratinocytes have evolved DNA repair mechanisms to correct the damage
 8      induced by UV; however, mutations can occur, leading to skin cancers that are appearing with
 9      increasing frequency (Hildesheim and Fornace, 2004). The relationship between skin cancer and
10      chronic exposure to UV radiation is further explored below, followed by discussion of the
11      influence of O3 on the incidence of skin cancer.
12
13      10.2.3.3  Ultraviolet Radiation Exposure and  the Incidence of Nonmelanoma Skin Cancers
14           The incidence of all three types of cancers  has been shown to rise with increasing UV
15      radiation concentrations across the U.S. (de Gruijl, 1999); however, the most convincing
16      evidence for a causal relationship exists between UV radiation and squamous cell carcinoma.
17      Squamous cell carcinoma occurs almost exclusively on skin that is regularly exposed to the sun,
18      such as the face, neck, arms, and hands. The incidence is higher among whites in areas of lower
19      latitudes, where solar flux is greater (Kricker et al., 1994). The risk of squamous cell carcinoma
20      was shown to increase with life-long accumulated exposure to UV radiation in one cross-
21      sectional study (Vitasa et al., 1990); however, increased risk was found to be associated only
22      with exposure 10 years prior to diagnosis in a case-control study (Gallagher et al., 1995a).  One
23      of the major concerns with both types of studies  is the potential for recall bias in reporting past
24      UV radiation exposure by individuals already aware of their disease status.
25           Ultraviolet radiation also has been linked to basal cell carcinoma.  Basal cell carcinoma is
26      common on the face and neck (80-90%) but rarely occurs on the back of the hands (de Gruijl,
27      1999).  While cumulative UV radiation exposure was not associated with an increased risk of
28      basal cell carcinoma (Vitasa et al., 1990), increased risk was observed in individuals with greater
29      recreational UV radiation exposure in adolescence and childhood (age <19 years) and
30      individuals with a history of severe erythema in childhood (Gallagher et al., 1995b). Once again,
31      consideration must be given to potential recall bias in  assessing these results.  Thus, there is

        August 2005                              10-22      DRAFT-DO NOT QUOTE OR CITE

-------
 1      suggestive evidence that UV radiation also plays a role in the development of basal cell
 2      carcinoma, but the etiologic mechanisms for squamous cell carcinoma and basal cell carcinoma
 3      likely differ. In an Australian study conducted in a subtropical community, the factors of having
 4      fair skin, a history of repeated sunburns, and nonmalignant solar skin damage diagnosed by
 5      dermatologists were strongly associated with both types of nonmelanoma skin cancer (Green
 6      et al., 1996). The authors attributed the finding that outdoor occupation was not associated with
 7      nonmelanoma skin cancer to self-selection. Individuals with fair or medium complexions and a
 8      tendency to sunburn accounted for more than 80% of the community study sample; however,
 9      they were systematically underrepresented among outdoor workers (Green et al., 1996).  Such
10      self-selection bias might partly explain the lack of consistent quantitative evidence of a causal
11      link between UV radiation and skin cancer in humans.
12           De Gruijl et al. (1993) assessed the action spectrum for nonmelanoma skin cancers using
13      hairless albino mice. Human data are not available regarding wavelength dependence of the
14      carcinogenicity of UV radiation.  After  adjusting for species differences, the Skin Cancer
15      Utrecht-Philadelphia action spectrum indicated the highest effectiveness in the UV-B range with
16      a maximum at 293 nm, which dropped to  10~4 of this maximum at the UV-A range above
17      340 nm (de Gruijl et al., 1993). The mutations commonly present in thep53 tumor suppressor
18      gene  in individuals with squamous cell  carcinoma and basal cell carcinoma are called the
19      "signature" mutations of UV-B radiation (de Gruijl, 2002). UV-B radiation is highly mutagenic,
20      because DNA is a chromophore for UV-B, but not for UV-A radiation (Ichihashi et al., 2003).
21      Nevertheless, other studies have found that UV-A radiation, in addition to UV-B radiation, can
22      induce DNA damage (Persson et al, 2002; Hunger et al., 2000). DNA damage by UV-A is
23      mediated by reactive oxygen species, making it indistinguishable from damage caused by other
24      agents that generate reactive oxygen species (de Gruijl, 2002).  Epidemiologic evidence of a
25      carcinogenic effect of UV-A was found in a study of psoriasis patients receiving oral psoralen
26      and UV-A radiation treatment (Stern et  al., 1998). High-dose exposure to oral psoralen and
27      UV-A radiation was associated with a persistent, dose-related increase in the risk of squamous
28      cell cancer. Risk of basal cell cancer also was increased in those patients exposed to very high
29      levels of UV-A radiation. Therefore, although UV-B radiation has long been considered the
30      main culprit for nonmelanoma skin cancer, limited evidence suggest that UV-A radiation may
31      also play a role.

        August 2005                              10-23        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Susceptible populations for nonmelanoma skin cancers include individuals with reduced
 2      capacity for nucleotide excision repair, the primary repair mechanism for UV radiation-induced
 3      DNA lesions (Ichihashi et al., 2003).  At particular risk are individuals with xeroderma
 4      pigmentosum, as they have defective nucleotide excision repair in all tissues (Kraemer, 1997;
 5      Sarasin, 1999).  Skin type also largely affects susceptibility to skin cancer. Of the six skin
 6      phenotypes, the  most sensitive individuals are those with skin types I and II, who have a fair
 7      complexion, blue or green eyes, and red or blond hair (Diepgen and Mahler, 2002).  These
 8      individuals tend to sunburn easily, tan poorly, and freckle with sun exposure.  A history of
 9      repeated sunburns also appears to increase the risk of both cancers, while sunburns during
10      childhood are more associated with increased basal cell carcinoma (Gallagher et al., 1995b;
11      Green etal., 1996).
12
13      Ultraviolet Radiation and the Incidence of Cutaneous Malignant Melanoma
14           From 1973 to 1994, the incidence rate of melanoma increased 120.5% along with an
15      increased mortality rate of 38.9% among whites in the U.S. (Hall et al., 1999).  The ICNIRP
16      (2004) states that during the past 40 years  or so, each decade has seen a twofold increase in the
17      incidence of malignant melanoma in white populations, with increased incidence observed more
18      prominently in individuals living in lower latitudes.  Cutaneous malignant melanoma has a
19      mutifactorial etiology with environmental, genetic, and host factors (Lens and Dawes, 2004).
20      The major environmental factor of malignant melanoma has been identified as UV radiation
21      exposure (Diepgen and Mahler, 2002); therefore, the increased incidence of melanoma
22      throughout the years might be partially attributable to changes in human activity patterns (e.g.,
23      increased outdoor activity) that influence UV exposure or increased UV radiation at the ground
24      level.  The risk of melanoma appears to depend on the interaction between the nature of the
25      exposure and skin type (Lens and Dawes, 2004).
26           Fears et al. (2002) examined the association between invasive cutaneous melanoma and
27      UV radiation in  non-Hispanic whites using a case-control study design.  Lifetime residential
28      history was  coupled with mid-range UV-B radiation flux measurements to reduce exposure
29      misclassification and recall bias. A 10% increase in the average annual UV-B flux was
30      significantly associated with a 19% (95% CI: 5, 35) increase in individual odds for melanoma in
31      men and a 16% (95% CI: 2, 32) increase in women. Whiteman et al. (2001) conducted a

        August 2005                              10-24        DRAFT-DO NOT QUOTE OR CITE

-------
 1      systematic review of studies that examined the association between childhood UV radiation
 2      exposure and risk of melanoma. Researchers found that ecological studies assessing ambient
 3      sun exposure consistently reported higher risks of melanoma among people who resided in an
 4      environment with high UV radiation during their childhood (Whiteman et al., 2001). The lack of
 5      consistency among the case-control studies was likely due to the varying methods used to assess
 6      UV radiation dose.
 7           While the evidence is  generally suggestive of a causal relationship between UV radiation
 8      and malignant melanoma, possibly conflicting data also has been observed. For example, the
 9      highest occurrence of malignant melanoma is on men's backs and women's legs, areas that do
10      not have prolonged exposure to the sun (Rivers, 2004). This indicates that, unlike nonmelamona
11      skin cancers, malignant melanoma tends to occur in sites of intermittent, intense sun exposure
12      (trunk and legs), rather than in areas of cumulative sun damage (head, neck, and arms) (Swetter,
13      2003).  A study by Whiteman et al. (2003) observed that individuals with melanomas of the
14      trunk had more melanocyte  nevi and less solar keratoses compared to individuals with head and
15      neck melanomas and, suggesting that cutaneous melanomas may arise through two pathways,
16      one associated with melanocyte proliferation and the other with chronic exposure to sunlight.
17      Green et al.  (1999) also found that melanomas of the soles and palms resembled other cutaneous
18      melanomas in their association with sun exposure, but were distinguished from them by their
19      strong positive associations with nevi on the soles.
20           The available data conflict with regard to the relative importance of UV-A versus UV-B in
21      inducing melanomas.  UV-A has a much higher flux rate at the Earth's surface, as it is not
22      absorbed by O3 and it is  able to penetrate more deeply into the skin surface due to its longer
23      wavelength. However, UV-B, as mentioned earlier, is much more energetic and, therefore, more
24      effective in photochemically altering DNA. The individual roles of UV-A and UV-B in the
25      development of cutaneous malignant melanoma have been examined in several studies.
26      A case-control study of 571 patients and 913 matched controls found an elevated odds ratio of
27      1.8 (95% CI: 1.2, 2.7), after adjusting for skin type, hair color, raised nevi, and number of
28      sunburns, for developing malignant melanoma in individuals who regularly used tanning beds,
29      which typically are UV-A sources  (Westerdahl et al., 2000). In a  study by Setlow et al. (1993),
30      an action spectrum using the tropical fish Xiphophorus indicated that UV-A range wavelengths
31      were especially important in malignant melanoma induction.  However, an action spectrum

        August 2005                             10-25       DRAFT-DO NOT QUOTE OR CITE

-------
 1     using the opossum Monodelphis domestica found that the potency of UV-A for melanoma
 2     induction was extremely low compared to that of UV-B (Robinson et al., 2000). A recent study
 3     by De Fabo et al. (2004) examined the differences in wavelength effectiveness using a
 4     hepatocyte growth factor/scatter factor-transgenic mouse model.  The epidermal tissue of these
 5     transgenic mice behaves similar to the human epidermis in response to UV exposure.  Given the
 6     absence of a mammalian melanoma action spectrum, the standardized CIE erythema action
 7     spectrum was used to deliver identical erythemally effective doses. Only UV-B radiation was
 8     found to initiate mammalian cutaneous malignant melanoma.  UV-A radiation, even at doses
 9     considered physiologically relevant, were ineffective at inducing melanoma (De Fabo et al.,
10     2004). Overall, current evidence suggests that UV-B, and not UV-A, is the primary risk factor
11     for malignant melanoma (ICNIRP, 2004).
12           The populations susceptible for malignant melanoma are similar to those for nonmelanoma
13     skin cancers.  Once again, individuals with xeroderma pigmentosum or a reduced capacity for
14     nucleotide excision repair are at increased risk (Tomescu et al., 2001; Wei et al., 2003).
15     Individuals with skin types I and II, or the fair-skin phenotype (blue or green eyes; blond or red
16     hair; skin that freckles, sunburns easily, and does not tan), have increased susceptibility to
17     malignant melanoma (Evans et al., 1988;  Swetter, 2003; Veier0d et al., 2003).  However, the
18     incidence of melanoma was also positively associated with UV radiation in Hispanics and blacks
19     (Hu et al., 2004).  Although the incidence of melanoma is much lower in Hispanics and blacks
20     compared to whites, melanomas in these populations are more likely to metastasize and have  a
21     poorer prognosis (Black et al., 1987; Bellows et al., 2001).  Among children, malignant
22     melanoma appears to have similar epidemiologic characteristics to the  adult form of the disease
23     (Whiteman et al.,  1997). Individuals with intermittent, intense sun exposure, particularly during
24     childhood, were found to have increased risk of melanoma (Whiteman et al., 2001), in contrast
25     to the association between cumulative exposure and increased risk of squamous cell carcinoma.
26     One study found that a personal history of nonmelanoma skin cancer or precancer, higher
27     socioeconomic status, and increased numbers of nevocytic nevi also were associated with
28     increased incidence of melanoma (Evans  et al., 1988).
29
30
        August 2005                             10-26       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Effect of Changes in Tropospheric O3 Levels on Skin Cancer Incidence
 2           The current evidence strongly suggests a causal link between exposure to UV radiation and
 3      the incidence of both nonmelanoma and melanoma skin cancer. Genetic factors, including skin
 4      phenotype and ability to repair DNA, affect an individual's susceptibility to skin cancer.
 5      Quantifying the relationship between UV radiation and skin cancer is complicated by the
 6      uncertainties involved in the selection of an action spectrum and appropriate characterization of
 7      dose (e.g., peak or cumulative levels of exposure, childhood or lifetime exposures). In addition,
 8      there are multiple complexities in attempting to quantify the effect of tropospheric O3 levels on
 9      UV-radiation exposure, as described in Section 10.2.  The absence of published studies that
10      critically examine increased incidence of skin cancer attributable to decreased tropospheric O3
11      levels reflects the significant challenges in determining ground-level O3-related changes in UV
12      radiation exposure. An analysis by Lutter and Wolz (1997) attempted to examine the effects of a
13      nationwide 10 ppb reduction in seasonal average tropospheric O3 on the incidence of
14      nonmelanoma and melanoma skin cancers and cataracts. Their estimate, however, depended
15      upon several simplifying assumptions, ranging from an assumed generalized 10 ppb reduction
16      in O3 column density, national annual average incidence rates for the two types of skin cancer,
17      and simple, linear biological amplification factors.  Further, the methodologies used in this
18      analysis inherently have ignored area-specific factors that are important in estimating the extent
19      to which small, variable changes in ground-level O3 mediate long-term exposures to UV-B
20      radiation.  More reasonable estimates of the human health impacts of enhanced UV-B
21      penetration following reduced surface O3 concentrations require both a solid understanding of
22      the multiple factors that define the extent of human exposure to UV-B at present, and well-
23      defined and quantifiable links between human disease and UV-B exposure. The reader is
24      referred to the U. S. EPA 2002 Final Response to Court Remand (Federal Register, 2003) for
25      detailed discussions of the data and scientific issues associated with the determination of public
26      health benefits resulting from the attenuation of UV-B by surface-level O3.
27           In the absence of studies specifically addressing the reduction of tropospheric O3 (by
28      assuming that the key variable is total column O3 density), inferences could be made concerning
29      the effects of reduced tropospheric O3-related increases in UV-B exposure on the basis of studies
30      focused on stratospheric O3 depletion. Several studies have examined the potential effect of
31      stratospheric O3 depletion on the incidence of skin cancer (de Gruijl, 1995; Longstreth et al.,

        August 2005                              10-27      DRAFT-DO NOT QUOTE OR CITE

-------
 1      1995; Madronich and de Gruijl, 1993; Slaper et al., 1996; Urbach, 1997). Note that several of
 2      the concerns expressed for Lutter and Wolz (1997) are relevant here as well. Stratospheric O3
 3      depletion is likely to increase the ground-level UV-B flux, as O3 absorbs radiation in that
 4      wavelength range with high efficiency.  Because UV-B radiation is primarily implicated in the
 5      induction of skin cancer, especially among persons with skin phenotypes I and II, there is
 6      concern that the depletion of the O3 layer would result in significantly increased incidence of
 7      skin cancers.
 8          Estimation of the increased risk in melanoma associated with stratospheric O3 depletion
 9      cannot be done adequately due to the lack of a mammalian action spectrum for melanoma.
10      Furthermore, the complexity of the UV-related induction mechanism of melanoma adds an
11      additional layer of uncertainty to the calculations. The excess risk in nonmelanoma skin cancers
12      associated with a decrease in stratospheric O3 was estimated using the Skin Cancer Utrecht-
13      Philadelphia action spectrum based on hairless albino mice (Longstreth et al.,  1995).
14      Quantification of how much more UV radiation would reach ground level with each percentage
15      decrease in O3 required several assumptions:  (1) annual doses were an appropriate measure; (2)
16      personal doses were proportional to ambient doses; and, most notably, (3) each percentage
17      decrease in O3 was associated with a 1.2% increase in UV radiation.  Next, the relationship
18      between UV radiation and nonmelanoma skin  cancer incidence was determined: each percent
19      increase in annual UV radiation dose was estimated to cause a 2.5% increase in squamous cell
20      carcinoma and  1.4% increase in basal cell carcinoma over a human lifetime. Incorporating all
21      these factors, Longstreth et al. (1995) calculated that  a sustained 10% decrease in stratospheric
22      O3 concentration would result in 250,000 additional nonmelanoma skin cancer cases per year.
23      Madronich and de Gruijl (1993) noted that the largest percent of O3-induced nonmelanoma skin
24      cancer increases would be at high  latitudes, where baseline incidence of skin cancer is usually
25      small.  Assuming a phaseout of primary O3-depleting substances by 1996, as established by the
26      Copenhagen Amendments in 1992, Slaper et al. (1996) estimated that the number of excess
27      nonmelanoma skin cancers in the U.S. caused by O3 depletion would exceed 33,000 per year (or
28      approximately 7 per  100,000) around the year  2050.
29          However, estimating the increase in nonmelanoma skin cancer incidence attributable to the
30      depletion of the stratospheric O3 layer is marred by uncertainty.  The following statement by
        August 2005                              10-28       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Madronich and de Gruijl (1994) describes the uncertainty of estimating the effect of
 2      stratospheric O3 depletion on the incidence of skin cancer:
 3
 4                Extrapolating trends and effects of UV into the future is very hypothetical due
 5                to uncertainties that arise from atmospheric chemistry, epidemiology, and related
 6                disciplines. The values that we calculated are one plausible measure of the
 7                magnitude of the 03-UV effects... .The timescales for atmospheric change and
 8                skin-cancer development are still far from certain: 03 reductions are expected
 9                to continue well into next century, and the time between UV exposure and
10                development of skin cancer is essentially unknown.
11
12      Therefore, much caution is necessary when assessing and interpreting the quantitative results of
13      excess nonmelanoma skin cancer incidence due to stratospheric O3 depletion. Although the
14      effect of reductions in tropospheric or ground-level O3 concentrations on skin cancer incidence
15      has not been assessed, it would be  expected to be much less compared to the effect from the
16      depletion of the stratospheric O3 layer, given that tropospheric O3 makes up < 10% of the total
17      atmospheric O3.
18
19      10.2.3.4  Ocular Effects of Ultraviolet Radiation Exposure
20      Ultraviolet Radiation Exposure and Risk of Ocular Damage
21           Ocular damage from UV radiation exposures includes effects on the cornea, lens, iris, and
22      associated epithelial and conjunctival tissues. Absorption of UV radiation differs by
23      wavelength, with short wavelengths (<300 nm) being almost completely absorbed by the cornea,
24      whereas longer wavelengths are transmitted  through the cornea and absorbed by the lens
25      (McCarty and Taylor, 2002).  The  most common acute ocular effect of environmental UV
26      exposure is photokeratitis, also known as snowblindness,  caused by absorption of short
27      wavelength UV radiation by the cornea. The action spectrum indicated that maximum
28      sensitivity of the human eye was found to occur at 270 nm (ICNIRP, 2004; Pitts, 1993). The
29      threshold for photokeratitis in humans varied from 4 to 14 mJ/cm2 for wavelengths 220 to
30      310nm.
31           Exposure to longer wavelengths has been shown to cause both transient and permanent
32      opacities of the lens, or cataracts. Extensive toxicologic and epidemiologic evidence supports
33      the causal association between UV radiation and cataracts (Hockwin et al.,  1999; McCarty and
34      Taylor, 2002).  Ultraviolet radiation-induced cataracts are hypothesized to be caused by

        August 2005                               10-29       DRAFT-DO NOT QUOTE OR CITE

-------
 1      oxidative stress leading to increased reactive species in the lens, which then causes damage to
 2      lens DNA and cross-linking of proteins.  Exposure time to low-dose UV radiation was found to
 3      strongly influence cataract formation (Ayala et al., 2000). An action spectrum determined using
 4      young female rats indicated that the rat lens was most sensitive to 300 nm, correcting for corneal
 5      transmittance (Merriam et al., 2000).  Oriowo et al. (2001) examined the action spectrum for
 6      cataract formation using whole cultured lens from pigs. As pigs lens are similar in shape and
 7      size to the human lens, some inferences may be made. Results indicated that the 270 to 315 nm
 8      waveband was most effective in producing UV-induced cataracts in vitro.  However, the
 9      threshold values varied widely within that range, from 0.02 J/cm2 for 285 nm to 0.74 J/cm2 for
10      315 nm (Oriowo et al., 2001). At wavelengths >325 nm, the threshold levels were orders of
11      magnitude larger, with a minimum threshold value of 18.7 J/cm2.
12           An epidemiologic study examined the effects of UV radiation on cataract formation in
13      watermen (e.g., commericial fishermen, boat workers) who worked on Chesapeake Bay, MD
14      (Taylor et al., 1988). Among the 838 individuals surveyed in this study, 111 had cortical
15      cataracts and 229 had nuclear cataracts. Results indicated that UV-B radiation was significantly
16      associated with cortical, but not nuclear, cataract formation.  For a given age, a doubling of
17      cumulative UV-B exposure was associated with a 60% excess risk (95% CI: 1, 164) of cortical
18      cataracts. No association was observed between cataracts and UV-A radiation in this outdoor-
19      working population.
20
21      Risk of Ocular Damage from Changes in  Tropospheric Ozone Levels
22           Cataracts are the  most common cause of blindness in the world.  McCarty et al. (2000)
23      calculated that ocular UV  radiation exposure accounted for  10% of the cortical cataracts in an
24      Australian cohort of 4,744 individuals from both urban and rural areas.  A study by Javitt and
25      Taylor (1994-1995) found that the probability of cataract surgery in the U.S. increased by 3% for
26      each 1° decrease in latitude. These results  suggest that depletion of the stratospheric O3 layer
27      may increase UV radiation-induced cataract formation. After assuming a certain wavelength
28      dependency along with several additional assumptions, every 1% decrease in the stratospheric O3
29      layer was estimated to  be associated with a 0.3 to 0.6% increase in cataracts (Longstreth et al.,
30      1995).  Longstreth et al. (1995) noted that this estimate has  a high degree of uncertainty due to
31      inadequate information on the action spectrum and dose-response relationships. Quantitative

        August 2005                              10-30       DRAFT-DO NOT QUOTE OR CITE

-------
 1      estimates have not been possible for photokeratitis, pterygium, or other UV-related ocular effects
 2      due to lack of epidemiologic and experimental data.
 3           As is the case for all of the other UV-related health outcomes, there is no published
 4      information on the potential effects on cataract formation due to any changes in surface-level
 5      UV flux resulting from decreases in tropospheric O3.
 6
 7      10.2.3.5  Ultraviolet Radiation and Immune System Suppression
 8           Experimental studies have suggested that exposure to UV radiation may suppress local and
 9      systemic immune responses to a variety of antigens (Clydesdale et  al., 2001; Garssen and van
10      Loveren, 2001; Belgrade et al., 1997).  In rodent models, these effects have been shown to
11      worsen the course and outcome of some infectious diseases and cancers (Granstein and Matsui,
12      2004; Norval et al., 1999).  Granstein and Matsui (2004) stated that exposure to UV-B radiation
13      caused immunosuppression in mice ultimately by releasing cytokines that prevent antigen-
14      presenting cells from performing their normal functions and causing direct damage to epidermal
15      Langerhans cells. Noonan et al. (2003) investigated UV skin cancer induction in two strains of
16      reciprocal Fl hybrid mice and found that genetically determined differences in susceptibility to
17      UV-induced immunosuppression was a risk factor for skin cancer.  At high UV radiation doses,
18      mice with greater susceptibility to immune suppression had a larger proportion of skin tumors
19      compared to those with lower susceptibility (Noonan et al., 2003).  In a study by Yoshikawa
20      et al. (1990), development of contact hypersensitivity to dinitrochlorobenzene on irradiated
21      buttock skin was examined. Individuals who failed to develop contact hypersensitivity were
22      considered to be  susceptible to UV-B radiation.  Virtually all skin cancer patients (92%) were
23      susceptibility to UV-B radiation-induced suppression of contact hypersensitivity, compared to
24      approximately 40% of healthy volunteers.  Others studies have observed increased skin cancer in
25      immune suppressed organ transplant patients (Caforio et al., 2000;  Lindelof et al., 2000).
26      Collectively, results from these studies suggest that immune suppression induced by UV
27      radiation may be a risk factor for skin cancer induction (Ullrich, 2005).
28           There is also some  evidence that UV radiation has indirect involvement in viral
29      oncogenesis through the  human papillomavirus (Pfister, 2003). Additional evidence of
30      UV-related immunosuppression comes from an epidemiologic study of 919 patients with rare
31      autoimmune muscle diseases from 15 cities on four continents with variable UV radiation

        August 2005                               10-31       DRAFT-DO NOT QUOTE OR CITE

-------
 1      intensity (Okada et al., 2003).  Ultraviolet radiation was strongly associated with the prevalence
 2      of dermatomyositis, an autoimmune disease distinguished by the presence of photosensitive
 3      pathognomonic rashes (Okada et al., 2003). In patients with the human immunodeficiency virus,
 4      UV-B radiation lead to activation of the virus in their skin through the release of cytoplasmic
 5      nuclear factor kappa B (Breuher-McHam et al., 2001).  In a study by Belgrade et al. (2001),
 6      UV-induced immunosuppression was examined in 185 subjects with different skin
 7      pigmentations. To assess immune suppression, dinitrochlorobenzene was applied to irradiated
 8      buttock skin 72 hours after irradiation.  Differences in sensitivity were unrelated to skin type
 9      based on the Fitzpatrick classification or minimal erythemal dose (Belgrade et al., 2001).
10      However, erythemal reactivity, assessed by the steepness of the erythemal dose-response curve,
11      was shown to be significantly associated with UV-induced immunosuppression. Only subjects
12      with steep erythemal responses, which included individuals with skin types I through V, showed
13      a dose-response relationship between UV exposure and immune suppression (Belgrade et al.,
14      2001).
15           In other studies, UV radiation was associated with decreased autoimmune diseases.
16      Several ecologic studies observed a decreased prevalence of multiple sclerosis, insulin-
17      dependent diabetes mellitus, and rheumatoid arthritis in regions with lower latitude (i.e., higher
18      UV radiation exposure) (Ponsonby et al., 2002). These results may be attributable to UV
19      radiation-induced immunosuppression and UV-B-related production of vitamin D, which has
20      immunomodulatory effects (Cantorna et al., 2000). The protective effects of UV radiation
21      resulting from its active role in vitamin D production are further discussed in the next section.
22           Most action spectrum investigations have concluded that immunosuppression is caused
23      most effectively by the UV-B waveband (Garssen and van Loveren, 2001).  The effects of UV-A
24      on local and systemic immunosuppression have been unclear and inconsistent.  There is some
25      evidence that high doses  of UV-A is protective of immunosuppression induced by UV-B
26      exposure (Halliday et al., 2004).  Given the variety of outcomes of immune suppression and
27      possible mechanisms of effect, little detailed information exists on UV radiation action
28      spectrums and dose-response relationships. The available data are insufficient to conduct a
29      critical risk assessment of UV radiation-induced immunosuppression in humans.
30
        August 2005                              10-32       DRAFT-DO NOT QUOTE OR CITE

-------
 1      10.2.3.6  Protective Effects of Ultraviolet Radiation - Production of Vitamin D
 2           Any risk assessment that attempts to quantify the consequences of increased UV-B
 3      exposure on humans due to reduced ground-level O3 must include consideration of both negative
 4      and positive effects.  A potential health benefit of increased UV-B exposure relates to the
 5      production of vitamin D in humans. Most humans depend on sun exposure to satisfy their
 6      requirements for vitamin D (Holick, 2004).  UV-B photons are absorbed by
 7      7-dehydrocholesterol in the skin, leading to its transformation to previtamin D3, which is rapidly
 8      converted to vitamin D3.  Vitamin D3 is metabolized in the liver, then in the kidney to its
 9      biologically active form of 1.25-dihydroxyvitamin D3.  One minimal erythemal dose produces
10      vitamin D equivalent to an oral dose of 20,000 IU vitamin D, which is 100 times the
11      recommended dietary allowance for adults under 50 years of age (Giovannucci, 2005;
12      Holick, 2004).
13           Vitamin D deficiency can cause metabolic bone disease among children and adults,  and
14      also may increase the risk of many common chronic diseases, including type I diabetes mellitus
15      and rheumatoid arthritis (Holick, 2004). Substantial in vitro and toxicologic evidence also
16      support a role for vitamin D activity against the incidence or progression of various cancers
17      (Giovannucci, 2005; Studzinski and Moore, 1995).  Large geographical gradients in mortality
18      rates for a number of cancers in the U.S. are not explained by dietary or other risk factors;
19      therefore, it has been hypothesized that some carcinomas are due to insufficient UV-B radiation.
20           Published literature indicates that solar UV-B radiation, by increasing vitamin D
21      production, is associated with a reduced risk of cancer. Most of these studies used an ecologic
22      study design, in which latitude gradient was examined in relation to cancer rates. Kimlin  and
23      Schallhorn (2004)  observed that latitude was a valid predictor of vitamin D-producing UV
24      radiation. The strongest evidence exists for an association between UV radiation and reduced
25      risk of colorectal cancer (Giovannucci, 2005; Grant and Garland, 2004; Freedman et al., 2002).
26      Several other  studies also have found an inverse relationship between UV radiation and various
27      other cancers, including cancer of the breast (Freedman et al., 2002; Garland et al., 1990;
28      Gorham et al., 1990; Grant, 2002a; John et al., 1999), ovary (Freedman et al., 2002; Lefkowitz
29      and Garland, 1994), and prostate (Freedman et al., 2002; Hanchette and Schwartz, 1992),  as well
30      as non-Hodgkin lymphoma (Hughes et al, 2004; Hartge et al., 1996).  Eight other cancers (i.e.,
        August 2005                              10-33       DRAFT-DO NOT QUOTE OR CITE

-------
 1      bladder, esophageal, kidney, lung, pancreatic, rectal, stomach, and corpus uteri) have been found
 2      to exhibit an inverse correlation between mortality rates and UV-B radiation (Grant, 2002b).
 3           Using UV-B data from July 1992 and U.S. cancer mortality rates from 1970 to 1994,
 4      premature cancer deaths attributable to insufficient UV-B exposure were analyzed in an ecologic
 5      study (Grant, 2002b). The minimum mortality rate, which was determined as the value
 6      corresponding to the maximum UV-B  dose, was used to calculate the number of premature
 7      deaths. This analysis observed that the annual number of premature deaths from various cancers
 8      due to inadequate UV-B exposures was 21,700 (95% CI: 20,400, 23,400) for white Americans;
 9      1,400 (95% CI:  1,100, 1,600) for black Americans; and 500 (95% CI: 400, 600) for Asian
10      Americans and other minorities.  Uncertainty in the estimations of UV-B exposure limits the
11      confidence for the estimates of excess  cancer deaths attributable to insufficient exposure.
12      Caution is required in interpreting results from ecologic data; however, no strong alternative
13      explanation is indicated in the association observed between UV radiation and the decreased risk
14      of cancer (Giovannucci, 2005). No study has assessed the decreased risk of cancer mortality
15      resulting from increased UV radiation  attributable to decreased tropospheric O3 levels, but the
16      change in risk is expected to be unappreciable.
17           In establishing guidelines on limits of exposure to UV radiation, the ICNIRP agreed that
18      some low-level exposure to UV radiation has health benefits (ICNIRP, 2004). However, the
19      adverse health effects of higher UV exposures necessitated the development of exposure limits
20      for UV radiation.  The ICNIRP recognized the challenge in establishing exposure limits that
21      would achieve a realistic balance between beneficial and adverse health effects.
22
23      10.2.4   Summary and Conclusions for Ozone Effects on UV-B Flux
24           Latitude and altitude are primary variables in defining UV-B flux at the Earth's surface,
25      immediately followed in importance by clouds, surface albedo, particulate matter concentration
26      and composition, and then by gas phase pollution. Of all of these, only latitude and altitude can
27      be defined with small uncertainty in any effort to develop a UV climatology for use in a public
28      health benefits analysis relevant to the areas not presently attaining the NAAQS for O3. Cloud
29      cover, and its effect on surface UV flux,  continues to be extremely difficult to define and predict.
30      Particulate matter and gas-phase tropospheric pollutants are subject to similarly high degrees of
31      uncertainty in predicting their relative  concentration distributions. Land cover and,

        August 2005                              10-34       DRAFT-DO NOT QUOTE OR CITE

-------
 1      consequently, surface albedos are highly variable at the geographic scales relevant to NAAQS
 2      attainment.
 3           Within the uncertain context of presently available information on UV-B surface fluxes, a
 4      risk assessment of UV-B-related health effects would need to factor in human habits (e.g., daily
 5      activities, recreation, dress, and skin care) in order to adequately estimate UV-B exposure levels.
 6      Little is known about the impact of variability in these human factors on individual exposure to
 7      UV radiation. Furthermore, detailed information does not exist regarding the relevant type (e.g.,
 8      peak or cumulative) and time period (e.g., childhood, lifetime, or current) of exposure,
 9      wavelength dependency of biological responses, and interindividual variability in UV resistance.
10      Recent reports of the necessity of UV-B in the production of vitamin D - a vitamin in which
11      many individuals are deficient - suggests that increased risks of human disease due to a slight
12      excess in UV-B radiation exposure may be offset by the benefits of enhanced vitamin D
13      production. However, as with other impacts of UV-B on human health, this beneficial effect of
14      UV-B has not been studied in sufficient detail to allow for a credible health benefits assessment.
15      In conclusion, the effect of changes in surface-level O3 concentrations on UV-induced health
16      outcomes cannot yet be critically assessed within reasonable uncertainty.
17
18
19      10.3  TROPOSPHERIC OZONE AND CLIMATE CHANGE
20           Water vapor, CO2, O3, N2O, CH4, CFCs, and other polyatomic gases present in the Earth's
21      troposphere, trap infrared radiation emitted by the Earth's surface, leading to  surface warming.
22      This phenomenon is widely known as the "Greenhouse Effect" (Arrhenius, 1896), and the gases
23      involved are known as "greenhouse gases" (GHGs).  The term used  for the role a particular
24      atmospheric component, or any other component of the greater climate system, plays in altering
25      the Earth's radiative balance is "forcing." In the past decade, the global atmospheric  sciences
26      and climate communities have made significant progress in determining the specific role O3
27      plays in forcing climate.
28           The Intergovernmental Panel on Climate Change (IPCC) was founded in 1988 by the
29      World Meteorological Society (WMO) and the United Nations Environmental Program (UNEP)
30      to support the work of the Conference of Parties (COP) to the United Nations Framework
31      Convention on Climate Change (UNFCCC). Drawing from the global climate and atmospheric

        August 2005                              10-3 5       DRAFT-DO NOT QUOTE OR CITE

-------
 1      sciences community for its authors and reviewers, the IPCC produces reports containing
 2      thorough assessments of the available peer-reviewed science regarding the physical climate
 3      system, past and present climate, and evidence of human-induced climate change.  This section
 4      will summarize the reviews of the available information on the forcing properties of tropospheric
 5      O3 as provided by the IPCC Third Assessment Report (IPCC, 200la), and will also describe
 6      some of the more recent developments on the  subject.
 7           The projected effects of global climate change will be briefly explained to provide the
 8      context within which O3 serves as a regional, and possibly global, anthropogenic pollutant.
 9      The concept of climate forcing is also explained, along with the factors that influence the extent
10      of climate forcing by O3.  The section concludes with a summary of the various estimates that
11      have been placed on the amount of globally averaged forcing due to O3.
12
13      10.3.1  The Projected Impacts of Global  Climate Change
14           The study of the atmospheric processes involved in global  climate change, and its potential
15      consequences for human health and global ecosystems, is an  area of active research.  The IPCC
16      Third Assessment Report (TAR) is the most thorough evaluation available of the science
17      concerning climate change. In addition to the first and  second IPCC assessments in 1990 and
18      1995, along with other IPCC reports, earlier assessments included those conducted by the UNEP
19      (1986), the WMO (1988), the U.S. Environmental Protection Agency (1987), and others (e.g.,
20      Patz et al., 2000a,b).  The reader is referred to those documents for a complete discussion of
21      climate change science. An abbreviated list of the  IPCC conclusions to date and a short
22      discussion of the potential impacts of climate change on human health and welfare is provided
23      here to serve as the context for the discussion  of the role of the increasing tropospheric O3
24      concentration in climate change.
25           According to various historic and modern measurement records, atmospheric GHG
26      concentrations have increased dramatically in the past century, and have been attributed to
27      human activities. The IPCC TAR describes the scientific theory and evidence linking increases
28      in GHGs to human activities  (IPCC, 200la).
29           An increasing body  of geophysical observations shows that the Earth is warming and that
30      other climate changes are underway.  These observations include the global surface temperature
31      record assembled since the year 1860, the satellite temperature record begun in 1979, recorded

        August 2005                              10-36       DRAFT-DO NOT QUOTE OR CITE

-------
 1      changes in snow and ice cover since the 1950s, sea level measurements taken throughout the
 2      20th century, and sea surface temperature observations recorded since the 1950s.
 3           Observations (Levitus et al., 2005) show that -84% of the total heating of the Earth System
 4      (oceans, atmosphere, continents, and cryosphere) over the last 40 years has gone into warming
 5      the oceans.  Barnett et al. (2005) have reported the emergence of a clear pattern of ocean surface
 6      warming associated with anthropogenic GHGs. The authors constructed a model-based
 7      fingerprint (i.e., a map of predicted changes in the vertical temperature profile of the Earth's six
 8      major oceans), and compared this map to the newly upgraded and expanded ocean temperature
 9      data set (Levitus et al., 2005).  They concluded that the warming signal far exceeds what would
10      be expected from natural variability, a finding that was in compelling agreement with GHG-
11      forced model profiles. Other evidence of ocean warming includes a marked increase in the
12      frequency, intensity, and persistence of the zonal atmospheric circulation shifts known as the El
13      Nino-Southern Oscillation (ENSO) over the past 100 years. ENSO events occur when the
14      tropical Pacific Ocean has accumulated a large, localized mass of warm water that interrupts
15      cold surface currents along South America, altering precipitation and temperature patterns in the
16      tropics, subtropics, and the midlatitudes.
17           IPCC (1998, 200la) reports also describe the results of general circulation model (GCM)
18      studies predicting that human activities will alter the climate system in a manner likely to lead to
19      marked global and regional changes in temperature, precipitation, and other climate properties.
20      These changes are expected to increase the global mean sea level as well as increase the number
21      of extreme weather events such as floods and droughts, increased wind speeds and precipitation
22      intensity of tropical cyclones, and changes in soil moisture.  These predicted changes can be
23      expected to directly impact human health, ecosystems, and global economic sectors (e.g.,
24      hydrology and water resources, food and fiber production) (IPCC, 1998, 2001b). Table 10-1
25      summarizes these projected impacts.
26           Wide variations in the course and net impacts of climate change in different geographic
27      areas are expected.  In general, the projected changes constitute additional stressors on natural
28      ecosystems and human societal systems already impacted by increasing resource demands,
29      unsustainable resource management practices, and pollution. Some of the predicted changes
30      include alterations in ecological balances; in the availability of adequate food, water, clean air;
        August 2005                              10-37       DRAFT-DO NOT QUOTE OR CITE

-------
      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
 Likelihood3
Representative Examples of Projected Impacts*
(all high confidence of occurrence in some areas')
 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)6
 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 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
August 2005
10-38
DRAFT-DO NOT QUOTE OR CITE

-------
             Table 10-1 (cont'd). Examples of Impacts Resulting From Projected Changes in
                                         Extreme Climate Events
         Projected changes during the 21st Century
         in Extreme Climate Phenomena and their
         Likelihood3
Representative Examples of Projected Impacts'"
(all high confidence of occurrence in some areas')
         Complex Extremes (cont'd)
         Increased intensity of midlatitude storms
         (little agreement between current models)d
  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, 200la).
         bThese impacts can be lessened by appropriate response measures.
         "High confidence refers to probabilities between 67 and 95%.
         Information from TAR WGI, Technical Summary.
         eChanges in regional distribution of tropical cyclones are possible but have not been established.
         Source: IPCC(2001b).
 1      and in human health and safety. Poorer nations can be expected to suffer the most, given their
 2      limited adaptive capabilities.
 3           Although many regions are predicted to experience severe, possibly irreversible, adverse
 4      effects due to climate change, beneficial changes may also take place.  For example, certain
 5      regions may benefit from warmer temperatures or increased CO2 fertilization, e.g., U.S. West
 6      Coast coniferous forests, and some Western rangelands.  Specific benefits may include reduced
 7      energy costs for heating, reduced road salting and snow-clearance costs, longer open-water
 8      seasons in northern channels and ports, and improved agricultural  opportunities in the northern
 9      latitudes as well as in the Western interior and coastal areas.  For further details about the
10      projected effects of climate change on a U.S.-regional scale, the reader is also referred to several
11      regionally-focused reports (MARAT, 2000; Yarnal et al., 2000; NERAG, 2001; GLRAG, 2000),
12      as well as a report on potential impacts of climate change on human health (Bernard et al.,
13      2001a,b).  The IPCC report, "The Regional Impacts of Climate Change," (IPCC,  1998) describes
14      the projected effects of human-induced climate change on various regions of the globe including
15      Africa, the Arctic and Antarctic, the Middle East and arid Asia, Australasia, Europe, Latin
16      America, North America, the small island nations, temperate Asia, and tropical Asia.
        August 2005
10-39
DRAFT-DO NOT QUOTE OR CITE

-------
 1           While current climate models can successfully simulate the present-day annual mean
 2      global climate and the seasonal cycles on a continental scale, they have been less successful on a
 3      regional scale.  Clouds and humidity, essential factors in defining local and regional (sub-grid
 4      scale) climate, are significantly uncertain (IPCC, 200la).  Due to modeling uncertainties, both in
 5      reproducing regional climate and in predicting the future economic activity that will govern
 6      future GHG emissions, the projected impacts discussed above are also uncertain.
 7           Findings from the U.S. Global Change Research Program (USGCRP) (NAST, 2000) and
 8      related reports illustrate the considerable uncertainties and difficulties in projecting likely
 9      climate change impacts at the regional or local scale. The USGCRP findings also reflect the
10      mixed nature of projected potential climate change impacts, i.e., combinations of deleterious and
11      beneficial effects, for U.S. regions and the variation of projected impacts across different
12      regions. Difficulties in projecting region-specific climate change impacts are complicated by the
13      need to evaluate the potential effects of regional- or local-scale changes in key air pollutants not
14      only on global-scale temperature trends, but also on regional- or local-scale temperature and
15      precipitation patterns.  The EPA is currently leading a research effort that uses regional-scale
16      climate models with the goal of identifying changes to O3 and PM concentrations that may occur
17      in a warming climate.  An assessment of the results of this effort will be available by the next
18      review of the O3 NAAQS. This focused effort to determine the impact of a warming climate on
19      criteria air pollution requires regional-scale models with improved skill in reproducing climate
20      history  and predicting change. Among the innovations being employed in this effort is the
21      downscaling of global circulation model outputs to provide boundary conditions for model
22      calculations at the regional scale (Liang et al., 2005). While focusing on projecting the impact of
23      a warming climate on regional O3 concentrations, the effort applied to improving regional-scale
24      modeling will also lead to improved estimates of current and projected future impacts of
25      tropospheric O3 on climate.
26
27      10.3.2   Solar Energy Transformation and the Components of the Earth's
28               Climate System
29           Mass,  in any form, has the capacity to interact with solar and terrestrial radiation, but the
30      manner in which it interacts with radiation is governed by its particular physical form and/or
31      molecular properties. Water provides one of the most interesting examples of how physical form

        August 2005                              10-40       DRAFT-DO NOT QUOTE OR CITE

-------
 1      affects radiative properties. In its gaseous form, water is the most important GHG present in the
 2      climate system, due to its ability to absorb long-wave terrestrial radiation.  Conversely, in its
 3      frozen form as snow or sea ice, water plays a very important role in the climate system by
 4      scattering UV and visible solar radiation back to space, i.e., decreasing the Earth's net solar
 5      radiation receipts by increasing the Earth's reflective properties (albedo). In its liquid aerosol
 6      form as clouds, water also greatly increases the Earth's albedo.  In its bulk liquid form as ocean
 7      water, it absorbs  terrestrial radiation, and represents the Earth's most important reservoir of heat
 8      energy.
 9           The atomic composition and molecular structure of a gas determines the wavelengths of
10      light it can absorb and, therefore, its role in defining the heat capacity of the atmosphere.  Ozone
11      and O2 provide examples of the relative importance of these molecular properties. While these
12      molecules are both composed solely of oxygen atoms, their bond structures are distinct.  Ozone
13      has a three-atom, bent molecular structure, giving it the capacity to absorb terrestrial (infrared)
14      wavelengths - making it a GHG. At any altitude, i.e., in the stratosphere or troposphere, O3 has
15      the capacity to absorb UV radiation of 320 nm and shorter, further increasing the energy -
16      absorbing capacity of the troposphere. Conversely, O2, due to its diatomic, linear structure, is
17      limited to absorbing very short-wave UV light - and does so at altitudes too high to influence the
18      climate system significantly.
19           Each component of the climate system plays a role in absorbing, transforming, storing,
20      dispersing,  or scattering solar radiation. Weather is a tangible consequence of the transformation
21      and dispersion of terrestrial radiation within the atmosphere.  The term  "weather" refers to the
22      condition of the Earth's atmosphere at a specific time and place. It is defined by  several specific
23      variables: the air temperature, air pressure, humidity, clouds, precipitation, visibility, and wind
24      speed. The "climate" for a given place on the Earth's surface is a long-term average of these
25      variables accounting for daily and seasonal weather events. The frequency of extreme weather
26      events is used to  distinguish among climates that have similar averages (Ahrens,  1994).
27           Climate components, including GHGs, land, oceans, sea ice, land ice and snow,
28      atmospheric particles, vegetation, clouds, etc., all contribute to the Earth's heat capacity, i.e., its
29      ability to absorb  and retain solar energy. Changes in the properties (or mass) of these
30      components will  "force" the climate system in one direction or the other, i.e., warmer versus
31      cooler. The transformation of atmospheric O2 into O3 by way of air pollution chemistry,

        August 2005                               10-41       DRAFT-DO NOT QUOTE OR CITE

-------
 1      enhances the heat capacity of the atmosphere.  The principles behind the important concept of
 2      climate forcing are further described, below.
 3
 4      10.3.3  The Composition of the Atmosphere and the Earth's
 5              Radiative Equilibrium
 6           The Greenhouse Effect is the term given to the decreased rate of reemission of absorbed
 7      solar energy due to the heat-retaining properties of the Earth's atmosphere. According to simple
 8      radiative transfer theory, at thermal equilibrium, the Earth's temperature should be near -15 °C.
 9      This is the temperature of a theoretical "black body" that is receiving and then reemitting
10      342.5 WnT2, i.e., the globally averaged amount of full-spectrum solar energy absorbed and then
11      reemitted by the Earth as infrared terrestrial radiation per square meter. In fact, satellite
12      observations well above the atmosphere indicate that the Earth's average planetary temperature
13      is remarkably close to its theoretical black body value at -18 °C, a temperature at which liquid
14      water ordinarily does not exist.
15           At its surface, however, the Earth's average temperature is +15 °C.  The +33 °C
16      temperature differential between the Earth's planetary and  surface temperatures is due to the
17      presence of infrared (TR) radiation-absorbing components in the atmosphere such as water
18      vapor, CO2, CH4, several other trace gases, and some types of particles and clouds.
19           The atmosphere, when cloud-free, is largely transparent in the solar wavelength range.
20      A small fraction of this radiation is absorbed and reemitted as black body radiation by dark
21      atmospheric particles (IPCC, 2001a). However, the majority of clouds and particles, in part,
22      offset the greenhouse effect by increasing the Earth's albedo, thereby decreasing the overall
23      amount of solar radiation absorbed by the Earth system.
24           Ozone, SO2 and NO2 also absorb ultraviolet and near ultraviolet wavelengths, in addition to
25      infrared radiation.  Once absorbed by a gas molecule, the energy introduced by a photon may
26      induce a photochemical reaction with the residual energy thermally exciting (heating) the
27      products of the reaction. Alternatively, the energy introduced into the molecule by the photon
28      may be dispersed amongst neighboring molecules via intermolecular collisions, or reemitted in
29      part as a lower energy (i.e., IR) photon.
30           Radiation from the sun or the Earth's surface that is absorbed by gases and particles is
31      reemitted isotropically, i.e., it is equally likely to be emitted in all directions. Therefore, to a

        August 2005                              10-42       DRAFT-DO NOT QUOTE OR CITE

-------
 1      first approximation, half of the radiation trapped by the Earth's atmosphere is reflected back to
 2      its surface. A portion of this radiation is transformed into the heat energy that drives the
 3      atmospheric processes that form the basis of weather and climate.  Radiation that is not absorbed
 4      by gases and aerosols reaches the Earth's  surface where it is scattered (reflected) or absorbed,
 5      depending on the surface albedo.
 6           Successful modeling of the Earth's climate and, therefore, the assessment of the extent of
 7      human-induced climate change and development of appropriate policy depend on the quality of
 8      available information on the relative efficiencies, amounts, and spatial and temporal distributions
 9      of the various radiatively active components of the atmosphere that absorb and/or reflect solar
10      and terrestrial radiation, along with all the other nonatmospheric components of the Earth
11      system.
12
13      10.3.3.1  Forcing of the Earth's Radiative Balance
14           As mentioned earlier, the commonly used measure of the relative influence of a given
15      component of the climate system on the Earth's radiative balance is its radiative forcing (IPCC,
16      2001a; Houghton et al., 1990).  Radiative forcing, in WnT2, is a quantity that was developed by
17      the climate modeling community as a first-order-only means of estimating relative effects of
18      individual anthropogenic and natural processes on the energy balance within the climate system.
19           When the effect of a particular component of the climate system is to reduce the amount of
20      solar energy absorbed, usually by increasing the Earth's albedo, this component is said to
21      provide a "negative" forcing, measured in WnT2.  The convention assigns a positive value to the
22      forcing induced by climate system components that enhance the Greenhouse Effect or otherwise
23      act to increase the heat-absorbing capacity of the Earth system.  Purely reflective atmospheric
24      aerosol, clouds, white rooftops, snow-covered land surfaces, and dense sea ice provide a
25      negative forcing, while highly absorbing dark-colored atmospheric aerosols, GHGs, and
26      increases dark ocean surface area, due to the melting of sea ice sheets,  positively force the
27      climate system.
28           Global and regional climate are roughly defined by the balance between the large number
29      of positive and negative forcings induced  by the many different components of the Earth system
30      and any changes in the properties of those components due to natural processes or anthropogenic
31      activities.  Following a perturbation or added forcing, such as an increase in GHG concentrations

        August 2005                              10-43       DRAFT-DO NOT QUOTE OR CITE

-------
 1      or modification the Earth's albedo through changes in land use, this balance is re-established via
 2      a complex redistribution of energy within the Earth system. Feedback mechanisms that are
 3      theorized but difficult to resolve at the quantitative level further complicate the prediction of the
 4      sensitivity of variables,  such as surface temperature, to changes in forcing.
 5           A simple example of a positive feedback would be melting sea ice. As sea ice melts with
 6      increasing ocean temperatures, the dark ocean surface that is revealed is more efficient at
 7      absorbing IR radiation, further increasing the rate of warming. A negative feedback would be
 8      the formation of clouds  over a moist, warming surface. As clouds form, less radiation is
 9      available to warm the surface, leading to cooling. The role of feedbacks in determining the
10      sensitivity of climate to  changes in radiative forcing is described in detail in the IPCC TAR
11      (IPCC, 200la).
12           Discussions are presently underway within the climate community regarding a metric to
13      replace forcing as the standard measure of climate impact - one that will account for more of the
14      factors that determine the effectiveness of a specific change in altering climate. However,
15      forcing remains the current standard (NRC, 2005).
16           The IPCC has reported estimated values for forcing by individual radiatively active gases,
17      and by particle-phase components of the atmosphere that were derived primarily through expert
18      judgment incorporating  the results of peer-reviewed modeling studies. The forcing estimates,
19      shown in Figure 10-6, are global averages attributed to known GHGs, including O3, particles,
20      anthropogenic cirrus clouds, land-use change, and natural solar flux variations.  Uncertainty
21      ranges are assigned to reflect the range of modeled values reported in those studies. The current
22      estimate of forcing due to long-lived, well-mixed, GHGs accumulated in the atmosphere from
23      the preindustrial era (ca., 1750) through the year 2000 is +2.4 WnT2 ± 10% (IPCC, 2001a). An
24      indication of the level of confidence in each of these estimates is given along the bottom of this
25      figure, again reflecting the expert judgment of the IPCC.
26           The IPCC reported a global  average forcing value of 0.35 ± 0.15 WnT2 for tropospheric O3,
27      based on model calculations constrained by climatological observations. The considerations and
28      studies used to estimate this value will be outlined below. Hansen and Sato (2001), accounting
29      for uncertainties in pre-industrial emissions levels, more recently estimated a value of 0.5 ± 0.2
30      WnT2 for forcing by O3.
31

        August 2005                              10-44       DRAFT-DO NOT QUOTE OR CITE

-------
V
E
£  a,
re  .g
                    2-
               
-------
 1      system depends upon its geospatial and temporal distribution, but it also depends upon the
 2      albedo of the underlying surface and its vertical position (altitude) in the atmosphere.
 3
 4      10.3.4.1 The Global Burden of Tropospheric Ozone
 5           Little historical data exist that may be used to estimate the global O3 burden prior to
 6      industrialization, although a few late 19th-century measurements suggest that O3 has more than
 7      doubled in Europe during the 20th century.  The insufficient data record on preindustrial
 8      tropospheric O3 distributions introduces a major uncertainty in the estimation of the change
 9      in O3-induced forcing since that period (IPCC, 2001a; Mickley et al., 2004a; Mickley et al.,
10      2001; Shindell and Faluvegi, 2002).
11           Ozone reacts photochemically at time scales that are generally shorter than those for large-
12      scale mixing processes in the atmosphere.  Concentrated O3 plumes evolve downwind of strong
13      sources of its precursor pollutants:  reactive nitrogen, CO, and non-methane hydrocarbons
14      (NMHCs).  The most important of these sources are midlatitude industrialized areas and tropical
15      biomass burning. When viewed from above the atmosphere by satellite-borne spectrometers, O3
16      enhancements appear as relatively localized air masses or regional-scale plumes, usually
17      originating from industrialized areas or areas in which active biomass burning is underway. The
18      IPCC (200la) describes the efforts  of several research teams who have analyzed data supplied by
19      the satellite-borne Total Ozone Mapping Spectrometer (TOMS) and other remote-sensing
20      instruments to map the global distribution of tropospheric O3 and to attempt to identify processes
21      that influence the global tropospheric O3 budget (IPCC, 200la). More recently, coincident
22      observations of total O3 by TOMS and the Solar Backscattered UV (SBUV) instrument were
23      used by Fishman et al.  (2003) to construct well-resolved spatial and temporal maps of the
24      regional distribution of tropospheric O3. Their results were consistent with those reported  by
25      others, but with higher regional-scale resolution. They reported large O3 enhancements in the
26      southern tropics in austral spring, and in the northern temperate latitudes in the summer. The
27      regional nature of high O3 concentrations was clearly visible in northeastern India, the eastern
28      United States, eastern China, and west and southern Africa, each coincident with high population
29      densities. Fishman et al. (2003) noted, as have the other groups cited above, significant
30      interannual variability in the concentrations observed over these regions. In situ measurements
31      of tropospheric O3 concentrations range from 10 ppb over remote oceans to 100 ppb in both the

        August 2005                              10-46       DRAFT-DO NOT QUOTE OR CITE

-------
 1      upper troposphere and in plumes downwind from polluted metropolitan regions (IPPC, 200la).
 2      Ground-level concentrations in urban areas are often >100 ppb. In the Southern Hemisphere,
 3      one of the largest sources of O3 precursors is biomass burning. Biomass burning elevates O3 on
 4      large spatial scales, particularly in the tropical Atlantic west of the coast of Africa and in
 5      Indonesia.
 6           In its third assessment report, the IPPC estimates placed the global burden of tropospheric
 7      O3 at a highly uncertain 370 Tg, equivalent to an average column density of 34 Dobson Units (1
 8      DU = 2.687 x 1016 molecules/cm"2) or  a mean concentration of 50 ppb (IPCC, 200la).
 9      Accounting for differences in levels of industrialization between the hemispheres, the average
10      column burden in the Northern Hemisphere is estimated to be 36 DU, with the Southern
11      Hemisphere estimated to average 32 DU. Due to its rapid photochemistry, individual surface
12      measurements of tropospheric O3 cannot capture large-scale concentrations, nor will they
13      represent the higher altitude concentrations. Dense surface and vertical measurements
14      (ozonesondes) would be required to supplement available output from remote sensing
15      instruments to provide the complete set of observations needed to derive a credible global O3
16      budget.  Such a measurement program  appears, at present, to be impractical.
17
18      10.3.4.2 Background Concentrations versus Regionally-Oriented Ozone Enhancements
19           Vingarzan (2004) surveyed the air quality literature and reported that annual average
20      background O3 concentrations at ground level in the Northern Hemisphere appear to range
21      between 20 and 45 ppb, depending upon geographic location, elevation, and the influence of
22      local sources. Fiore et al. (2003) modeled the U.S. continental O3 concentrations and found that
23      surface background levels overlap the lower end of the range reported by Vingarzan (2004), e.g.,
24      15 to 35 ppb, with higher levels  (40 to  50 ppb) arising at high-elevation sites due to the influence
25      of the upper troposphere (See Chapter 3 and its associated annexes for a complete discussion of
26      "policy relevant background [PRB]). Local- and regional-scale enhancements in O3 may be
27      thought of as roughly superimposed upon these background levels, with the exception of longer
28      stagnation  events in which preexisting background O3 reacts away or is deposited as fresh O3 is
29      produced from local precursors.
30           Lin et al. (2001) analyzed the EPA AIRS database for the 1980-1998 period and noted
31      that O3 concentrations have declined at the high end of the probability distribution, consistent

        August 2005                             10-47       DRAFT-DO NOT QUOTE OR CITE

-------
 1      with the effects of emissions controls, but had increased at the low end of the distribution by 3-5
 2      ppb. They divided the monitoring data for the continental U.S. into 4 quadrants by geography,
 3      and noted a pattern of increase for the Western states that might be attributed to the long-range
 4      transport of O3 precursors from Asia. They found, however, that the Northeastern quadrant had
 5      the highest increase in the low end of the concentration probability distribution, which could not
 6      be reasonably attributed to transboundary transport of O3 precursors.
 7           While not representing an ideal source of information for assessing the climatic effects
 8      of O3 within the continental United States, data from the large air-quality-focused ground-based
 9      monitoring network may be used to identify boundary-layer geospatial and temporal patterns
10      in O3 concentrations for comparison to regional-scale chemistry/climate models. Extensive
11      analysis of data available within the EPA AQS database can be found in Chapter 3 of this
12      document, including an analysis showing the diurnal O3 concentration patterns for several large
13      metropolitan areas with peak values ranging up to!60 ppb (Los Angeles). Lehman et al. (2004)
14      analyzed the AQS database of daily 8-h maximum O3 concentrations collected for 1,090 stations
15      in the eastern half of the United States for the 1993 to 2002 period. They applied a rotated
16      principle component analysis to a reasonably complete, spatially representative, nonurban subset
17      of the database in order to identify coherent, regionally oriented patterns in O3 concentrations.
18      Five spatially homogenous regions were identified: the U.S. Northeast, Great Lakes, Mid-
19      Atlantic, Southwest (including Alabama, Louisiana, Texas, Oklahoma), and Florida. The
20      Mid-Atlantic region displayed the highest mean concentration (52 ppb) of all of the regions
21      analyzed, with the Great Lakes, Southwest, and Northeast regions following with around 47 ppb.
22      The average concentration derived for Florida was 41 ppb.  The authors found strong
23      correlations in measured  concentrations among stations within the same region, suggesting that
24      the geospatial patterns of pollutant emissions and meteorological activity may also have a
25      regional orientation.  These results that these regions may define natural domains for regional
26      scale modeling studies of the influence  of O3 (as well as PM) on climate.
27
28      10.3.4.3  Ozone Trends: Globally and in North America
29           For the Northern Hemisphere, weekly continuous data are available from 1970 for only
30      nine stations in the latitude range 36° N to 59° N (IPCC, 200la). Available tropospheric O3
31      measurements do not reveal a clear trend in concentration, while trends in the stratosphere are

        August 2005                              10-48       DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
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 and increasing emissions in
rapidly growing economies in Asia. The measurements shown in Figure 10-7 are for surface
concentrations only. 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.
                          70

                          65

                          60
                      S"
                      Q.
                      -S:  55
                       M
                      O
                          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).
       August 2005
                                         10-49
              DRAFT-DO NOT QUOTE OR CITE

-------
 1          The IPCC (200la) surveyed the results of published chemistry transport model (CTM)
 2     modelling studies (see Table 10-2) that estimated the global average increases in total column O3
 3     since the preindustrial era. Model estimates ranged from +7 to +12 DU.  On the basis of these
 4     estimates, available measurements, and other analyses, the IPCC estimated that total column O3
 5     has increased by 9 DU with a 67% confidence range of+6 to +13 DU. In some of the modelling
 6     studies, emissions scenarios predicted a further increase in column O3 due to growing emissions
 7     of O3 precursors. Fusco and Logan (2003) stated that, according to models, increased NOX
 8     emissions from fossil fuel combustion have had the greatest effect on O3 in the lower
 9     troposphere since the 1970s. In addition, increases in background CH4 have also contributed as
10     much as 20% to the increase in tropospheric O3, in the northern latitudes.  Given its longer
11     atmospheric residence time, CH4 can serve as an O3 precursor at much longer distances from its
12     source than can other O3 precursors, and, therefore, has a more uniform effect across the globe.
13
14
            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)
Mickleyetal. (1999)
Roelofs etal. (1997)
Stevenson et al. (2000)
Berntsen et al. (1999)
VanDorland etal. (1997)
        Source: IPCC (200la)
 1          Fusco and Logan (2003) found a 10% increase in O3 concentrations year-round over
 2     Canada, Europe, and Japan and a 20% increase for Japan and Europe during spring and summer.
 3     It was expected — but not found — that O3 concentrations over Japan would increase in line

       August 2005                              10-50       DRAFT-DO NOT QUOTE OR CITE

-------
 1      with emissions from China.  The authors suggested that convective activity over Asia is stronger
 2      than that seen over other industrialized areas of the global.  Such a meteorological characteristic
 3      would result in an injection of pollutants into the free troposphere, allowing long-range transport
 4      to North America. Their suggestion is supported by evidence of increasing background
 5      concentrations within the United States (Fiore et al., 1998).
 6           NARSTO (2000), in its assessment of the available information on O3 pollution in North
 7      America, stated that no single pattern for trends in O3 over North America can be found in the
 8      available monitoring data. In the United States, the average 1-h concentration at surface
 9      monitoring sites decreased by 15% between  1986 and 1996, with most of the observed
10      declines occurring in urban and urban-influenced locations.  The largest declines occurred in
11      Los Angeles, New York, and Chicago. Free tropospheric O3 concentrations appeared to hold
12      steady, or only declined slightly, from the 1980s, forward.
13           In preparation for the IPCC TAR (IPCC,  200la), research groups engaged in modeling
14      global-scale tropospheric chemistry were invited to participate in a model intercomparison
15      focusing on potential changes in the oxidative capacity of the atmosphere (OxComp), which
16      included O3 concentrations, for the 2000 to 2100 period. Participating groups employed the
17      IPCC A2p scenario, i.e., including the highest emissions levels, to calculate the geospatial
18      distribution of O3 up to 20 km.  The predicted spatial distributions of O3 where quite variable,
19      but the predictions for total column O3 density  change fell within 9 DU of each other (11.4 to
20      20.5 DU) in all cases and that was considered to be encouraging by the authors.  Fusco and
21      Logan (2003) pointed out that several  unresolved issues may limit the ability of models in
22      reproducing observed trends in tropospheric O3. Among these are the use of different
23      meteorological inputs, photochemical  reaction  schemes, and predicted cloud cover — each
24      contributing to different predictions in O3 production and loss rates.
25
26      10.3.4.4  The Sensitivity of Ozone-Related Forcing Surface to Albedo
27           The characteristics of the surface underlying an O3 enhancement play a role in the O3
28      forcing effect.  Highly reflective surfaces, such as light-colored deserts, sea ice and snow, scatter
29      solar short wave (UV and visible) radiation.  UV and visible radiation can then be absorbed,
30      transformed into long-wave radiation,  and reemitted in part back to the surface by tropospheric
31      O3. Studies by two groups, Hauglustaine et al.  (1998) and Mickley et al. (1999), have shown

        August 2005                              10-51       DRAFT-DO NOT QUOTE OR CITE

-------
 1      that industrial pollution that has been transported to the Arctic induces a high, regional O3-
 2      related forcing due to the highly reflective underlying ice and snow surface.
 3           Liao et al. (2004) calculated that the maximum change in O3-related top-of-the-atmosphere
 4      forcing occurs over high albedo regions in high northern latitudes.  Surface forcing was
 5      calculated to be greatest at high northern latitudes as well as at dust-source regions, which also
 6      tend to have high surface albedos.
 7
 8      10.3.4.5  The Altitude Dependence of Forcing by Tropospheric Ozone
 9           Altitude plays an important role in the forcing effect of tropospheric O3 (IPCC, 1992;
10      Gauss et al., 2003). The efficiency of IR absorption by O3 depends upon its temperature - at
11      atmospheric temperatures that are low, relative to the Earth's surface, it has the capacity to
12      absorb more IR radiation than O3 at temperatures close to that of the surface. While this
13      temperature effect applies to all GHGs, it introduces a  complication for estimating forcing by O3,
14      because O3 is not homogeneously mixed within the troposphere. Ozone forcing estimates must
15      account for these difficult-to-predict vertical inhomogeneities. However, as part of the OxCornp
16      modeling intercomparison, Gauss et al. (2003) found that the overall forcing by O3 can be
17      calculated within reasonable uncertainty simply on the basis of total column density.
18
19      10.3.4.6  Co-occurrence of Ozone with Particulate Matter
20           Analysis of the 2001 data from the AQS database showed infrequent co-occurrence of
21      high PM25 and O3 concentrations (Chapter 3 of this document). For those cases when O3
22      production is high, in combination with high PM concentrations, there is a suggestion in the
23      literature that heterogeneous chemistry on PM surfaces may lead to reduced gas-phase O3.  Liao
24      et al. (2004) modeled heterogeneous chemistry taking place on PM, and found a significant
25      titration of O3 and its NOX precursors. The importance of this titration effect remains an open
26      question, given the difficulty in obtaining in situ measurements to validate model calculations.
27           Liao et al. (2004) also estimated that forcing by BC,  mineral dust, and organic carbon
28      aerosols substantially offsets forcing by tropospheric O3, yielding an overall negative globally
29      averaged forcing at both the top of the atmosphere and at the Earth's surface.  However, such
30      estimates neglect the regional aspects of forcing by these individual pollutants. Elevated
31      concentrations of these very different types  of pollutants often appear independently of the

        August 2005                               10-52       DRAFT-DO NOT QUOTE OR CITE

-------
 1      others, such as with biomass burning plumes, Saharan dust, and organic aerosols associated
 2      with biogenic terpene emissions by forests. It is unlikely that a global average of the forcing
 3      effects of these individual pollutants will adequately capture their impacts on climate at the
 4      regional scale.
 5
 6      10.3.5  Estimated Forcing by Tropospheric Ozone
 7      10.3.5.1 Direct Climate Forcing Due to Ozone
 8           The inhomogeneous distribution of O3 within the troposphere, coupled with the large
 9      uncertainly in the global O3 budget, significantly complicates the matter of estimating the global
10      average direct forcing due to O3.  The IPCC TAR (200la) lists the results of several modeling
11      studies that estimated the annual change in the relative forcing by O3 from preindustrial times.
12      It was noted that the differences among the estimates were most likely due to differences in
13      predicted O3 chemistry, including the emissions inventories used and the chemical process
14      and transport mechanisms incorporated into the models, rather than by factors relating to
15      radiative transfer.  The IPCC intercomparison of the models and their results indicate that the
16      uncertainties in estimated forcings due to O3 have decreased since the IPCC Second Assessment
17      Report (1996).
18           The O3-related forcings estimated by studies considered by the IPCC (200la) are listed in
19      Table 10-3. Ten of the listed estimates are based on global CTM calculations.  One study was
20      constrained by a climatology derived from observations. Given the differences in calculated
21      total column O3 among the models, a normalized forcing (WnT2 per Dobson Unit of
22      tropospheric O3 change) is listed in addition to the absolute forcing (WnT2) estimated by each
23      model.  Both clear sky (cloud-free) and total sky (including clouds) forcing estimates are listed.
24           The largest O3-related forcings coincide with the strongest sources of tropospheric O3,
25      which the models predict occur in the northern midlatitude regions (40° to 50°  N) and  reach as
26      much as 1 WnT2 in the summer as well  as in the tropics, and are related to biomass burning.
27      In general, the estimates are comparable in magnitude and show similarity in geographic
28      distribution.  For total sky conditions, the range in globally and annually averaged
29      tropospheric O3 forcing from all of these models is from 0.28 to 0.43 WnT2, while the
30      normalized forcing is 0.033 to 0.056 WnT2 per DU.  As expected, they are opposite in sign to
31      the forcing estimated for sulfate aerosols, which scatter radiation. The range in normalized

        August 2005                              10-53        DRAFT-DO NOT QUOTE OR CITE

-------
          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 et al. (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).
1
2
3
4
5
forcings emphasizes the differences in assumptions used by the different models.  The
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.
      August 2005
                                        10-54
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Since the publication of the IPCC TAR (200 la), new studies have been published that
 2      illuminate some of the regionally-relevant details associated with direct forcing by O3
 3      (Mickley et al., 2004a; Liao et al., 2004). Forcing by O3, due to its capacity for absorbing solar
 4      UV as well as solar and terrestrial IR radiation, can be divided into "shortwave" forcing and
 5      "long-wave" forcing.  These forcings occur under different conditions.  Shortwave forcing can
 6      only take place during daytime, while long-wave forcing can occur at all hours as a function of
 7      the diurnally varying concentration of atmospheric O3.  As noted, earlier, unlike CO2, the
 8      absorption spectrum for O3 is distinct from that of water vapor — meaning that O3 will absorb
 9      and reemit long-wave radiation without interference by water under high humidity conditions.
10      Mickley et al. (2004a) reported that, according to their modeling study,  surface temperature
11      response to the predicted O3 enhancement since the preindustrial period differs greatly from that
12      of the CO2 response, and that this difference can only be explained by the geographical
13      distribution and absorption properties of O3. Liao et al. (2004) estimated globally averaged
14      top-of-the-atmosphere separate short- and long-wave forcings to O3 to be 0.21 W/m2 and
15      0.32 W/m2, respectively.
16
17      10.3.5.2  Indirect Forcing Due to Ozone
18           Ozone has an indirect climate forcing effect due to its role in the oxidative removal of
19      other reactive GHGs,  including CH4, hydrofluorocarbons (HFCs), and other reactive NMHCs.
20      The primary actor in this effect is a second generation product of the photolysis of O3, the
21      hydroxyl (OH) radical. Hydroxyl radicals are produced by way of a pair of reactions that start
22      with the photodissociation of O3 by solar UV.
23
24
25                                    O3 + hv-» O(!D) + O2                               (10-5)
26
27
28
29                                   O^D) + H2O-» OH + OH                             (10-6)
30
31
32      Reactions with OH are the primary removal mechanism for CH4 and NMHCs as well as the
33      pollutants NOX and CO. Methane and CO are in especially high abundance in the global
34      atmosphere.  OH is estimated to react with these two gases within 1 second of its formation.

        August 2005                              10-55        DRAFT-DO NOT QUOTE OR CITE

-------
 1      In addition to CH4, NOX, CO, and the NMHCs, OH concentrations are controlled by local
 2      concentrations of H2O (i.e., humidity) and the intensity of solar UV. Different atmospheric
 3      concentrations of the required precursors suggest that preindustrial OH concentrations were
 4      likely to have been different from present-day concentrations, but there is no consensus on the
 5      magnitude of this difference. Observations of global atmospheric concentrations of
 6      methylchloroform (CH3CC13), a well-mixed tropospheric species that also reacts with OH, have
 7      been used to estimate OH abundances.  Independent studies have shown overlapping trends for
 8      the period 1978 to 1994, but none of the trends are outside the given uncertainty ranges (0.5 ±
 9      0.6%/year) (Prinn et al., 1995; Krol et al., 1998).  The IPCC (2001a) reported a range of+5% to
10      -20% for predicted changes in global OH abundances.
11           Given the difficulty in estimating  global OH abundances in the past, present, and future,
12      estimates of indirect forcing due to O3 have been  difficult to obtain and are highly uncertain.
13
14      10.3.5.3  Predictions for Future Climate Forcing by Anthropogenic Ozone
15           The rate of increase in surface O3  in Europe and North America since 1980 appears to be
16      slowing, likely due to control measures intended to improve urban air quality.  Not surprisingly,
17      CTM modeling attempts to predict  future precursor emissions and resulting O3 abundances
18      indicate that the largest future O3-related forcings will be related to population growth and
19      economic development in Asia (van Borland et al., 1997; Brasseur et al., 1998). The results of
20      these modeling studies predict that the globally averaged total radiative forcing due to O3 from
21      preindustrial times 0.66  WnT2 will  rise  to 0.63 WnT2 by 2050. Chalita et al. (1996) predicted a
22      change in the globally averaged radiative forcing from preindustrial times to 2050  of 0.43 WnT2.
23      Stevenson et al. (1998) predicted an O3-related forcing of 0.48 WnT2 in 2100. Applying the
24      SRES scenario projecting the highest emissions out to the year 2100 (IPCC, 2000), the OxComp
25      model intercomparison study yielded a  projected  O3-induced forcing ranging from 0.40 to
26      0.78 WnT2. The authors concluded, given their prediction for forcing by well-mixed GHGs of
27      5.6 WnT2, that O3 would remain an important contributor to overall anthropogenic forcing well
28      into the future. However, all of these predictions must be viewed with caution given the
29      considerable uncertainties associated with the long-term economic  activity projections required
30      for such estimates.
31

        August 2005                               10-56       DRAFT-DO NOT QUOTE OR CITE

-------
 1      10.3.6  The Impact of a Warming Climate on Atmospheric Ozone
 2              Concentrations
 3           Evaluation of the potential impact of climate warming on U.S. air quality is currently
 4      underway.  Initial modeling results reported by Mickley et al. (2004b) suggest that reduced
 5      cyclone frequency in a warmer climate will lead to increases in the severity of summertime
 6      pollution episodes. Cyclonic weather patterns are known to play an important role in ventilating
 7      pollution away from the surface.  They note that compelling evidence is accumulating that the
 8      frequency of these cyclones has decreased over the past few decades.  An early study by Jacob
 9      et al. (1993) found a correlation between O3 concentrations and temperature was due to the effect
10      of O3 on atmospheric chemistry, biogenic emissions, and stagnation.
11
12      10.3.7  Conclusion
13           The general consensus within the atmospheric sciences community, as represented by the
14      United Nations Intergovernmental Panel on Climate Change (IPCC), is that human activities
15      have a discernable effect on the Earth's climate. However, quantifying the extent of human-
16      induced forcing on climate requires detailed information about human-induced change on the
17      components of the Earth System that govern climate.  Tropospheric O3 is a well-known GHG,
18      but information regarding its historical trends in concentration, its current and future
19      atmospheric burden, and other critical details needed for estimating its direct and indirect
20      forcing effects on the climate system are highly uncertain.
21           The IPCC has estimated that the globally averaged forcing due to O3 is approximately
22      0.35 ± 0.15 WnT2, with an updated value of 0.5 ± 0.2 WnT2 provided by Hansen and Sato
23      (2001). However, the most important role of O3 in climate is likely to be at the regional scale,
24      adjacent to the sources of its chemical precursors.  This expectation is consistent with satellite
25      observations of high regional scale column O3 densities near large urban areas and large-scale
26      biomass burning activity. Modeling studies evaluated by the IPCC have estimated that regional-
27      scale forcing due to O3 can approach 1 WnT2, or as much as 40% of the globally averaged
28      forcing due to the well-mixed GHGs.  While more certain estimates of the overall importance of
29      global-scale forcing due to tropospheric O3 await further advances in monitoring and chemical
30      transport modeling, the overall  body of scientific evidence suggests that high  concentrations
31      of O3 on the regional scale could have a discernable influence on climate, leading to surface

        August 2005                             10-57       DRAFT-DO NOT QUOTE OR CITE

-------
1     temperature and hydrological cycle changes. Confirming this effect requires improvement in
2     regional-scale modeling — an activity that is currently underway.
3
      August 2005                             10-58       DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES
  2       Ahrens, D. C. (1994) Meteorology today: an introduction to weather, climate and the environment. Minneapolis,
  3             MN: West Publishing Co., p. 17.
  4       Airey, D. K.; Wong, J. C.; Fleming, R. A.; Meldrum, L. R. (1997) An estimate of the UV-B exposure for outdoor
  5             workers during a south-east Queensland summer. Health Phys. 72: 544-549.
  6       Ancellet, G.; Beekmann, M. (1997) Evidence for changes in the ozone concentrations in the free troposphere over
  7             southern France from 1976 to 1995. Atmos. Environ. 31: 2835-2851.
  8       Arrhenius, S. (1896) On the influence of carbonic acid in the air upon the temperature of the ground. Philos. Mag.
  9             41:237-276.
10       Autier, P.; Dore, J. F.; Negrier, S.; Lienard, D.; Panizzon, R.; Lejeune, F. J.; Guggisberg, D.; Eggermont, A. M.
11             (1999) Sunscreen use and duration of sun exposure: a double-blind, randomized trial. J. Natl. Cancer Inst. 91:
12             1304-1309.
13       Autier, P.; Dore, J.-F.; Reis, A. C.; Grivegnee, A.; Ollivaud, L.; Truchetet, F.; Chamoun, E.; Rotmensz, N.; Severi,
14             G.; Cesarini, J. P.; EORTC Melanoma Co-operative Group. (2000) Sunscreen use and intentional exposure to
15             ultraviolet A and B radiation: a double blind randomized trial using personal dosimeters. Br. J. Cancer 83:
16             1243-1248.
17       Ayala, M. N.; Michael, R.; Soderberg, P. G. (2000) Influence of exposure time forUV radiation-induced cataract.
18             Invest. Ophthalmol. Visual Sci. 41: 3539-3543.
19       Barnard, W. F.; Saxena, V. K.; Wenny, B. N.; DeLuisi, J. J. (2003) Daily surface UV exposure and its relationship
20             to surface pollutant measurements. J. Air Waste Manage. Assoc. 53: 237-245.
21       Barnett, T. P.; Pierce, D. W.; AchutaRao, K. M.; Gleckler, P. J.; Santer, B. D.; Gregory, J. M.; Washington, W. M.
22             (2005) Penetration of human-induced warming into the world's oceans. Science 309: 284-287.
23       Bellows, C. F.; Belafsky, P.; Fortgang, I. S.; Beech, D. J. (2001) Melanoma in African-Americans: trends in
24             biological behavior and clinical characteristics over two decades. J. Surg. Oncol. 78: 10-16.
25       Bernard, S. M.; McGeehin, M. G.; Patz, J. A., eds. (2001a) Health: the potential consequences of climate variability
26             and change. Environ. Health Perspect. 109(suppl. 2): 175-233.
27       Bernard, S. M.; Samet, J. M.; Grambsch, A.; Ebi, K. L.; Romieu,  I. (2001b) The potential impacts of climate
28             variability and change on air pollution-related health effects in the United States. Environ. Health Perspect.
29             109[suppl. 2]: 199-209.
30       Berntsen, T. K.; Isaksen, I. S. A.  (1999) Effects of lightning and convection on changes in tropospheric ozone due to
31             NOX emissions from aircraft. Tellus 5IB: 766-788.
32       Berntsen, T. K.; Isaksen, I. S. A.; Myhre, G.; Fuglestvedt, J. S.; Stordal, F.; Larsen, T. A.; Freckleton, R. S.; Shine,
33             K. P. (1997) Effects of anthropogenic emissions  on tropospheric ozone and its radiative forcing. J. Geophys.
34             Res. [Atmos.] 102: 28,101-28,126.
35       Berntsen, T. K.; Myhre, G.; Stordal, F.; Isaksen, I.  S. A. (2000) Time evolution of tropospheric ozone and its
36             radiative forcing. J. Geophys. Res. [Atmos.] 105: 8915-8930.
37       Berret, J.;  Liardet, S.; Scaletta, C.; Panizzon, R.; Hohlfeld, P.; Applegate, L. A. (2002) Use of sunscreens in families
38             living in Switzerland. Dermatology 204: 202-208.
39       Black, W. C.; Goldhahn, R. T., Jr.; Wiggins, C. (1987) Melanoma within a southwestern Hispanic population. Arch.
40             Dermatol.  123: 1331-1334.
41       Brasseur, G. P.; Kiehl, J. T.; Mtiller, J.-F.; Schneider, T.; Granier, C.; Tie, X.-X.; Hauglustaine, D. (1998) Past and
42             future changes in global tropospheric ozone: impact on radiative forcing. Geophys. Res. Lett. 25: 3807-3810.
43       Breuer-McHam, J.; Simpson, E.; Dougherty, L; Bonkobara, M.; Ariizumi, K.; Lewis, D. E.; Dawson, D. B.; Duvic,
44             M.; Cruz, P. D., Jr. (2001) Activation of HIV in human skin by ultraviolet B radiation and its inhibition by
45             NFKB blocking agents. Photochem. Photobiol. 74: 805-810.
46       Bruhl, C.; Crutzen, P. J. (1988) On the disproportionate role of tropospheric ozone as a filter against solar UV-B
47             radiation. Geophys. Res. Lett. 16: 703-706.
48       Caforio, A. L. P.; Fortina, A. B.;  piaserico, S.; Alaibac, M.; Tona, F.; Feltrin, G.; Pompei, E.;  Testolin, L.; Gambino,
49             A.; Volta, S. D.; Thiene, G.; Casarotto, D.; Peserico, A.  (2000) Skin cancer in heart transplant recipients: risk
50             factor analysis and relevance  of immunosuppressive therapy. Circulation 102(suppl. Ill): 111-222 -111-221.
51       Cantorna,  M. T. (2000) Vitamin D and autoimmunity: is vitamin D status an environmental factor affecting
52             autoimmune disease prevalence? Proc. Soc. Exp. Biol. Med. 223: 230-233.
53       Chalita, S.; Hauglustaine, D. A.;  Le  Treut, H.; Mtiller, J. F.; Penkett, S., eds. (1996) Radiative forcing due to
54             increased tropospheric ozone concentrations. Atmos. Environ. 30:  1641-1646.
55       Clydesdale, G. J.;  Dandie, G. W.; Muller, H. K. (2001)  Ultraviolet light induced injury: immunological and
56             inflammatory effects. Immunol. Cell Biol. 79:  547-568.


         August 2005                                   10-59        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Cockell, C. S. (2001) A photobiological history of Earth. In: Cockell, C. S.; Blaustein, A. R., eds. UV radiation flux
  2             into ecosystems. New York, NY: Springer-Verlag.
  3       Cutchis, P. (1974) Stratospheric ozone depletion and solar ultraviolet radiation on Earth. Science (Washington, DC)
  4             184: 13-19.
  5       Davis, K. J.; Cokkinides, V. E.; Weinstock, M. A.; O'Connell, M. C.; Wingo, P. A. (2002) Summer sunburn and sun
  6             exposure among US youths ages 11 to 18: national prevalence and associated factors. Pediatrics 110: 27-35.
  7       De Fabo, E. C.; Noonan, F. P.; Fears, T.; Merlino, G. (2004) Ultraviolet B but not ultraviolet A radiation initiates
  8             melanoma. Cancer Res. 64: 6372-6376.
  9       De Gruijl, F. R. (1995) Action spectrum for photocarcinogenesis. Recent Results Cancer Res. 139: 21-30.
10       De Gruijl, F. R. (1999) Skin cancer and solar UV radiation. Eur. J. Cancer 35: 2003-2009.
11       De Gruijl, F. R. (2002) Photocarcinogenesis:  UVA vs. UVB radiation. Skin Pharmacol. Appl. Skin Physiol. 15:
12             316-320.
13       De Gruijl, F. R.; Henricus, J. C. M.; Sterenborg, H.; Forbes, P. D.; Davies, R. E>; Cole, C.; Kelfkens, G.; Van
14             Weelden, H.; Slaper, H.; VanderLeun, J. C. (1993) Wavelength dependence of skin cancer induction by
15             ultraviolet irradiation of albino hairless mice. Cancer Res. 53: 53-60.
16       Diepgen, T. L.; Mahler, V  (2002) The epidemiology of skin cancer. Br. J. Dermatol. 146(suppl. 61): 1-6.
17       Diffey, B. L. (1990) Human exposure to ultraviolet radiation. Semin. Dermatol. 9: 2-10.
18       Diffey, B. L. (1994) Observed and predicted minimal erythema doses: a comparative study. Photochem. Photobiol.
19             60:380-382.
20       Diffey, B. L.; Gibson, C. J.; Haylock, R.; McKinlay, A. F. (1996) Outdoor ultraviolet exposure of children and
21             adolescents. Br. J. Dermatol. 134: 1030-1034.
22       Evans, R. D.; Kopf, A. W.; Lew, R. A.; Rigel, D. S.; Bart, R. S.; Friedman, R. J.; Rivers,  J. K. (1988) Risk factors
23             for the development of malignant melanoma--!: review of case-control studies. J. Dermatol. Surg. Oncol. 14:
24             393-408.
25       Fears, T. R.; Bird, C. C.; Guerry, D., IV; Sagebiel, R. W.; Gail, M. H.; Elder, D. E.; Halpern, A.; Holly, E. A.;
26             Hartge, P.; Tucker M. A. (2002) Average midrange ultraviolet radiation flux and time outdoors predict
27             melanoma risk. Cancer Res.  62: 3992-3996.
28       Federal Register. (2003) National ambient air quality standards for ozone: final response to remand; final rule. F. R.
29             (January 6) 68: 614-645.
30       Fiore, A. M.; Jacob, D. J.; Logan, J. A.; Yin,  J. H. (1998) Long-term trends in ground level ozone over the
31             contiguous United States, 1980-1995. J. Geophys. Res. (Atmos.) 103: 1471-1480.
32       Fiore, A.; Jacob, D. J.; Liu, H.; Yantosca, R. M.; Fairlie, T. D.; Li, Q. (2003) Variability in surface ozone
33             background over the United States: implications for air quality policy. J. Geophys. Res. (Atmos.) 108(D24):
34             10.1029/2003 JD003855.
35       Fishman, J.; Wozniak, A. E.; Creilson, J. K. (2003) Global distribution of tropospheric ozone from satellite
36             measurements using the empirically corrected tropospheric ozone residual technique: identification of the
37             regional aspects of air pollution. Atmos. Chem. Phys. 3: 893-907.
38       Frederick, J. E.; Koob, A. E.; Weatherhead, E. C. (1993) Empirical studies of tropospheric transmission in the
39             ultraviolet: roadband measurements. J. Appl. Meteorol. 32: 1883-1892.
40       Freedman, D. M.; Dosemeci, M.; McGlynn, K. (2002) Sunlight and mortality from breast, ovarian, colon, prostate,
41             and non-melanoma skin cancer: a composite death certificate based case-control study. Occup. Environ. Med.
42             59:257-262.
43       Fusco, A. C.; Logan, J. A. (2003) Analysis of 1970-1995 trends in tropospheric ozone at Northern Hemisphere
44             midlatitudes with the GEOS-CHEM model. J. Geophys. Res. (Atmos.) 108(D15): 10.1029/2002JD002742.
45       Gallagher, R. P.; Hill, G. B.; Bajdik, C. D.; Coldman, A. J.; Fincham, S.; McLean, D. I.; Threlfall, W. J. (1995a)
46             Sunlight exposure, pigmentation factors and risk of nonmelanocytic skin cancer. II. Squamous cell carcinoma.
47             Arch. Dermatol. 131: 164-169.
48       Gallagher, R. P.; Hill, G. B.; Bajdik, C. D.; Fincham, S.; Coldman, A. J.; McLean, D. I.; Threlfall, W. J. (1995b)
49             Sunlight exposure, pigmentary factors, and risk of nonmelanocytic skin cancer. I. Basal cell carcinoma. Arch.
50             Dermatol. 131: 157-163.
51       Garland, F. C.; Garland, C. F.; Gorham, E. D.; Young, J. F. (1990) Geographic variation in breast cancer mortality in
52             the United States: a hypothesis involving exposure to solar radiation. Prev. Med. 19: 614-622.
53       Garssen, J.; Van Loveren, H. (2001) Effects of ultraviolet exposure on the immune system. Crit. Rev. Immunol. 21:
54             359-397.
55       Gauss, M.; Myhre, G.; Pitari, G.; Prather, M.  J.; Isaksen, I. S. A.; Berntsen, T. K.; Brasseur, G. P.; Dentener, F. J.;
56             Derwent, R. G.; Hauglustaine, D. A.; Horowitz, L. W.; Jacob, D. J.; Johnson, M.; Law, K. S.; Mickley, L. J.;
         August 2005                                    10-60        DRAFT-DO NOT QUOTE OR CITE

-------
  1             Mtiller, J.-F.; Plantevin, P. H.; Pyle, J. A.; Rogers, H. L.; Stevenson, D. S.; Sundet, J. K.; Van Weele, M;
  2             Wild, O. (2003) Radiative forcing in the 21st century due to ozone changes in the troposphere and the lower
  3             stratosphere. J. Geophys. Res. (Atmos.) 108(D9): 10.1029/2002JD002624.
  4       Gies, P.; Wright, J. (2003) Measured solar ultraviolet radiation exposures of outdoor workers in Queensland in the
  5             building and construction industry. Photochem. Photobiol. 68: 78-83.
  6       Gies, P.; Roy, C.; Toomey, S.; MacLennan, R.; Watson, M. (1998) Solar UVR exposures of primary school children
  7             at three locations in Queensland. Photochem. Photobiol. 68: 78-83.
  8       Giovannucci, E. (2005) The epidemiology of vitamin D and cancer incidence and mortality: a review (United
  9             States). Cancer Causes Control 16: 83-95.
10       Gloster, H. M., Jr.; Brodland, D. G. (1996) The epidemiology of skin cancer. Dermatol. Surg. 22: 217-226.
11       Godar, D. E. (2001) UV doses of American children and adolescents. Photochem. Photobiol. 74: 787-793.
12       Godar, D. E.; Wengraitis, S. P.;  Shreffler, J.; Sliney,  D. H. (2001) UV doses of Americans. Photochem. Photobiol.
13             73:621-629.
14       Godar, D. E.; Urbach, F.; Gasparro, F. P.; Van Der Leun, J. C. (2003) UV doses of young adults. Photochem.
15             Photobiol. 77: 453-457.
16       Gorham, E. D.; Garland, F. C.; Garland, C. F. (1990) Sunlight and breast cancer incidence in the USSR. Int. J.
17             Epidemiol.  19: 820-824.
18       Granstein, R. D.; Matsui,  M. S. (2004) UV radiation-induced immunosuppression and skin cancer. Cutis 74(suppl.
19             5): 4-9.
20       Grant, W. B. (2002a) An ecologic study of dietary and solar ultraviolet-B links to breast carcinoma mortality rates.
21             Cancer 94: 272-281.
22       Grant, W. B. (2002b) An estimate of premature cancer mortality in the U.S. due to inadequate doses of solar
23             ultraviolet-B radiation. Cancer 94: 1867-1875.
24       Grant, W. B.; Garland, C. F.  (2004) A critical review of studies on vitamin D in relation to colorectal cancer. Nutr.
25             Cancer 48:  115-123.
26       Great Lakes Regional Assessment Group (GLRAG). (2000) Preparing for a changing climate: the potential
27             consequences of climate variability and change. Great Lakes overview. Washington, DC: U.S. Environmental
28             Protection Agency; Office of Research and Development; U.S. Global Change Research Program (USGCRP)
29             and Ann Arbor, MI: University of Michigan; Atmospheric, Oceanic and Space Sciences Department.
30             Available: http://www.geo.msu.edu/glra/assessment/assessment.html [17  April 2002].
31       Green, A.; Battistutta, D.; Hart, V.; Leslie, D.; Weedon, D. (1996) Skin cancer in a subtropical Australian
32             population:  incidence  and lack of association with occupation. The Nambour Study Group. Am. J. Epidemiol.
33             144:  1034-1040.
34       Green, A.; McCredie, M.; MacKie, R.; Giles, G.; Young, P.; Morton, C.; Jackman, L.; Thursfield, V (1999) A
35             case-control study  of melanomas of the soles and palms (Australia and Scotland). Cancer Causes Control 10:
36             21-25.
37       Hall, H. L; Miller,  D. R.; Rogers, J. D.; Bewerse, B. J. (1999) Update on the incidence and mortality from melanoma
38             in the United States. J. Am. Acad. Dermatol. 40: 35-42.
39       Halliday, G. M.; Byrne, S. N.; Kuchel, J. M.; Poon, T. S.; Barnetson, R.  S. (2004) The suppression of immunity by
40             ultraviolet radiation: UVA, nitric oxide and DNA damage. Photochem. Photobiol. Sci. 3: 736-740.
41       Hanchette, C. L.; Schwartz, G. G. (1992) Geographic patterns of prostate cancer mortality. Cancer 70: 2861-2869.
42       Hansen, J. E.;  Sato, M. (2001) Trends of measured climate forcing agents. Proc. Natl. Acad. Sci. 98: 14778-14783.
43       Harrison, G. I.; Young, A. R.  (2002) Ultraviolet radiation-induced erythema in human skin. Methods (San Diego,
44             CA, U.S.) 28: 14-19.
45       Hartge, P.; Devesa, S. S.; Grauman, D.; Fears, T. R.; Fraumeni, J. F., Jr. (1996) Non-Hodgkin's lymphoma and
46             sunlight. J. Natl. Cancer Inst. 88: 298-300.
47       Hartmann, D. L. (1994) Global physical climatology. New York, NY: Academic Press. (International Geophysics: v.
48             56).
49       Hauglustaine, D. A.; Brasseur, G. P.; Walters, S.; Rasch, P. J.; Mtiller, J.-F.; Emmons, L. K.; Carroll, M. A. (1998)
50             MOZART,  a global chemical transport model for ozone and related chemical tracers 2.  model results and
51             evaluation. J. Geophys. Res. (Atmos.) 103: 28291-28335.
52       Haywood, J. M.; Schwarzkopf, M. D.; Ramaswamy, V. (1998) Estimates of radiative forcing due to modeled
53             increases in tropospheric ozone. J. Geophys. Res. [Atmos.] 103: 16,999-17,007.
54       Henriksen, M.; Na, R.; Agren, M. S.; Wulf, H. C. (2004) Minimal erythema dose after umltiple UV exposures
55             depends on  pre-exposure skin pigmentation. Photodermatol. Photoimmunol. Photomed. 20:  163-169.
         August 2005                                    10-61        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Hildesheim, I; Fornace, A. I, Jr. (2004) The dark side of light: the damaging effects of UV rays and the protective
  2             efforts of MAP kinase signaling in the epidermis. DNA Repair 3: 567-580.
  3       Hockwin, O.; Kojima, M; Sakamoto, Y.; Wegener, A.; Shui, Y. B.; Sasaki, K. (1999) UV damage to the eye lens:
  4             further results from animal model studies: a review. J. Epidemiol. 9(suppl. 6): S39-S47.
  5       Holick, M. F. (2004) Sunlight and vitamin D for bone health and prevention of autoimmune diseases, cancers, and
  6             cardiovascular disease. Am. J. Clin. Nutr. 80(6 suppl.): 1678S-1688S.
  7       Houghton, J. T.; Jenkins, G. J.; Ephraums, J. J., eds. (1990) Climate change: the IPCC scientific assessment.
  8             Cambridge, MA:  Cambridge University Press; p. 55.
  9       Hu, S.; Ma, F.; Collado-Mesa, F.; Kirsner, R. S. (2004) UV radiation, latitude, and melanoma in US Hispanics and
10             blacks. Arch. Dermatol. 140: 819-824.
11       Hughes, A. M.; Armstrong, B. K.; Vajdic, C. M.; Turner, J.; Grulich, A. E.; Fritschi, L.; Milliken, S.; Kaldor, J.;
12             Benke, G.; Kricker, A. (2004) Sun exposure may protect against non-Hodgkin lymphoma: a case-control
13             study. Int. J. Cancer 112: 865-871.
14       Ichihashi, M.; Ueda, M.; Budiyanto, A.; Bito, T.;  Oka, M.; Fukunaga, M.; Tsuru, K.; Horikawa, T. (2003)
15             UV-induced skin damage. Toxicology 189: 21-39.
16       Intergovernmental Panel on Climate Change (IPCC). (1992) Climate change 1992 - the supplementary report to the
17             IPCC scientific assessment. Cambridge, United Kingdom: Cambridge University Press.
18       Intergovernmental Panel on Climate Change. (1996) Climate change 1995: the science of climate change. Summary
19             for policymakers  and technical summary of the working group I report. Presented at: the fifth session of the
20             IPCC Working Group I; November; Madrid, Spain. Geneva, Switzerland: World Meteorological Association.
21       Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
22             of vulnerability, a special report of IPCC Working Group II. Watson, R. T.; Zinyowera, M. C.; Moss, R. H.,
23             eds. Cambridge, United Kingdom: Cambridge University Press.
24       Intergovernmental Panel on Climate Change (IPCC). (2000) Emissions scenarios: a special report of working group
25             III of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom: Cambridge University
26             Press. Available:  http://www.grida.no/climate/ipcc/emission/ [26 August, 2005].
27       Intergovernmental Panel on Climate Change (IPCC). (2001a) Climate change 2001: the scientific basis. Contribution
28             of working group I to the third assessment report of the Intergovernmental Panel on Climate Change.
29             Cambridge, United Kingdom: Cambridge University Press.
30       Intergovernmental Panel on Climate Change (IPCC). (200 Ib) Climate change 2001:  impacts, adaptation, and
31             vulnerability. Contribution of working group II to the third assessment report of the Intergovernmental Panel
32             on Climate Change. Cambridge, United Kingdom: Cambridge University Press.
33       International Agency for Research on Cancer. (1992) Solar and ultraviolet radiation. Lyon, France: International
34             Agency for Research on Cancer. (IARC monographs on the evaluation of carcinogenic risks to humans: v.
35             55).
3 6       International Commission on Non-Ionizing Radiation Protection. (2004) Guidelines on limits of exposure to
3 7             ultraviolet radiation of wavelengths between 180 nm and 400 nm (incoherent optical radiation).  Health Phys.
38             87:  171-186.
39       Jacob, D. J.; Logan, J. A.; Yevich, R. M.; Gardner, G. M.; Spivakovsky, C. M.; Wofsy,  S. C.; Munger, J. W.;
40             Sillman, S.; Prather, M. J.; Rodgers, M. O.; Westberg, H.; Zimmerman, P. R.  (1993) Simulation of
41             summertime ozone over North America. J. Geophys. Res. [Atmos.] 98: 14,797-14,816.
42       Javitt, J. C.; Taylor, H. R. (1994-95) Cataract and latitude. Doc. Ophthalmol. 88: 307-325.
43       Jemal, A.;  Tiwari,  R. C.; Murray, T.; Ghafoor, A.; Samuels, A.; Ward, E.; Feuer, E. J.; Thun, M. J. (2004) Cancer
44             statistics, 2004. CA-Cancer J. Clin. 54: 8-29.
45       John, E. M.; Schwartz, G. G.; Dreon, D.  M.; Koo, J. (1999) Vitamin D and breast cancer risk: the NHANES I
46             Epidemiologic Follow-up Study, 1971-1975 to 1992.  Cancer Epidemiol. Biomarkers Prev. 8:  399-406.
47       Kaidbey, K. H.; Kligman, A.  M. (1981) Cumulative effects from repeated exposures to ultraviolet radiation. J.
48             Invest. Dermatol. 76: 352-355.
49       Kerr, J. B.; McElroy, C. T. (1993) Evidence for large upward trends of ultraviolet-B radiation linked to ozone
50             depletion. Science (Washington, DC) 262: 1032-1034.
51       Kiehl, J. T.; Schneider, T. L.; Portmann,  R. W.; Solomon, S. (1999) Climate forcing due to tropospheric and
52             stratospheric ozone. J. Geophys. Res. [Atmos.] 104: 31,239-31,254.
53       Kimlin, M. G.; Schallhorn, K. A. (2004)  Estimations of the human Vitamin D' UV exposure in the USA. Photochem.
54             Photobiol. Sci. 3: 1067-1070.
         August 2005                                   10-62        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Kimlin, M. G.; Parisi, A. V.; Wong, J. C. (1998a) Quantification of personal solar UV exposure of outdoor workers,
  2             indoor workers and adolescents at two locations in Southeast Queensland. Photodermatol. Photoimmunol.
  3             Photomed. 14:7-11.
  4       Kimlin, M. G.; Wong, J. C.; Parisi, A. V. (1998b) Simultaneous comparison of the personal UV exposure of two
  5             human groups at different altitudes. Health Phys. 74: 429-434.
  6       Kollias, N.; Stamatas, G. N.; Youn, J. I. (2001) Suppression of UVB-induced cutaneous erythema by a previous
  7             UVB exposure. Photochem. Photobiol. 74: 471-476.
  8       Koloutsou-Vakakis, S.; Helmis, C. G.; Assimakopoulos, V.; Glisten, H. (2001) Middle and lower troposphere
  9             aerosol characteristics and ozone concentrations over northwestern Greece during STAAARTE 1997. Atmos.
10             Environ. 35: 1517-1526.
11       Kraemer, K. H. (1997) Sunlight and skin cancer: another link revealed. Proc. Natl. Acad. Sci. U. S. A. 94: 11-14.
12       Kricker, A.; Armstrong, B. K.; English, D. R. (1994) Sun exposure and non-melanocytic skin cancer.  Cancer Causes
13             Control  5: 367-392.
14       Krol, M.; Van Leeuwen, P. J.; Lelieveld, J. (1998) Global OH trend inferred from methylchloroform measurements.
15             J. Geophys. Res. [Atmos.] 103: 10,697-10,711.
16       Krotkov, N. A.; Bhartia, P. K.; Herman, J. R.; Fioletov, V; Kerr, J. (1998) Satellite estimation of spectral surface
17             UVirradiance in the presence of tropospheric aerosols 1. cloud-free case. J. Geophys. Res. (Atmos.) 103:
18             8779-8793.
19       Lehman, J.; Swinton, K.; Bortnick, S.; Hamilton, C.; Baldridge, E.; Ender, B.; Cox, B. (2004) Spatio-temporal
20             characterization of tropospheric ozone across the eastern United States. Atmos. Environ. 38: 4357-4369.
21       Lefkowitz, E. S.; Garland, C. F. (1994) Sunlight, vitamin D, and ovarian cancer mortality rates in U.S. women. Int.
22             J. Epidemiol. 23: 1133-1136.
23       Lelieveld, J.; Dentener, F.  J. (2000) What controls tropospheric  ozone? J. Geophys. Res. [Atmos.] 105: 3531-3551.
24       Lens, M. B.; Dawes, M. (2004) Global perspectives of contemporary epidemiological trends of cutaneous malignant
25             melanoma. Br. J. Dermatol. 150:  179-185.
26       Levitus, S.; Antonov, J.; Boyer, T. (2005) Warming of the world ocean, 1955-2003. Geophys. Res. Lett.
27             32(L02604): 10.1029/2004GL021592.
28       Levy, H., II; Kasibhatla, P. S.; Moxim, W. J.; Klonecki, A. A.; Hirsch, A. I.; Oltmans, S. J.; Chameides, W. L.
29             (1997) The global impact of human activity on tropospheric ozone. Geophys. Res. Lett. 24: 791-794.
30       Liang, X.-Z.; Pan, J. P.; Zhu, J. H.; Kunkel, K. E.; Dai, A.; Meehl, J. (2005)  Regional climate model downscaling of
31             U.S. climate. J. Geophys. Res.: submitted.
32       Liao, H.; Seinfeld, J. H.; Adams, P. J.; Mickley, L. J. (2004) Global radiative forcing of coupled tropospheric ozone
33             and aerosols in a unified general circulation model.  J. Geophys. Res. [Atmos.] 109(D16):
34             10.1029/2003 JD004456.
35       Lin, C.-Y. C.; Jacob, D. J.; Fiore, A. M. (2001) Trends in exceedances of the ozone air quality standard in the
36             continental United States, 1980-1998. Atmos. Environ. 35: 3217-3228.
37       Lindelof, B.; Sigurgeirsson, B.; Ga'bel, H.; Stern, R. S.  (2000) Incidence of skin cancer in 5356 patients following
38             organ transplantation. Br. J. Dermatol. 143: 513-519.
39       Logan, J. A. (1999) An analysis of ozonesonde data for the troposphere: recommendations for testing 3-D models
40             and development of a gridded climatology for tropospheric ozone.  J. Geophys. Res. 104: 16,115-16,149.
41       Longstreth, J. D.; De Gruijl; Kripke, M. L.; Takizawa, Y.; VanderLeun, J. C. (1995) Effects of increased solar
42             ultraviolet radiation on human health. Ambio 24: 153-165.
43       Longstreth, J.; De Gruijl, F. R.; Kripke, M. L.; Abseck, S.; Arnold, F.; Slaper, H. I.; Velders, G.; Takizawa, Y.; Van
44             Der Leun, J. C. (1998) Health risks. J. Photochem. Photobiol.  B 46: 20-39.
45       Lovato, C. Y.;  Shoveller, J. A.; Peters, L.; Rivers, J. K. (1998a)  Canadian National Survey on Sun Exposure &
46             Protective Behaviours: parent's reports on children.  Cancer Prev. Control. 2:  123-128.
47       Lovato, C. Y.;  Shoveller, J. A.; Peters, L.; Rivers, J. K. (1998b) Canadian National Survey on Sun Exposure &
48             Protective Behaviours: youth at leisure. Cancer Prev. Control  2: 117-122.
49       Lutter, R.; Wolz, C. (1997) UV-B screening by tropospheric ozone: implications for the national ambient air quality
50             standard. Environ. Sci. Technol. 31: 142A-146A.
51       Madronich, S.; De Gruijl, F. (1993) Skin cancer and UV radiation. Nature (London, UK) 366(6450): 23.
52       Madronich, S.; De Gruijl, F. (1994) Reply to "Kricker, A.; Armstrong, B. K.; McMichael, A. J. (1994) Skin cancer
53             and ultraviolet. Nature (London, UK) 368: 594." Nature (London, UK) 368(6472): 594.
54       Matsumura, Y.; Ananthaswamy, H. N. (2004) Toxic effects of ultraviolet radiation on the skin. Toxicol. Appl.
55             Pharmacol. 195: 298-308.
         August 2005                                    10-63         DRAFT-DO NOT QUOTE OR CITE

-------
  1       McCarty, C. A.; Nanjan, M. B.; Taylor, H. R. (2000) Attributable risk estimates for cataract to prioritize medical and
  2             public health action. Invest. Ophthalmol. Vis. Sci. 41: 3720-3725.
  3       McCarty, C. A.; Taylor, H. R. (2002) A review of the epidemiologic evidence linking ultraviolet radiation and
  4             cataracts. Dev. Ophthalmol. 35:21-31.
  5       McKenzie, R. L.; Seckmeyer, G.; Bais, A. F.; Kerr, J. B.; Madronich, S. (2001) Satellite retrievals of erythemal UV
  6             dose compared with ground-based measurements at northern and southern midlatitudes. J. Geophys. Res.
  7             [Atmos.] 106:24,051-24,062.
  8       McKenzie, R.; Smale, D.; Bodeker, G.; Claude, H. (2003) Ozone profile differences between Europe and New
  9             Zealand: effects  on surface UV irradiance and its estimation from satellite sensors. J. Geophys. Res. [Atmos.]
10             108(D6): 10.1029/2002JD002770.
11       McKinlay, A. F.; Diffey, B. L. (1987) A reference action spectrum for ultraviolet induced erythema in human skin.
12             In: Passchier, W. F.; Bosnijakovic, B. F. M., eds. Human exposure to ultraviolet radiation: risks and
13             regulations. Amsterdam, The Netherlands: Elsevier Science Publishers; pp. 83-87.
14       Merriam, J. C.; Lofgren, S.; Michael, R.; Soderberg, P.; Dillon, J.; Zheng, L.; Ayala, M. (2000) An action spectrum
15             for UV-B radiation and the rat lens. Invest. Ophthalmol. Vis. Sci. 41: 2642-2647.
16       Mickley, L. J.; Murti, P. P.; Jacob, D. J.; Logan, J. A.; Koch, D. M.; Rind, D. (1999) Radiative forcing from
17             tropospheric ozone calculated with a unified chemistry-climate model. J. Geophys. Res. [Atmos.] 104:
18             30,153-30,172.
19       Mickley, L. J.; Jacob, D. J.; Rind, D. (2001) Uncertainty in preindustrial abundance of tropospheric ozone:
20             implications for  radiative forcing calculations. J. Geophys. Res. (Atmos.) 106(D4): 3389-3399.
21       Mickley, L. J.; Jacob, D. J.; Field, B. D.; Rind, D. (2004a) Climate response to the increase in tropospheric ozone
22             since preindustrial times: a comparison between ozone and equivalent CO2 forcings. J. Geophys. Res.
23             [Atmos.] 109(05): 10.1029/2003 JD003653.
24       Mickley, L. J.; Jacob, D. J.; Field, B. D.; Rind, D. (2004b) Effects of future climate change on regional air pollution
25             episodes in the United States. Geophys. Res. Lett. 31(L24103):  10.1029/2004GL021216.
26       Mid-Atlantic Regional  Assessment (MARA) Team. (2000) Preparing for a changing climate: the potential
27             consequences of climate variability and change. Mid-Atlantic overview. Washington, DC: U.S.
28             Environmental Protection Agency; Office of Research and Development; U.S. Global Change Research
29             Program (USGCRP)  and University Park, PA: Pennsylvania State University. Available:
30             http://www.usgcrp.gov/usgcrp/nacc/midatlantic.htm [17 April 2002].
31       Moehrle, M. (2001) Ultraviolet exposure in the Ironman triathlon. Med. Sci. Sports Exercise 33: 1385-1386.
32       Moehrle, M.; Heinrich, L.; Schmid, A.; Garbe, C. (2000) Extreme UV exposure of professional cyclists.
33             Dermatology 201: 44-45.
34       Moise, A. F.; Buttner, P. G.; Harrison, S. L. (1999) Sun exposure at school. Photochem. Photobiol. 70: 269-274.
3 5       National Assessment Synthesis Team (NAST). (2000) Climate change impacts on the United States: the potential
36             consequences of climate variability and change. Overview. Washington, DC: U.S. Global Change Research
37             Program. Available at:  http://www.usgcrp.gov/usgcrp/Library/nationalassessment/overview.htm (17 January
38             2003).
3 9       National Research Council.  (2005) Radiative forcing of climate change: expanding the concept and addressing
40             uncertainties. Washington, DC: The National Academies Press.
41       New England Regional Assessment Group (NERAG). (2001) Preparing for a changing climate: the potential
42             consequences of climate variability and change. New England regional overview. Washington, DC: U.S.
43             Environmental Protection Agency; Office of Research and Development; U.S. Global Change Research
44             Program (USGCRP)  and Durham, NH: University of New Hampshire; Institute for the Study of Earth,
45             Oceans, and Space. Available: http://www.necci.sr.unh.edu/2001-NERA-report.html [17 April 2002].
46       Newchurch, M. J.; Ayoub, M. A.; Oltmans, S.; Johnson, B.; Schmidlin, F. J. (2003) Vertical distribution of ozone at
47             four sites in the United States. J. Geophys. Res. 108(D1): 10.1029/2002JD002059.
48       Nole, G.; Johnson, A. W. (2004) An analysis of cumulative lifetime solar ultraviolet radiation exposure and the
49             benefits of daily  sun protection. Dermatol.  Ther. 17(suppl. 1): 57-62.
50       Noonan, F. P.; Muller, H. K.; Fears, T. R.; Kusewitt, D. F.; Johnson, T. M.; De Fabo, E. D. (2003) Mice with
51             genetically determined high susceptibility to ultraviolet (UV)-induced immunosuppression show enhanced
52             UV carcinogenesis. J. Invest. Dermatol. 121: 1175-1181.
53       North American Research Strategy for Tropospheric Ozone (NARSTO) Synthesis Team. (2000) An assessment of
54             tropospheric ozone pollution: a North American perspective. Palo Alto, CA: Electric Power Research
55             Institute.
         August 2005                                   10-64        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Norval, M; Garssen, J.; Van Loveren, H.; El-Ghorr, A. A. (1999) UV-induced changes in the immune response to
  2             microbial infections in human subjects and animal models. J. Epidemiol. 9(suppl. 6): S84-S92.
  3       Okada, S.; Weatherhead, E.; Targoff, I. N.; Wesley, R.; Miller, F. W.; International Myositis Collaborative Study
  4             Group. (2003) Global surface ultraviolet radiation intensity may modulate the clinical and imunologic
  5             expression of autoimmune muscle disease. Arthritis Rheum. 48: 2285-2293.
  6       Oriowo, O. M.; Cullen, A. P.; Chou, B. R.; Sivak, J. G. (2001) Action spectrum and recovery for in vitro
  7             UV-induced cataract using whole lenses. Invest. Ophthalmol. Vis. Sci. 42: 2596-2602.
  8       Parisi, A. V.; Willey, A.; Kimlin, M. G.; Wong, J. C. (1999) Penetration of solar erythema! UV radiation in the
  9             shade of two common Australian trees. Health Phys. 76: 682-686.
10       Patz, J. A.; Engelberg,  D.; Last, J. (2000a) The effects of changing weather on public health. Annu. Rev. Public
11             Health 21: 271-307.
12       Patz, J. A.; McGeehin, M. A.; Bernard, S. M.; Ebi, K. L.; Epstein, P. R.; Grambsch, A.; Gubler, D. J.; Reiter, P.;
13             Romieu, L; Rose, J. B.; Samet, J. M.; Trtanj, J. (2000b) The potential health impacts of climate variability and
14             change for the United States: executive summary of the report of the health sector of the U.S. national
15             assessment. Environ. Health Perspect. 108: 367-376.
16       Persson, A. E.; Edstrom, D. W.; Backvall, H.; Lundeberg, J.; Ponten, F.; Ros, A. M.; Williams, C. (2002) The
17             mutagenic effect of ultraviolet-A1 on human skin demonstrated by sequencing the p53 gene in single
18             keratinocytes. Photodermatol. Photoimmunol. Photomed.  18: 287-293.
19       Pfister, H. (2003) Human papillomavirus and skin cancer. J. Natl. Cancer Inst. Monogr. 31: 52-56.
20       Pitts, D. G. (1993) Ocular effects of radiant energy. In: Pitts, D. G.; Kleinstein, R. N., eds. Environmental Vision.
21             Stoneham, MA: Butterworth-Heinemann, pp. 151-220.
22       Ponsonby, A.-L.; McMichael, A.; Van der Mei, I. (2002) Ultraviolet radiation and autoimmune disease:  insights
23             from epidemiological research. Toxicology 181/182: 71-78.
24       Prinn, R. G.; Weiss, R. F.; Miller, B. R.; Huang, J.; Alyea, F. N.; Cunnold, D. M.; Fraser, P. J.; Hartley,  D. E.;
25             Simmonds, P. G. (1995) Atmospheric trends and lifetime of CH3CC13 and global OH concentrations. Science
26             (Washington, DC) 269: 187-192.
27       Rigel, D. S.; Rigel, E. G.; Rigel, A. C. (1999) Effects of altitude and latitude on ambient UVB radiation. J. Am.
28             Acad. Dermatol. 40: 114-116.
29       Rivers, J. K. (2004) Is there more than one road to melanoma? Lancet 363: 728-730.
30       Robinson, E. S.; Hill, R. H., Jr.; Kripke, M. L.; Setlow, R. B. (2000) The Monodelphis melanoma model: initial
31             report on large ultraviolet A exposures of suckling young. Photochem. Photobiol. 71: 743-746.
32       Roelofs, G. J.;  Lelieveld, J.; Van Borland, R. (1997) A three-dimensional chemistry/general circulation model
3 3             simulation of anthropogenically derived ozone in the troposphere and its radiative climate forcing. J.
34             Geophys. Res. [Atmos.] 102: 23,389-23,401.
35       Rosenthal, F. S.; Phoon, C.; Bakalian, A. E.; Taylor, H. R. (1988) The ocular dose of ultraviolet radiation to outdoor
36             workers. Invest. Ophthalmol. Vis. Sci. 29: 649-656.
37       Rozema, J.; Manetas, Y.; Bjorn, L. O., eds. (2001) Responses of plants to UV-B radiation. Dordrecht, The
38             Netherlands: Kluwer Academic. [Advances in Vegetation Science, v. 18].
39       Rtinger, T. M.; Moller, K.; Jung, T.; Dekant, B. (2000) DNA damage formation, DNA repair, and survival after
40             exposure of DNA repair-proficient and nucleotide excision repair-deficient human lymphoblasts to UVA1
41             and UVB. Int. J. Radiat. Biol. 76: 789-797.
42       Salby, M. L. (1996) Fundamentals of atmospheric physics. New York, NY: Academic Press. (The International
43             Geophysics Series, Monographs and Textbooks, v. 61).
44       Sarasin, A. (1999) The molecular pathways of ultraviolet-induced carcinogenesis. Mutat. Res. 428: 5-10.
45       Sarkar, A. K. (2004) An evaluation of UV protection imparted by cotton fabrics dyed with natural colorants. BMC
46             Dermatol. 4: 15.
47       Schenker, M. B.; Orenstein, M.  R.; Samuels, S. J. (2002) Use of protective equipment among California farmers.
48             Am. J. Ind.  Med. 42: 455-464.
49       Selgrade, M. K.; Repacholi, M. H.; Koren, H. S. (1997) Ultraviolet radiation-induced immune modulation: potential
50             consequences for infectious, allergic, and autoimmune disease. Environ. Health Perspect. 105: 332-334.
51       Selgrade, M. K.; Smith, M. V.; Oberhelman-Bragg, L. J.; LeVee, G. J.; Koren, H. S.; Cooper, K. D. (2001) Dose
52             response for UV-induced immune suppression in people of color: differences based on erythema! reactivity
53             rather than skin pigmentation. Photochem. Photobiol. 74: 88-95.
54       Setlow, R. B.; Grist, E.; Thompson, K.; Woodhead, A. D. (1993) Wavelengths effective in induction of malignant
55             melanoma. Proc. Natl. Acad. Sci. U. S. A. 90: 6666.6670.
         August 2005                                    10-65        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Shindell, D. T.; Faluvegi, G. (2002) An exploration of ozone changes and their radiative forcing prior to the
  2             chlorofluorocarbon era. Atmos. Chem. Phys. 2: 363-374.
  3       Shoveller, J. A.; Lovato, C. Y.; Peters, L.; Rivers, J. K. (1998) Canadian National Survey on Sun Exposure &
  4             Protective Behaviours: adults at leisure. Cancer Prev. Control 2: 111-116.
  5       Slaper, H.; Velders, G. J. M; Daniel, J. S.; de Gruijl, F. R.; Van derLeun, J. C.  (1996) Estimates of ozone depletion
  6             and skin cancer incidence to examine the Vienna Convention achievements. Nature (London, U.K.) 384:
  7             256-258.
  8       Solanki, S. K.; Fligge, M. (2000) Reconstruction of past solar irradiance. Space Sci. Rev. 94:  127-138.
  9       Stern, R. S.; Liebman, E. J.; Vakeva, L. (1998) Oral psoralen and ultraviolet-A light (PUVA)  treatment of psoriasis
10             and persistent risk of nonmelanoma skin cancer. PUVA Follow-up Study. J. Natl. Cancer Inst. 90: 1278-1284.
11       Stevenson, D. S.; Johnson, C. E.; Collins, W. J.; Derwent, R. G.; Shine, K. P.; Edwards, J. M. (1998) Evolution of
12             tropospheric ozone radiative forcing. Geophys. Res. Lett. 25: 3819-3822.
13       Stevenson, D. S.; Johnson, C. E.; Collins, W. J.; Derwent, R. G.; Edwards, J. M. (2000) Future estimates of
14             tropospheric ozone radiative forcing and methane turnover - the impact of climate change. Geophys. Res.
15             Lett. 27: 2073-2076.
16       Studzinski, G. P.; Moore, D. C. (1995) Sunlight-can it prevent as well as cause cancer? Cancer Res. 55:  4014-4022.
17       Swelter, S. M. (2003) Dermatological perspectives of malignant melanoma. Surg. Clin. N. Am. 83: 77-95.
18       Taylor, H. R.; West,  S. K.; Rosenthal, F.  S.; Munoz, B.; Newland, H. S.; Abbey, H.; Emmett, E. A. (1988) Effect of
19             ultraviolet radiation on cataract formation. N. Engl. J. Med. 319: 1429-1433.
20       Thieden, E.; Agren, M. S.; Wulf, H. C. (2001) Solar UVR exposures of indoor workers in a working and a holiday
21             period assessed by personal dosimeters and sun exposure diaries. Photodermatol. Photoimmunol. Photomed.
22             17:249-255.
23       Thieden, E.; Philipsen, P. A.; Heydenreich, J.; Wulf, H. C. (2004a) UV radiation exposure related to age, sex,
24             occupation, and sun behavior based on time-stamped personal dosimeter readings. Arch. Dermatol. 140:
25             197-203.
26       Thieden, E.; Philipsen, P. A.; Sandby-Moller, J.; Heydenreich, J.; Wulf, H. C. (2004b) Proportion of lifetime UV
27             dose received by children, teenagers and adults based on time-stamped personal dosimetry. J. Invest.
28             Dermatol. 123: 1147-1150.
29       Tomescu, D.; Kavanagh, G.; Ha, T.; Campbell, H.; Melton, D. W. (2001) Nucleotide excision repair geneXPD
30             polymorphisms and genetic predisposition to melanoma. Carcinogenesis 22: 403-408.
31       Ullrich, S. E. (2005)  Mechanisms underlying UV-induced immune suppression. Mutat. Res. 571: 185-205.
32       U.S. Environmental Protection Agency. (1987) Assessing the  risks of trace gases that can modify the stratosphere, v.
33             1-8. Washington, DC: Office of Air and Radiation; report no. EPA 400/1-87/001A-H.
34       U.S. Environmental Protection Agency. (2004) Air quality criteria for particulate matter. Research Triangle Park,
3 5             NC: National  Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
36             http://cfpub.epa.gov/ncea/ [9 November, 2004].
37       United Nations Environment Programme (UNEP). (1986) Report of the international conference on the assessment
38             of the role of carbon dioxide and of other greenhouse gases in climate variations and associated impacts;
39             October 1985; Villach, Austria. Geneva, Switzerland: World Meteorological Organization; WMO no. 661.
40       Urbach, F. (1997) Ultraviolet radiation and skin cancer of humans.  J. Photochem. Photobiol. B 40: 3-7.
41       Van der Leun, J. C.;  De Gruijl, F. R. (1993) Influences of ozone depletion on human and animal health. In: Tevini,
42             M., ed. UV-B radiation and ozone depletion: Effects on humans, animals, plants, microorganisms, and
43             materials. Ann Arbor: Lewis Publisher; pp. 95-123.
44       Van Dorland, R.; Dentener, F. J.; Lelieveld, J. (1997)  Radiative forcing due to tropospheric ozone and sulfate
45             aerosols.  J. Geophys. Res. [Atmos.] 102: 28,079-28,100.
46       Veiered, M. B.;  Weiderpass, E.; Thorn, M.; Hansson, J.; Lund, E.; Armstrong, B.; Adami, H.-O. (2003) A
47             prospective study of pigmentation, sun exposure, and risk of cutaneous malignant melanoma in women. J.
48             Natl. Cancer Inst. 95: 1530-1538.
49       Vingarzan, R. (2004) A review of surface ozone background levels and trends. Atmos. Environ. 38: 3431-3442.
50       Vishvakarman, D.; Wong, J. C.; Boreham, B. W. (2001) Annual occupational exposure to ultraviolet radiation in
51             central Queensland. Health Phys. 81: 536-544.
52       Vitasa, B. C.; Taylor, H. R.; Strickland, P. T.; Rosenthal, F. S.; West, S.; Abbey, H.; Ng, S. K.; Munoz, B.; Emmett,
53             E. A. (1990) Association on nonmelanoma skin cancer and actinic keratosis with cumulative solar ultraviolet
54             exposure in Maryland watermen. Cancer 65:  2811-2817.
55       Waterston, K.; Nay smith, L.; Rees. J. L. (2004) Physiological variation in the erythemal response to ultraviolet
56             radiation and photoadaptation. J. Invest. Dermatol.  123: 958-964.
         August 2005                                    10-66         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Wayne, R. P. (2000) Chemistry of Atmospheres: an introduction to the chemistry of the atmospheres of Earth, the
  2             planets, and their satellites. 3rd ed. New York, NY: Oxford University Press, Inc.
  3       Wei, Q.; Lee, J. E.;  Gershenwald, J. E.; Ross, M. I.; Mansfield, P.  F.; Strom, S. S.; Wang, L. E.; Quo, Z.; Qiao, Y.;
  4             Amos, C. L;  Spitz, M. R.; Duvic, M. (2003) Repair of UV light-induced DNA damage and risk of cutaneous
  5             malignant melanoma. J. Natl. Cancer Inst. 95: 308-315.
  6       Wendisch, M.; Mayer, B. (2003)  Vertical distribution of spectral solar irradiance in the cloudless sky: a case study.
  7             Geophys. Res. Lett. 30: 10.1029/2002GL016529.
  8       Westerdahl, J.; Ingvar, C.; Masback, A.; Jonsson, N.; Olsson, H. (2000) Risk of cutaneous malignant melanoma in
  9             relation to use of sunbeds:  further evidence for UV-A carcinogenicity. Br. J. Cancer 82: 1593-1599.
10       Whiteman, D. C.; Valery, P.; McWhirter, W.; Green, A. C. (1997) Risk factors for childhood malanoma in
11             Queensland,  Australia. Int. J. Cancer 70: 26-31.
12       Whiteman, D. C.; Whiteman,  C. A.; Green, A.  C. (2001) Childhood sun exposure as a risk factor for melanoma: a
13             systematic review of epidemiologic studies. Cancer Causes Control 12: 69-82.
14       Whiteman, D. C.; Watt,  P.; Purdie, D. M.; Hughes, M.  C.; Hayward, N. K.; Green, A. C. (2003) Melanocytic nevi,
15             solar keratoses, and divergent pathways to cutaneous melanoma. J. Natl. Cancer Inst.
16       Wong, J. C.; Airey, D. K.; Fleming, R. A.  (1996) Annual reduction of solar UV exposure to the facial aea of outdoor
17             workers in southeast Queensland by wearing a hat. Photodermatol. Photoimmunol. Photomed.  12: 131-135.
18       World Meteorological Organization. (1988) Developing policies for responding to climatic change: a  summary of
19             the discussions and recommendations of workshops; September-October 1987; Villach, Austria; and
20             November 1987;  Bellagio, Austria.  Geneva, Switzerland: World Meteorological Organization;  report no.
21             WMO/TD; no. 225. [World Climate Impact Programme series report no. WCIP-1 ].
22       World Meteorological Organization (WMO). (1999) Scientific assessment of ozone depletion: 1998. Geneva,
23             Switzerland:  World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
24       World Meteorological Organization (WMO). (2002) Scientific assessment of ozone depletion: 2002. Geneva,
25             Switzerland:  United Nations Environment Programme. [Global Ozone and Monitoring Project; report no. 47].
26       Xenopoulos, M. A.; Schindler, D. W. (2001) Physical factors  determining ultraviolet radiation flux into ecosystems.
27             In: Cockell, C.; Blaustein,  A. R. Ecosystems, evolution, and ultraviolet radiation. New York, NY:  Springer;
28             pp. 36-62.
29       Yarnal, B.; Kalkstein, L. S.; Scheraga, J. D. (2000) Mid-Atlantic regional assessment of climate change impacts.
30             Clim. Res. (CR Special 7)  14: 153-269.
31       Yoshikawa, T.; Rae, V.; Bruins-Slot, W.; Van den Berg, J. W.; Taylor, J. R.; Streilein, J. W. (1990) Susceptibility to
3 2             effects of UVB radiation on induction of contact hypersensitivity as a risk factor for skin cancer in humans. J.
33             Invest. Dermatol. 95: 530-536.
34
         August 2005                                    10-67         DRAFT-DO NOT QUOTE OR CITE

-------
 i       11.  EFFECT OF OZONE ON MAN-MADE MATERIALS
 2
 3
 4          Ozone (O3) and other photochemical oxidants react with many economically important
 5     man-made materials, decreasing their useful life and aesthetic appearance.  Some materials
 6     known to be damaged by ozone include elastomers, fibers, dyes, and paints. This chapter
 7     provides a brief discussion of O3 effects on man-made materials, including denoting of damage
 8     mechanisms and, where possible, concentration-response relationships.  Much of what is known
 9     about ozone effects on man-made materials is derived from research conducted in the 1970's,
10     1980's, and early 1990's, with very little new research on the subject having been conducted
11     since then. Since only very limited new information has been published on effects of ozone on
12     materials, this chapter mainly summarizes key information assessed in the previous 1996 Air
13     Quality Criteria Document for Ozone and other photochemical oxidants (1996 O3 AQCD)
14     (U.S. Environmental Protection Agency, 1996) and provides detailed discussion of the very
15     limited new information that has become available since then. In the ensuing sections,
16     discussion is focused on ozone effects on: elastomers (Sect 11.1); textiles and fabrics (11.2);
17     dyes, pigments, and inks (11.3); artist's pigments (11.4); and  surface coatings (11.5). Evaluation
18     of the relevance and economic importance of O3 materials damage information, as it affects
19     productivity or cultural resources (such as museums), is beyond the scope of this chapter.  The
20     reader is referred to the previous criteria document (1996 O3 AQCD) for a more detailed
21     discussion of the earlier studies summarized below.
22
23
24     11.1  ELASTOMERS
25          The elastomeric compounds, natural rubber and synthetic polymers and copolymers of
26     butadiene, isoprene, and styrene, are particularly susceptible to even low levels of ozone.
27     Elastomeric compounds are long chain unsaturated organic molecules. Ozone damages these
28     compounds by breaking the molecular chain at the carbon-carbon double bond; a chain of three
29     oxygen atoms is added directly across the double bond, forming a five-membered ring structure
30     (Mueller and Stickney, 1970). The change in structure promotes the characteristic cracking of
31     stressed/stretched rubber called "weathering."  A 5% tensile strain will produce cracks on the

       August 2005                              11-1       DRAFT-DO NOT QUOTE OR CITE

-------
 1      surface of the rubber that increase in number with increased stress/stretching. The rate of crack
 2      growth is dependent on the degree of stress, the type of rubber compound, concentration, time of
 3      exposure, velocity, and temperature (Bradley and Haagen-Smit, 1951; Lake and Mente, 1992)
 4      (Gent and McGrath, 1965). Once cracking occurs, there is further penetration, additional
 5      cracking, and eventually mechanical weakening or stress relaxation (U.S. Environmental
 6      Protection Agency, 1996).  Razumovskii et al. (1988) demonstrated the effect of ozone on stress
 7      relaxation of polyisoprene vulcanizates.  A decrease in stress (stress relaxation) is caused by
 8      ozone-induced cracks in exposed elastomers resulting in irreversible changes in the elastomer
 9      dimensions and decreased tensile strength.
10           To counteract ozone effects on elastomers, antiozonants and wax are often added to the
11      elastomeric formulations during processing. An antiozonant is an additive used to protect a
12      polymer against the effects of ozone-induced degradation and, hence, is used mainly in diene
13      rubbers. Antiozonant protection works either (a) by providing a physical barrier to ozone
14      penetration via forming a thin surface film of an ozone-resisting wax or (b) by chemically
15      reacting with ozone or polymer ozonolysis products, as do aromatic diamines such as
16      p-phenylene diamine derivatives.  The antiozonant diffuses to the surface of the elastomeric
17      material, where it reacts with ozone faster than ozone reacts to break the molecular chain and the
18      carbon-carbon double bond, or the antiozonant diffuses to the surface of the material but is not
19      reactive with ozone and serves as a protective coating against ozone attack. The antiozonant
20      may also serve to scavenge ozone while also providing protective film against ozone attack
21      (Andries et al., 1979; Lattimer et al., 1984).
22           Most of the studies  on ozone effects on elastomers were designed to evaluate the
23      effectiveness of antiozonants in counteracting the rubber cracking produced by ozone exposure.
24      Consequently, many of the studies were conducted using ozone concentrations higher than those
25      typically found in the ambient air. Natural rubber strips exposed to high concentrations of ozone
26      (20,000 ppm) under stressed conditions cracked almost instantaneously and were broken within
27      1 sec. When the ozone concentration was lowered (0.02 to 0.46 ppm), the time to required to
28      produce cracks in the exposed rubber material was increased (Bradley and Haagen-Smit, 1951).
29      Lake and Mente (1992) studied the effect of temperature on ozone-induced elastomer cracking
30      and antiozonant protection on natural rubber, epoxidised natural rubber, and two acrylonitrile-
31      butadiene copolymers under constant strain. Temperatures ranged from -20 °C  to +70 °C. The

        August 2005                               11 -2        DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
elastomers were exposed to 0.05 to 1,000 ppm ozone for 16 h. Ozone cracking decreased at
lower ambient temperatures; however, diffusing of both chemical and wax antiozonants also
were 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 ozone-induced elastomer
cracking to estimate
were exposed for 24
concentrations were
ambient ozone concentrations. Two vulcanized natural rubber compounds
h to varying ozone concentrations under stressed conditions. Ozone
60, 80, 90, 100, and 120 ppb for durations of 2, 4, or 6 h. The 24 h average
ozone concentrations ranged from 3 1 to 57.5 ppb. There was a clear relationship between the
24-h average ozone concentration and the distribution of crack length frequencies on the rubber
surface. Table 11-1
natural rubber strips.


gives the average 24-h ozone concentration and lengths for two vulcanized



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




Crack Length (mm)
0.05-0.10
0.10-0.15
0.15-0.20
0.20 - 0.40
1% Antiozonant 4010NA # 0.5% Antiozonant 4010NA
Middle Zones Central Zones Middle Zones Central Zones
37.5 37.5 40.0 42.5
45.0 48.0 48.0 53.0
48.0 >57.5 >57.5 >57.5
>57.5 >57.5 >57.5 >57.5
        Ozone concentrations given in ppb.
        Adapted from Serrano et al. (1993).
1      11.2  TEXTILES AND FABRICS
2           Ozone can damage textiles and fabrics by methods similar to those associated with
3      elastomers.  Generally, synthetic fibers are less affected by ozone than natural fibers; however,
4      ozone contribution to the degradation of textiles and fabrics is not considered significant (U.S.
      August 2005
DRAFT-DO NOT QUOTE OR CITE

-------
1     Environmental Protection Agency, 1996).  A study reported by Bogaty et al. (1952) showed that
2     ozone affects moistened cloth more than dry cloth.  Scoured cotton duck cloth and commercially
3     bleached cotton print cloth were exposed to 20 to 60 ppb for 1,200 h (50 days). The rate of
4     deterioration was measured by the changes in cuprammonium fluidity values and the fabric
5     breaking strength.  At the end of the 1,200-h exposure, there was a 20% loss in breaking
6     strength. Table 11-2 list the changes in cuprammonium fluidity values for both fabrics.
7
                 Table 11-2. Cuprammonium Fluidity of Moist Cotton Cloth Exposed
                                       to 20 to 60 ppb Ozone

Duck Cloth




Bleached Print Cloth





Duration of Exposure (h)
0
200
680
960
1200
0
200
510
650
865
1500
Cuprammonium Fluidity (rhes)
2.6
2.8
4.0
6.8
9.5
8.2
8.7
9.4
12.0
12.7
16.5
        Adapted from Bogaty et al. (1952).
1      11.3  DYES, PIGMENTS, AND INKS
2           Ozone fading of textile dyes is diffusion-controlled; the rate of fading is controlled by the
3      diffusion of the dye to the fiber surface. Many textile dyes react with ozone; however, the rate
4      and severity of the ozone attack is influenced by the chemical nature of the textile fiber and the
5      manner in which the dye is applied. Ozone molecules break the aromatic ring portion of the dye
6      molecule, oxidizing the dye (U.S. Environmental Protection Agency, 1996). In case of aromatic
7      azo dyes, ozone attacks the aromatic rings and electron rich nitrogen atoms (Matsui et al., 1988).
8      Grosjean et al. (1987; 1988a,b) proposed a mechanism for reactions of ozone with indigo,
9      thioindigo, and dibromoindigo, alazarin, and curcumin dyes under dark conditions. Ozone

       August 2005                              11 -4        DRAFT-DO NOT QUOTE OR CITE

-------
 1      attaches to the dye molecule at the unsaturated carbon = carbon bond.  An ozone adduct is
 2      formed (1,2,3-trioxolane), followed by scission of the carbon-carbon bond and the subsequent
 3      formation of the corresponding Criegee biradical. A similar mechanism was proposed for the
 4      reaction of ozone with triphenylmethane colorant Basic Violet 14.  Ozone attacked Basic Violet
 5      14 at the carbon=carbon unsaturated bond and at the carbon-nitrogen unsaturated bond under
 6      dark conditions. Other members of the group of triphenylmethane colorants with unsaturated
 7      carbon-carbon bonds also are expected to be subject to ozone fading.  Tripheylmethane
 8      colorants that are expected to be ozone-fugitive include the amino-substituted cationic dyes
 9      (Malachite Green, Brilliant Green, Crystal Violet, Pararosaniline Chloride, Methyl Green, and
10      others) (Grosjean et al., 1989).
11          An indication that ozone caused textile dye fading was first reported by Salvin and Walker
12      (1955). The researchers found that the fading was primarily the result of the destruction of the
13      blue dye molecule. Drapes made of acetate, Arnel,  and Dacron and dyed with  anthraquinone
14      blue dye exhibited a decrease in shade that was not  accompanied by the characteristic reddening
 15      caused by NOX.  Figures 11-1  and 11-2 demonstrate the effect of ozone exposure on nylon 6 yarn
16      colored with several blue dyes. Nylon samples inside the home were located on a wall away
17      from sunlight. Outside nylon  samples were placed  on a covered patio or under the eaves of the
18      house to minimize exposure to sunlight and rain. Ozone  concentrations ranged from 2 to 5 ppb
19      outside and  0 to 2 ppb inside.  The percent change in dye color was determined monthly by
20      extraction and analysis of the remaining dye or by instrumental measurement of the color change
21      (Haylock and Rush, 1978).
22
23
24      11.4  ARTISTS'PIGMENTS
25          Several artists' pigments are sensitive to fading and oxidation by ozone when exposed to
26      concentrations found in urban areas (Shaver et al., 1983;  Drisko et al.,  1985; Whitmore et al.,
27      1987; Whitmore and Cass, 1988; Grosjean et al., 1993).  The organic pigments that are ozone
28      fugitive include alizarin red pigments  containing lakes of the poly cyclic aromatic compound
29      1,2-dihydroxyanthraquinone, blue-violet pigments containing substituted triphenylmethane
30      lakes, indigo, and yellow coloring agents containing polyfunctional, polyunsaturated compounds
31      such as curcumin (Grosjean et al., 1987). Because of the potential of ozone to damage works of

        August 2005                              11-5         DRAFT-DO NOT  QUOTE OR CITE

-------
                                   100 r
                                      J  FMAMJ  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).
 1     art, recommended limits on ozone concentrations in museums, libraries, and archives are
 2     relatively low, ranging from 0.013 to 0.01 ppm.
 3          Experimental studies demonstrate a concentration x time (C x T) relationship between
 4     ozone concentration, exposure time, and pigment fading.  Cass et al. (1991) summarized some of
 5     the earlier research on the effects of ozone on artists' pigments.  In studies evaluating the effect
 6     of ozone on organic and inorganic watercolors and traditional organic pigments, only the
 7     traditional organic pigments showed measurable fading from ozone exposure.  Of the inorganic
 8     pigments tested, only the arsenic sulfides showed ozone-related  changes. The pigments were
 9     exposed to 0.3 to 0.4 ppm ozone for 3 mo in the absence of light, at 22 °C and 50% RH. The
10     authors equated this exposure to a C x T of 6 to 8 years inside a  Los Angeles museum with air
11     conditioning but without a pollutant removal system.
       August 2005
11-6
DRAFT-DO NOT QUOTE OR CITE

-------
                                    100 r
                                    90 -
                                       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).
 1          Whitmore and Cass (1988) studied the effect of ozone on traditional Japanese colorants.
 2     Most of these compounds are insoluble metal salts that are stable in light and air. Suspensions or
 3     solutions of the colorants were airbrushed on hot-pressed watercolor paper or silk cloths.
 4     A sample of Japanese woodblock print also was included in the analysis. Samples were exposed
 5     to 0.4 ppm ozone at 22 °C, 50% relative humidity, in the absence of light for 12 wk. Changes in
 6     reflectance spectra were used to evaluate the effect of ozone exposure on colorant fading.
 7     Among the colorants tested on paper, curmin, indigo, madder lake, and lac lake were the most
 8     sensitive to ozone exposure. Gamboge was relatively insensitive to ozone.  The blue and green
 9     areas of the sample from the woodblock print was very reactive due to the indigo dye ozone
10     sensitivity. The other colorants, red, yellow, and purple, showed very little sensitivity to ozone.
       August 2005
11-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1      The textiles dyes that reacted with ozone were indigo, alone or in combination with several
 2      yellow dyes.
 3           Ye et al. (2000) reported the rate of ozone fading of traditional Chinese plant dyes. Twelve
 4      different colorants were applied to watercolor paper and silk and exposed to 0.4 ppm ozone at
 5      25 °C, at 50% RH, in the absence of light for 22 wks.  Dye fading was greater when the colorant
 6      was applied to the watercolor paper compared to the silk cloth due to the darker initial depth of
 7      the shade, the greater saturation of the colorant throughout the cloth. Turn eric, gromwell, and
 8      violet on paper was particularly reactive. Tangerine peel was moderately reactive and sappan
 9      wood, dalbergia wood, Chinese gall, indigo, and Chinese yellow cork tree were slightly reactive
10      to ozone.  Black tea was not reactive to ozone. The colorants on silk samples showing color
11      changes were gromwell, sappan wood, gardenia, tummeric, and violet.  Figures  11-3 and 11-4
12      demonstrate the color change of the various colorants on watercolor paper and silk.
13           Artists' pigments also have exhibited fading when exposed to a mixture of photochemical
14      oxidants.  Grosjean et al. (1993) exposed 35 artists' pigments to a mixture of photochemical
15      oxidants consisting of ozone, nitrogen dioxide (NO2), and peroxyacetyl nitrate (PAN) for
16      12 wks. Weekly average photochemical concentrations were 200 ppb for ozone, 56 ± 12 to
17      99 ± 24 for NO2, and 11 ± 3 to 18 ± 2 for PAN. All exposures were carried out at room
18      temperature in the absence of light. To determine the effect of humidity on pigment fading, the
19      relative humidity was increased from 46% after 8 weeks of exposure to 83% for a 2 week period
20      and then returned to 46% for the remainder of the exposure.
21           Table  11-3 lists the artists' pigment and degree of fading. Eleven of the pigments
22      exhibited negligible color change, 12 had small color changes, 3 had modest color changes, and
23      9 exhibited substantial color changes.  Fading of Disperse Blue 3  and Reactive Blue 2 were
24      likely the result of NO2 exposure, the fading of triphenylmethanes is consistent with exposure to
25      nitric acid formed under high humidity conditions. Fading of the indigos was dominated by
26      ozone exposure and curcumin was faded by all of the photochemicals studied. Increasing the
27      relative humidity resulted in a substantial color change for all of the pigments, with the
28      exception of curcumin and indigo.
29
30
        August 2005                              11-8        DRAFT-DO NOT QUOTE OR CITE

-------
                                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 ozone at 25 °C ± 1 °C, 50% RH, in the absence of light.

Source: Ye et al. (2000).
August 2005
11-9
DRAFT-DO NOT QUOTE OR CITE

-------
                              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 (A£" 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 3)


 Disperse Blue 3

 French Ultramarine Blue

 Gamboge (Natural Yellow 24)

 Hooker's Green Light


 Indigo (a formulation)


 Indigo carmine *

 Indigo (73000)*

 Mauve


 New Gamboge
       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
August 2005
             11-11
DRAFT-DO NOT QUOTE OR CITE

-------
                   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).
1      11.5  SURFACE COATINGS

2           Ozone will act to erode some surface coatings (paints, varnishes, and lacquers). However,

3      many of the available studies on ozone degradation of surface coatings do not separate the

4      effects of ozone from other pollutants or environmental factors such as weather, humidity, and

5      temperature.  Campbell et al. (1974) attempted to demonstrate an ozone related effect on oil

6      house paint, acrylic latex coating, alkyd industrial maintenance coating, urea alkyd coil coating,

7      and nitrocellulose/acrylic automotive paint.  Painted test panels were exposed to 100 and
       August 2005
             11-12
DRAFT-DO NOT QUOTE OR CITE

-------
 1      1,000 ppb ozone in a xenon arc accelerated weathering chamber for up to 1,000 h. Using weight
 2      loss as a measure of ozone-induced erosion the researchers concluded that all of the paints tested
 3      suffered degradation in the presence of ozone and that the automotive finish suffered the most
 4      ozone-induce degradation. When ozone degradation was measured using scanning electron
 5      microscopy, the oil house paint and latex coating samples showed erosion above that seen with
 6      clean air but only at the highest exposure level.  No effects were noted for the automotive paint.
 7      The other painted surfaces were not evaluated.
 8           Spence et al. (1975) studied the effect of air pollutants and relative humidity on oil based
 9      house paint, acrylic latex house paint, acrylic coil coating, and vinyl coil coating under
10      laboratory conditions. Test panels were exposed in weathering chambers equipped with a xenon
11      are light for simulating sunlight to low and high levels of ozone (0.08 and 0.5 ppm), sulfur
12      dioxide (0.03 and 0.5 ppm), and nitrogen dioxide (0.05 and 0.5 ppm) and relative humidity
13      (50 and 90%).  Samples were exposed for a total of 1000 h. The exposure cycle consisted of
14      20 min of dew and 20 min of light.  The effects of the exposure on the painted surfaces were
15      measured by  weight loss and loss in film thickness.  The acrylic coil coating had the lowest
16      erosion rate of the surface coatings tested. However, ozone was the only pollutant that had a
17      significant effect on the surface erosion.  Sulfur dioxide and relative humidity were significant
18      factors in the erosion of oil base house paints and vinyl coil coating.  The findings for acrylic
19      latex house paint were not reported.
20
21
22      11.6  CONCLUSIONS
23           Ozone and other photochemical oxidants react with many economically important
24      man-made materials, decreasing their useful life and aesthetic appearance. Some materials
25      known to be damaged by ozone include elastomers, fibers and dyes, and paints. Most studies
26      have been on single compounds rather than complex materials.
27           The elastomeric compounds, natural rubber and synthetic polymers and copolymers of
28      butadiene, isoprene, and styrene, are particularly susceptible to even low concentrations of
29      ozone.  Ozone damages these compounds by breaking the molecular chain at the carbon-carbon
30      double bond; a  chain of three oxygen atoms is added directly across the double bond.  The
31      change in structure promotes the characteristic cracking of stressed/stretched rubber called

        August 2005                               11-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1      "weathering." Tensile strain produces cracks on the surface of the rubber that increase in
 2      number with increased stress/stretching.
 3           The rate of crack growth is dependent on the degree of stress, the type of rubber
 4      compound, ozone concentration, time of exposure, ozone velocity, and temperature. After initial
 5      cracking, there is further ozone penetration, resulting in additional cracking and, eventually,
 6      mechanical weakening or stress relaxation.
 7           Ozone can damage textiles and fabrics by methods similar to those associated with
 8      elastomers.  Generally, synthetic fibers are less affected by ozone than natural fibers; however,
 9      ozone contribution to the degradation of textiles and fabrics is not considered significant.
10           Ozone fading of textile dyes is a diffusion-controlled process; the rate of fading is
11      controlled by the diffusion of the dye to the fiber surface. Many textile dyes react with ozone.
12      The rate and severity of the ozone attack is influenced by the chemical nature of the textile fiber
13      and the manner in which the dye is applied.
14           Several artists' pigments are also sensitive to fading and oxidation by ozone when exposed
15      to concentrations found in urban areas.
16
        August 2005                               11-14        DRAFT-DO NOT QUOTE OR CITE

-------
  1       REFERENCES

  2       Andries, J. C.; Rhee, C. K.; Smith, R. W.; Ross, D. B.; Diem, H. E. (1979) A surface study of ozone attack and
  3             antiozonant protection of carbon black loaded natural rubber compounds. Rubber Chem. Technol.
  4             52: 823-837.
  5       Bogaty, H.; Campbell, K. S.; Appel, W. D. (1952) The oxidation of cellulose by ozone in small concentrations.
  6             Text. Res. J. 22: 81-83.
  7       Bradley, C. E.; Haagen-Smit, A. J. (1951) The application of rubber in the quantitative determination of ozone.
  8             Rubber Chem. Technol. 24: 750-755.
  9       Campbell, G. G.; Schurr, G. G.; Slawikowski, D. E.; Spence, J. W. (1974) Assessing air pollution damage to
10             coatings. J. Paint Technol. 46: 59-71.
11       Cass, G. R.; Nazaroff, W. W.; Tiller, C.; Whitmore, P. M. (1991) Protection of works of art from damage due to
12             atmospheric ozone. Atmos. Environ. Part A 25: 441-451.
13       Drisko, K.; Cass, G. R.; Whitmore, P. M.; Druzik, J. R. (1985) Fading of artists' pigments due to atmospheric ozone.
14             In: Vendl, A.; Pichler, B.; Weber, J.; Banik, G., eds. Wiener Berichte tiber Naturwissenschaft in der Kunst:
15             Doppelband 2/3. Vienna, Austria: Verlag ORAC; pp. 66-87.
16       Gent, A. N.; McGrath, J. E. (1965) Effect of temperature on ozone cracking of rubbers. J. Polymer Sci. A
17             3:  1473-1482.
18       Grosjean, D.; Whitmore, P. M.; De Moor, C. P.; Cass, G. R.; Druzik, J. R. (1987) Fading of alizarin and related
19             artists' pigments by atmospheric ozone: reaction products and mechanisms. Environ. Sci. Technol.
20             21:635-643.
21       Grosjean, D.; Whitmore, P. M.; Cass, G. R.; Druzik, J. R. (1988a) Ozone fading of natural organic colorants:
22             mechanisms and products of the reaction of ozone with indigos. Environ. Sci. Technol. 22: 292-298.
23       Grosjean, D.; Whitmore, P. M.; De Moor, C. P.; Cass, G. R.; Druzik, J. R. (1988b) Ozone fading of organic
24             colorants: products and mechanism of the reaction of ozone with curcumin. Environ.  Sci. Technol.
25             22: 1357-1361.
26       Grosjean, D.; Whitmore, P. M.; Cass, G. R.; Druzik, J. R. (1989) Ozone fading of triphenylmethane colorants:
27             reaction products and mechanisms. Environ. Sci. Technol. 23: 1164-1167.
28       Grosjean, D.; Grosjean, E.; Williams, E. L., II. (1993) Fading of artists' colorants by a mixture of photochemical
29             oxidants. Atmos. Environ. Part A 27: 765-772.
30       Haylock, J. C.; Rush, J. L. (1978) Studies on the ozone fading of anthraquinone dyes on nylon fibers. Part II:
31             In-service performance. Text. Res. J. 48: 143-149.
32       Lake, G. J.; Mente, P. G. (1992) Ozone cracking and protection of elastomers at high and low temperatures. J. Nat.
33             Rubber Res. 7: 1-13.
34       Lattimer, R. P.; Layer, R. W.; Rhee, C. K. (1984) Mechanisms of antiozonant protection: antiozonant-rubber
3 5             reactions during ozone exposure. In: 32nd annual conference on mass spectrometry and allied topics.
36             May/June; San Antonio, TX. Bethesda, MD:  American Society for Mass Spectrometry; pp. 357-358.
37       Matsui, M.; Koike, T.; Shibata, K. (1988) Ozone fading of phenolphthalein and aurin. J. Soc. Dyers Colour
38             104:482-486.
39       Mueller, W. J.; Stickney, P. B. (1970) A survey and economic assessment of the effects of air pollution on
40             elastomers:  final report. Columbus, OH: Battelle Memorial Institute, Columbus Laboratories; National Air
41             Pollution Control Administration contract no. CPA-22-69-146.
42       Razumovskii, S. D.; Podmasteriev, V. V.; Zaikov, G. E.  (1988) Kinetics and mechanism of stress relaxation of
43             polyisoprene vulcanizates under ozone ageing. Polym. Degrad. Stab. 20: 37-47.
44       Salvin, V. S.; Walker, R. A. (1955) Service fading of disperse dyestuffs by chemical agents  other than the oxides of
45             nitrogen. Text. Res. J. 25: 571-585.
46       Serrano, E.; Castro, M.; Macias, A. (1993) An improved direct method of rubber cracking analysis for estimating
47             24-hour ozone levels. Atmos. Environ. Part A 27: 431-442.
48       Shaver, C. L.; Cass, G. R.; Druzik, J. R. (1983) Ozone and the deterioration of works of art. Environ. Sci. Technol.
49             17:748-752.
50       Spence, J. W.; Haynie, F.; Upham, J. B. (1975) Effects of gaseous pollutants on paints: a chamber study. J. Paint
51             Technol. 47: 57-63.
52       U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
53             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF. 3v.
54             Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
55             http://cfpub2.epa.gov/ncea/.
         August 2005                                   11-15         DRAFT-DO NOT QUOTE OR CITE

-------
1      Whitmore, P. M; Cass, G. R. (1988) The ozone fading of traditional Japanese colorants. Stud. Conserv. 33: 29-40.
2      Whitmore, P. M.; Cass, G. R.; Druzik, J. R. (1987) The ozone fading of traditional natural organic colorants on
3            paper. J. Am. Inst. Conserv. 26: 45-58.
4      Ye, Y.; Salmon, L.  G.; Cass, G. R. (2000) The ozone fading of traditional Chinese plant dyes. J. Am. Inst. Conserv.
5            39:245-257.
       August 2005                                  11-16        DRAFT-DO NOT QUOTE OR CITE

-------
wEPA
                            Please make all necessary changes in the below label,           PRESORTED STANDARD
                            detach copy or copy, and return to the address in the upper left-      POSTAGE & FEES PAID
                            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-05/004bB
August 2005

-------